1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465666768697071727374757677787980818283848586878889909192939495969798991001011021031041051061071081091101111121131141151161171181191201211221231241251261271281291301311321331341351361371381391401411421431441451461471481491501511521531541551561571581591601611621631641651661671681691701711721731741751761771781791801811821831841851861871881891901911921931941951961971981992002012022032042052062072082092102112122132142152162172182192202212222232242252262272282292302312322332342352362372382392402412422432442452462472482492502512522532542552562572582592602612622632642652662672682692702712722732742752762772782792802812822832842852862872882892902912922932942952962972982993003013023033043053063073083093103113123133143153163173183193203213223233243253263273283293303313323333343353363373383393403413423433443453463473483493503513523533543553563573583593603613623633643653663673683693703713723733743753763773783793803813823833843853863873883893903913923933943953963973983994004014024034044054064074084094104114124134144154164174184194204214224234244254264274284294304314324334344354364374384394404414424434444454464474484494504514524534544554564574584594604614624634644654664674684694704714724734744754764774784794804814824834844854864874884894904914924934944954964974984995005015025035045055065075085095105115125135145155165175185195205215225235245255265275285295305315325335345355365375385395405415425435445455465475485495505515525535545555565575585595605615625635645655665675685695705715725735745755765775785795805815825835845855865875885895905915925935945955965975985996006016026036046056066076086096106116126136146156166176186196206216226236246256266276286296306316326336346356366376386396406416426436446456466476486496506516526536546556566576586596606616626636646656666676686696706716726736746756766776786796806816826836846856866876886896906916926936946956966976986997007017027037047057067077087097107117127137147157167177187197207217227237247257267277287297307317327337347357367377387397407417427437447457467477487497507517527537547557567577587597607617627637647657667677687697707717727737747757767777787797807817827837847857867877887897907917927937947957967977987998008018028038048058068078088098108118128138148158168178188198208218228238248258268278288298308318328338348358368378388398408418428438448458468478488498508518528538548558568578588598608618628638648658668678688698708718728738748758768778788798808818828838848858868878888898908918928938948958968978988999009019029039049059069079089099109119129139149159169179189199209219229239249259269279289299309319329339349359369379389399409419429439449459469479489499509519529539549559569579589599609619629639649659669679689699709719729739749759769779789799809819829839849859869879889899909919929939949959969979989991000100110021003100410051006100710081009101010111012101310141015101610171018101910201021102210231024102510261027102810291030103110321033103410351036103710381039104010411042104310441045104610471048104910501051105210531054105510561057105810591060106110621063106410651066106710681069107010711072107310741075107610771078107910801081108210831084108510861087108810891090109110921093109410951096109710981099110011011102110311041105110611071108110911101111111211131114111511161117111811191120112111221123112411251126112711281129113011311132113311341135113611371138113911401141114211431144114511461147114811491150115111521153115411551156115711581159116011611162116311641165116611671168116911701171117211731174117511761177117811791180118111821183118411851186118711881189119011911192119311941195119611971198119912001201120212031204120512061207120812091210121112121213121412151216121712181219122012211222122312241225122612271228122912301231123212331234123512361237123812391240124112421243124412451246124712481249125012511252125312541255125612571258125912601261126212631264126512661267126812691270127112721273127412751276127712781279128012811282128312841285128612871288128912901291129212931294129512961297129812991300130113021303130413051306130713081309131013111312131313141315131613171318131913201321132213231324132513261327132813291330133113321333133413351336133713381339134013411342134313441345134613471348134913501351135213531354135513561357135813591360136113621363136413651366136713681369137013711372137313741375137613771378137913801381138213831384138513861387138813891390139113921393139413951396139713981399140014011402140314041405140614071408140914101411141214131414141514161417141814191420142114221423142414251426142714281429143014311432143314341435143614371438143914401441144214431444144514461447144814491450145114521453145414551456145714581459146014611462146314641465146614671468146914701471147214731474147514761477147814791480148114821483148414851486148714881489149014911492149314941495149614971498149915001501150215031504150515061507150815091510151115121513151415151516151715181519152015211522152315241525152615271528152915301531153215331534153515361537153815391540154115421543154415451546154715481549155015511552155315541555155615571558155915601561156215631564156515661567156815691570157115721573157415751576157715781579158015811582158315841585158615871588158915901591159215931594159515961597159815991600160116021603160416051606160716081609161016111612161316141615161616171618161916201621162216231624162516261627162816291630163116321633163416351636163716381639164016411642164316441645164616471648164916501651165216531654165516561657165816591660166116621663166416651666166716681669167016711672167316741675167616771678167916801681168216831684168516861687168816891690169116921693169416951696169716981699170017011702170317041705170617071708170917101711171217131714171517161717171817191720172117221723172417251726172717281729173017311732173317341735173617371738173917401741174217431744174517461747174817491750175117521753175417551756175717581759176017611762176317641765176617671768176917701771177217731774177517761777177817791780178117821783178417851786178717881789179017911792179317941795179617971798179918001801180218031804180518061807180818091810181118121813181418151816181718181819182018211822182318241825182618271828182918301831183218331834183518361837183818391840184118421843184418451846184718481849185018511852185318541855185618571858185918601861186218631864186518661867186818691870187118721873187418751876187718781879188018811882188318841885188618871888188918901891189218931894189518961897189818991900190119021903190419051906190719081909191019111912191319141915191619171918191919201921192219231924192519261927192819291930193119321933193419351936193719381939194019411942194319441945194619471948194919501951195219531954195519561957195819591960196119621963196419651966196719681969197019711972197319741975197619771978197919801981198219831984198519861987198819891990199119921993199419951996199719981999200020012002200320042005200620072008200920102011201220132014201520162017201820192020202120222023202420252026202720282029203020312032203320342035203620372038203920402041204220432044204520462047204820492050205120522053205420552056205720582059206020612062206320642065206620672068206920702071207220732074207520762077207820792080208120822083208420852086208720882089209020912092209320942095209620972098209921002101210221032104210521062107210821092110211121122113211421152116211721182119212021212122212321242125212621272128212921302131213221332134213521362137213821392140214121422143214421452146214721482149215021512152215321542155215621572158215921602161216221632164216521662167216821692170217121722173217421752176217721782179218021812182218321842185218621872188218921902191219221932194219521962197219821992200220122022203220422052206220722082209221022112212221322142215221622172218221922202221222222232224222522262227222822292230223122322233223422352236223722382239224022412242224322442245224622472248224922502251225222532254225522562257225822592260226122622263226422652266226722682269227022712272227322742275227622772278227922802281228222832284228522862287228822892290229122922293229422952296229722982299230023012302230323042305230623072308230923102311231223132314231523162317231823192320232123222323232423252326232723282329233023312332233323342335233623372338233923402341234223432344234523462347234823492350235123522353235423552356235723582359236023612362236323642365236623672368236923702371237223732374237523762377237823792380238123822383238423852386238723882389239023912392239323942395239623972398239924002401240224032404240524062407240824092410241124122413241424152416241724182419242024212422242324242425242624272428242924302431243224332434243524362437243824392440244124422443244424452446244724482449245024512452245324542455245624572458245924602461246224632464246524662467246824692470247124722473247424752476247724782479248024812482248324842485248624872488248924902491249224932494249524962497249824992500250125022503250425052506250725082509251025112512251325142515251625172518251925202521252225232524252525262527252825292530253125322533253425352536253725382539254025412542254325442545254625472548254925502551255225532554255525562557255825592560256125622563256425652566256725682569257025712572257325742575257625772578257925802581258225832584258525862587258825892590259125922593259425952596259725982599260026012602260326042605260626072608260926102611261226132614261526162617261826192620262126222623262426252626262726282629263026312632263326342635263626372638263926402641264226432644264526462647264826492650265126522653265426552656265726582659266026612662266326642665266626672668266926702671267226732674267526762677267826792680268126822683268426852686268726882689269026912692269326942695269626972698269927002701270227032704270527062707270827092710271127122713271427152716271727182719272027212722272327242725272627272728272927302731273227332734273527362737273827392740274127422743274427452746274727482749275027512752275327542755275627572758275927602761276227632764276527662767276827692770277127722773277427752776277727782779278027812782278327842785278627872788278927902791279227932794279527962797279827992800280128022803280428052806280728082809281028112812281328142815281628172818281928202821282228232824282528262827282828292830283128322833283428352836283728382839284028412842284328442845284628472848284928502851285228532854285528562857285828592860286128622863286428652866286728682869287028712872287328742875287628772878287928802881288228832884288528862887288828892890289128922893289428952896289728982899290029012902290329042905290629072908290929102911291229132914291529162917291829192920292129222923292429252926292729282929293029312932293329342935293629372938293929402941294229432944294529462947294829492950295129522953295429552956295729582959296029612962296329642965296629672968296929702971297229732974297529762977297829792980298129822983298429852986298729882989299029912992299329942995299629972998299930003001300230033004300530063007300830093010301130123013301430153016301730183019302030213022302330243025302630273028302930303031303230333034303530363037303830393040304130423043304430453046304730483049305030513052305330543055305630573058305930603061306230633064306530663067306830693070307130723073307430753076307730783079308030813082308330843085308630873088308930903091309230933094309530963097309830993100310131023103310431053106310731083109311031113112311331143115311631173118311931203121312231233124312531263127312831293130313131323133313431353136313731383139314031413142314331443145314631473148314931503151315231533154315531563157315831593160316131623163316431653166316731683169317031713172317331743175317631773178317931803181318231833184318531863187318831893190319131923193319431953196319731983199320032013202320332043205320632073208320932103211321232133214321532163217321832193220322132223223322432253226322732283229323032313232323332343235323632373238323932403241324232433244324532463247324832493250325132523253325432553256325732583259326032613262326332643265326632673268326932703271327232733274327532763277327832793280328132823283328432853286328732883289329032913292329332943295329632973298329933003301330233033304330533063307330833093310331133123313331433153316331733183319332033213322332333243325332633273328332933303331333233333334333533363337333833393340334133423343334433453346334733483349335033513352335333543355335633573358335933603361336233633364336533663367336833693370337133723373337433753376337733783379338033813382338333843385338633873388338933903391339233933394339533963397339833993400340134023403340434053406340734083409341034113412341334143415341634173418341934203421342234233424342534263427342834293430343134323433343434353436343734383439344034413442344334443445344634473448344934503451345234533454345534563457345834593460346134623463346434653466346734683469347034713472347334743475347634773478347934803481348234833484348534863487348834893490349134923493349434953496349734983499350035013502350335043505350635073508350935103511351235133514351535163517351835193520352135223523352435253526352735283529353035313532353335343535353635373538353935403541354235433544354535463547354835493550355135523553355435553556355735583559356035613562356335643565356635673568356935703571357235733574357535763577357835793580358135823583358435853586358735883589359035913592359335943595359635973598359936003601360236033604360536063607360836093610361136123613361436153616361736183619362036213622362336243625362636273628362936303631363236333634363536363637363836393640364136423643364436453646364736483649365036513652365336543655365636573658365936603661366236633664366536663667366836693670367136723673367436753676367736783679368036813682368336843685368636873688368936903691369236933694369536963697369836993700370137023703370437053706370737083709371037113712371337143715371637173718371937203721372237233724372537263727372837293730373137323733373437353736373737383739374037413742374337443745374637473748374937503751375237533754375537563757375837593760376137623763376437653766376737683769377037713772377337743775377637773778377937803781378237833784378537863787378837893790379137923793379437953796379737983799380038013802380338043805380638073808380938103811381238133814381538163817381838193820382138223823382438253826382738283829383038313832383338343835383638373838383938403841384238433844384538463847384838493850385138523853385438553856385738583859386038613862386338643865386638673868386938703871387238733874387538763877387838793880388138823883388438853886388738883889389038913892389338943895389638973898389939003901390239033904390539063907390839093910391139123913391439153916391739183919392039213922392339243925392639273928392939303931393239333934393539363937393839393940394139423943394439453946394739483949395039513952395339543955395639573958395939603961396239633964396539663967396839693970397139723973397439753976397739783979398039813982398339843985398639873988398939903991399239933994399539963997399839994000400140024003400440054006400740084009401040114012401340144015401640174018401940204021402240234024402540264027402840294030403140324033403440354036403740384039404040414042404340444045404640474048404940504051405240534054405540564057405840594060406140624063406440654066406740684069407040714072407340744075407640774078407940804081408240834084408540864087408840894090409140924093409440954096409740984099410041014102410341044105410641074108410941104111411241134114411541164117411841194120412141224123412441254126412741284129413041314132413341344135413641374138413941404141414241434144414541464147414841494150415141524153415441554156415741584159416041614162416341644165416641674168416941704171417241734174417541764177417841794180418141824183418441854186418741884189419041914192419341944195419641974198419942004201420242034204420542064207420842094210421142124213421442154216421742184219422042214222422342244225422642274228422942304231423242334234423542364237423842394240424142424243424442454246424742484249425042514252425342544255425642574258425942604261426242634264426542664267426842694270427142724273427442754276427742784279428042814282428342844285428642874288428942904291429242934294429542964297429842994300430143024303430443054306430743084309431043114312431343144315431643174318431943204321432243234324432543264327432843294330433143324333433443354336433743384339434043414342434343444345434643474348434943504351435243534354435543564357435843594360436143624363436443654366436743684369437043714372437343744375437643774378437943804381438243834384438543864387438843894390439143924393439443954396439743984399440044014402440344044405440644074408440944104411441244134414441544164417441844194420442144224423442444254426442744284429443044314432443344344435443644374438443944404441444244434444444544464447444844494450445144524453445444554456445744584459446044614462446344644465446644674468446944704471447244734474447544764477447844794480448144824483448444854486448744884489449044914492449344944495449644974498449945004501450245034504450545064507450845094510451145124513451445154516451745184519452045214522452345244525452645274528452945304531453245334534453545364537453845394540454145424543454445454546454745484549455045514552455345544555455645574558455945604561456245634564456545664567456845694570457145724573457445754576457745784579458045814582458345844585458645874588458945904591459245934594459545964597459845994600460146024603460446054606460746084609461046114612461346144615461646174618461946204621462246234624462546264627462846294630463146324633463446354636463746384639464046414642464346444645464646474648464946504651465246534654465546564657465846594660466146624663466446654666466746684669467046714672467346744675467646774678467946804681468246834684468546864687468846894690469146924693469446954696469746984699470047014702470347044705470647074708470947104711471247134714471547164717471847194720472147224723472447254726472747284729473047314732473347344735473647374738473947404741474247434744474547464747474847494750475147524753475447554756475747584759476047614762476347644765476647674768476947704771477247734774477547764777477847794780478147824783478447854786478747884789479047914792479347944795479647974798479948004801480248034804480548064807480848094810481148124813481448154816481748184819482048214822482348244825482648274828482948304831483248334834483548364837483848394840484148424843484448454846484748484849485048514852485348544855485648574858485948604861486248634864486548664867486848694870487148724873487448754876487748784879488048814882488348844885488648874888488948904891489248934894489548964897489848994900490149024903490449054906490749084909491049114912491349144915491649174918491949204921492249234924492549264927492849294930493149324933493449354936493749384939494049414942494349444945494649474948494949504951495249534954495549564957495849594960496149624963496449654966496749684969497049714972497349744975497649774978497949804981498249834984498549864987498849894990499149924993499449954996499749984999500050015002500350045005500650075008500950105011501250135014501550165017501850195020502150225023502450255026502750285029503050315032503350345035503650375038503950405041504250435044504550465047504850495050505150525053505450555056505750585059506050615062506350645065506650675068506950705071507250735074507550765077507850795080508150825083508450855086508750885089509050915092509350945095509650975098509951005101510251035104510551065107510851095110511151125113511451155116511751185119512051215122512351245125512651275128512951305131513251335134513551365137513851395140514151425143514451455146514751485149515051515152515351545155515651575158515951605161516251635164516551665167516851695170517151725173517451755176517751785179518051815182518351845185518651875188518951905191519251935194519551965197519851995200520152025203520452055206520752085209521052115212521352145215521652175218521952205221522252235224522552265227522852295230523152325233523452355236523752385239524052415242524352445245524652475248524952505251525252535254525552565257525852595260526152625263526452655266526752685269527052715272527352745275527652775278527952805281528252835284528552865287528852895290529152925293529452955296529752985299530053015302530353045305530653075308530953105311531253135314531553165317531853195320532153225323532453255326532753285329533053315332533353345335533653375338533953405341534253435344534553465347534853495350535153525353535453555356535753585359536053615362536353645365536653675368536953705371537253735374537553765377537853795380538153825383538453855386538753885389539053915392539353945395539653975398539954005401540254035404540554065407540854095410541154125413541454155416541754185419542054215422542354245425542654275428542954305431543254335434543554365437543854395440544154425443544454455446544754485449545054515452545354545455545654575458545954605461546254635464546554665467546854695470547154725473547454755476547754785479548054815482548354845485548654875488548954905491549254935494549554965497549854995500550155025503550455055506550755085509551055115512551355145515551655175518551955205521552255235524552555265527552855295530553155325533553455355536553755385539554055415542554355445545554655475548554955505551555255535554555555565557555855595560556155625563556455655566556755685569557055715572557355745575557655775578557955805581558255835584558555865587558855895590559155925593559455955596559755985599560056015602560356045605560656075608560956105611561256135614561556165617561856195620562156225623562456255626562756285629563056315632563356345635563656375638563956405641564256435644564556465647564856495650565156525653565456555656565756585659566056615662566356645665566656675668566956705671567256735674567556765677567856795680568156825683568456855686568756885689569056915692569356945695569656975698569957005701570257035704570557065707570857095710571157125713571457155716571757185719572057215722572357245725572657275728572957305731573257335734573557365737573857395740574157425743574457455746574757485749575057515752575357545755575657575758575957605761576257635764576557665767576857695770577157725773577457755776577757785779578057815782578357845785578657875788578957905791579257935794579557965797579857995800580158025803580458055806580758085809581058115812581358145815581658175818581958205821582258235824582558265827582858295830583158325833583458355836583758385839584058415842584358445845584658475848584958505851585258535854585558565857585858595860586158625863586458655866586758685869587058715872587358745875587658775878587958805881588258835884588558865887588858895890589158925893589458955896589758985899590059015902590359045905590659075908590959105911591259135914591559165917591859195920592159225923592459255926592759285929593059315932593359345935593659375938593959405941594259435944594559465947594859495950595159525953595459555956595759585959596059615962596359645965596659675968596959705971597259735974597559765977597859795980598159825983598459855986598759885989599059915992599359945995599659975998599960006001600260036004600560066007600860096010601160126013601460156016601760186019602060216022602360246025602660276028602960306031603260336034603560366037603860396040604160426043604460456046604760486049605060516052605360546055605660576058605960606061606260636064606560666067606860696070607160726073607460756076607760786079608060816082608360846085608660876088608960906091609260936094609560966097609860996100610161026103610461056106610761086109611061116112611361146115611661176118611961206121612261236124612561266127612861296130613161326133613461356136613761386139614061416142614361446145614661476148614961506151615261536154615561566157615861596160616161626163616461656166616761686169617061716172617361746175617661776178617961806181618261836184618561866187618861896190619161926193619461956196619761986199620062016202620362046205620662076208620962106211621262136214621562166217621862196220622162226223622462256226622762286229623062316232623362346235623662376238623962406241624262436244624562466247624862496250625162526253625462556256625762586259626062616262626362646265626662676268626962706271627262736274627562766277627862796280628162826283628462856286628762886289629062916292629362946295629662976298629963006301630263036304630563066307630863096310631163126313631463156316631763186319632063216322632363246325632663276328632963306331633263336334633563366337633863396340634163426343634463456346634763486349635063516352635363546355635663576358635963606361636263636364636563666367636863696370637163726373637463756376637763786379638063816382638363846385638663876388638963906391639263936394639563966397639863996400640164026403640464056406640764086409641064116412641364146415641664176418641964206421642264236424642564266427642864296430643164326433643464356436643764386439644064416442644364446445644664476448644964506451645264536454645564566457645864596460646164626463646464656466646764686469647064716472647364746475647664776478647964806481648264836484648564866487648864896490649164926493649464956496649764986499650065016502650365046505650665076508650965106511651265136514651565166517651865196520652165226523652465256526652765286529653065316532653365346535653665376538653965406541654265436544654565466547654865496550655165526553655465556556655765586559656065616562656365646565656665676568656965706571657265736574657565766577657865796580658165826583658465856586658765886589659065916592659365946595659665976598659966006601660266036604660566066607660866096610661166126613661466156616661766186619662066216622662366246625662666276628662966306631663266336634663566366637663866396640664166426643664466456646664766486649665066516652665366546655665666576658665966606661666266636664666566666667666866696670667166726673667466756676667766786679668066816682668366846685668666876688668966906691669266936694669566966697669866996700670167026703670467056706670767086709671067116712671367146715671667176718671967206721672267236724672567266727672867296730673167326733673467356736673767386739674067416742674367446745674667476748674967506751675267536754675567566757675867596760676167626763676467656766676767686769677067716772677367746775677667776778677967806781678267836784678567866787678867896790679167926793679467956796679767986799680068016802680368046805680668076808680968106811681268136814681568166817681868196820682168226823682468256826682768286829683068316832683368346835683668376838683968406841684268436844684568466847684868496850685168526853685468556856685768586859686068616862686368646865686668676868686968706871687268736874687568766877687868796880688168826883688468856886688768886889689068916892689368946895689668976898689969006901690269036904690569066907690869096910691169126913691469156916691769186919692069216922692369246925692669276928692969306931693269336934693569366937693869396940694169426943694469456946694769486949695069516952695369546955695669576958695969606961696269636964696569666967696869696970697169726973697469756976697769786979698069816982698369846985698669876988698969906991699269936994699569966997699869997000700170027003700470057006700770087009701070117012701370147015701670177018701970207021702270237024702570267027702870297030703170327033703470357036703770387039704070417042704370447045704670477048704970507051705270537054705570567057705870597060706170627063706470657066706770687069707070717072707370747075707670777078707970807081708270837084708570867087708870897090709170927093709470957096709770987099710071017102710371047105710671077108710971107111711271137114711571167117711871197120712171227123712471257126712771287129713071317132713371347135713671377138713971407141714271437144714571467147714871497150715171527153715471557156715771587159716071617162716371647165716671677168716971707171717271737174717571767177717871797180718171827183718471857186718771887189719071917192719371947195719671977198719972007201720272037204720572067207720872097210721172127213721472157216721772187219722072217222722372247225722672277228722972307231723272337234723572367237723872397240724172427243724472457246724772487249725072517252725372547255725672577258725972607261726272637264726572667267726872697270727172727273727472757276727772787279728072817282728372847285728672877288728972907291729272937294729572967297729872997300730173027303730473057306730773087309731073117312731373147315731673177318731973207321732273237324732573267327732873297330733173327333733473357336733773387339734073417342734373447345734673477348734973507351735273537354735573567357735873597360736173627363736473657366736773687369737073717372737373747375737673777378737973807381738273837384738573867387738873897390739173927393739473957396739773987399740074017402740374047405740674077408740974107411741274137414741574167417741874197420742174227423742474257426742774287429743074317432743374347435743674377438743974407441744274437444744574467447744874497450745174527453745474557456745774587459746074617462746374647465746674677468746974707471747274737474747574767477747874797480748174827483748474857486748774887489749074917492749374947495749674977498749975007501750275037504750575067507750875097510751175127513751475157516751775187519752075217522752375247525752675277528752975307531753275337534753575367537753875397540754175427543754475457546754775487549755075517552755375547555755675577558755975607561756275637564756575667567756875697570757175727573757475757576757775787579758075817582758375847585758675877588758975907591759275937594759575967597759875997600760176027603760476057606760776087609761076117612761376147615761676177618761976207621762276237624762576267627762876297630763176327633763476357636763776387639764076417642764376447645764676477648764976507651765276537654765576567657765876597660766176627663766476657666766776687669767076717672767376747675767676777678767976807681768276837684768576867687768876897690769176927693769476957696769776987699770077017702770377047705770677077708770977107711771277137714771577167717771877197720772177227723772477257726772777287729773077317732773377347735773677377738773977407741774277437744774577467747774877497750775177527753775477557756775777587759776077617762776377647765776677677768776977707771777277737774777577767777777877797780778177827783778477857786778777887789779077917792779377947795779677977798779978007801780278037804780578067807780878097810781178127813781478157816781778187819782078217822782378247825782678277828782978307831783278337834783578367837783878397840784178427843784478457846784778487849785078517852785378547855785678577858785978607861786278637864786578667867786878697870787178727873787478757876787778787879788078817882788378847885788678877888788978907891789278937894789578967897789878997900790179027903790479057906790779087909791079117912791379147915791679177918791979207921792279237924792579267927792879297930793179327933793479357936793779387939794079417942794379447945794679477948794979507951795279537954795579567957795879597960796179627963796479657966796779687969797079717972797379747975797679777978797979807981798279837984798579867987798879897990799179927993799479957996799779987999800080018002800380048005800680078008800980108011801280138014801580168017801880198020802180228023802480258026802780288029803080318032803380348035803680378038803980408041804280438044804580468047804880498050805180528053805480558056805780588059806080618062806380648065806680678068806980708071807280738074807580768077807880798080808180828083808480858086808780888089809080918092809380948095809680978098809981008101810281038104810581068107810881098110811181128113811481158116811781188119812081218122812381248125812681278128812981308131813281338134813581368137813881398140814181428143814481458146814781488149815081518152815381548155815681578158815981608161816281638164816581668167816881698170817181728173817481758176817781788179818081818182818381848185818681878188818981908191819281938194819581968197819881998200820182028203820482058206820782088209821082118212821382148215821682178218821982208221822282238224822582268227822882298230823182328233823482358236823782388239824082418242824382448245824682478248824982508251825282538254825582568257825882598260826182628263826482658266826782688269827082718272827382748275827682778278827982808281828282838284828582868287828882898290829182928293829482958296829782988299830083018302830383048305830683078308830983108311831283138314831583168317831883198320832183228323832483258326832783288329833083318332833383348335833683378338833983408341834283438344834583468347834883498350835183528353835483558356835783588359836083618362836383648365836683678368836983708371837283738374837583768377837883798380838183828383838483858386838783888389839083918392839383948395839683978398839984008401840284038404840584068407840884098410841184128413841484158416841784188419842084218422842384248425842684278428842984308431843284338434843584368437843884398440844184428443844484458446844784488449845084518452845384548455845684578458845984608461846284638464846584668467846884698470847184728473847484758476847784788479848084818482848384848485848684878488848984908491849284938494849584968497849884998500850185028503850485058506850785088509851085118512851385148515851685178518851985208521852285238524852585268527852885298530853185328533853485358536853785388539854085418542854385448545854685478548854985508551855285538554855585568557855885598560856185628563856485658566856785688569857085718572857385748575857685778578857985808581858285838584858585868587858885898590859185928593859485958596859785988599860086018602860386048605860686078608860986108611861286138614861586168617861886198620862186228623862486258626862786288629863086318632863386348635863686378638863986408641864286438644864586468647864886498650865186528653865486558656865786588659866086618662866386648665866686678668866986708671867286738674867586768677867886798680868186828683868486858686868786888689869086918692869386948695869686978698869987008701870287038704870587068707870887098710871187128713871487158716871787188719872087218722872387248725872687278728872987308731873287338734873587368737873887398740874187428743874487458746874787488749875087518752875387548755875687578758875987608761876287638764876587668767876887698770877187728773877487758776877787788779878087818782878387848785878687878788878987908791879287938794879587968797879887998800880188028803880488058806880788088809881088118812881388148815881688178818881988208821882288238824882588268827882888298830883188328833883488358836883788388839884088418842884388448845884688478848884988508851885288538854885588568857885888598860886188628863886488658866886788688869887088718872887388748875887688778878887988808881888288838884888588868887888888898890889188928893889488958896889788988899890089018902890389048905890689078908890989108911891289138914891589168917891889198920892189228923892489258926892789288929893089318932893389348935893689378938893989408941894289438944894589468947894889498950895189528953895489558956895789588959896089618962896389648965896689678968896989708971897289738974897589768977897889798980898189828983898489858986898789888989899089918992899389948995899689978998899990009001900290039004900590069007900890099010901190129013901490159016901790189019902090219022902390249025902690279028902990309031903290339034903590369037903890399040904190429043904490459046904790489049905090519052905390549055905690579058905990609061906290639064906590669067906890699070907190729073907490759076907790789079908090819082908390849085908690879088908990909091909290939094909590969097909890999100910191029103910491059106910791089109911091119112911391149115911691179118911991209121912291239124912591269127912891299130913191329133913491359136913791389139914091419142914391449145914691479148914991509151915291539154915591569157915891599160916191629163916491659166916791689169917091719172917391749175917691779178917991809181918291839184918591869187918891899190919191929193919491959196919791989199920092019202920392049205920692079208920992109211921292139214921592169217921892199220922192229223922492259226922792289229923092319232923392349235923692379238923992409241924292439244924592469247924892499250925192529253925492559256925792589259926092619262926392649265926692679268926992709271927292739274927592769277927892799280928192829283928492859286928792889289929092919292929392949295929692979298929993009301930293039304930593069307930893099310931193129313931493159316931793189319932093219322932393249325932693279328932993309331933293339334933593369337933893399340934193429343934493459346934793489349935093519352935393549355935693579358935993609361936293639364936593669367936893699370937193729373937493759376937793789379938093819382938393849385938693879388938993909391939293939394939593969397939893999400940194029403940494059406940794089409941094119412941394149415941694179418941994209421942294239424942594269427942894299430943194329433943494359436943794389439944094419442944394449445944694479448944994509451945294539454945594569457945894599460946194629463946494659466946794689469947094719472947394749475947694779478947994809481948294839484948594869487948894899490949194929493949494959496949794989499950095019502950395049505950695079508950995109511951295139514951595169517951895199520952195229523952495259526952795289529953095319532953395349535953695379538953995409541954295439544954595469547954895499550955195529553955495559556955795589559956095619562956395649565956695679568956995709571957295739574957595769577957895799580958195829583958495859586958795889589959095919592959395949595959695979598959996009601960296039604960596069607960896099610961196129613961496159616961796189619962096219622962396249625962696279628962996309631963296339634963596369637963896399640964196429643964496459646964796489649965096519652965396549655965696579658965996609661966296639664966596669667966896699670967196729673967496759676967796789679968096819682968396849685968696879688968996909691969296939694969596969697969896999700970197029703970497059706970797089709971097119712971397149715971697179718971997209721972297239724972597269727972897299730973197329733973497359736973797389739974097419742974397449745974697479748974997509751975297539754975597569757975897599760976197629763976497659766976797689769977097719772977397749775977697779778977997809781978297839784978597869787978897899790979197929793979497959796979797989799980098019802980398049805980698079808980998109811981298139814981598169817981898199820982198229823982498259826982798289829983098319832983398349835983698379838983998409841984298439844984598469847984898499850985198529853985498559856985798589859986098619862986398649865986698679868986998709871987298739874987598769877987898799880988198829883988498859886988798889889989098919892989398949895989698979898989999009901990299039904990599069907990899099910991199129913991499159916991799189919992099219922992399249925992699279928992999309931993299339934993599369937993899399940994199429943994499459946994799489949995099519952995399549955995699579958995999609961996299639964996599669967996899699970997199729973997499759976997799789979998099819982998399849985998699879988998999909991999299939994999599969997999899991000010001100021000310004100051000610007100081000910010100111001210013100141001510016100171001810019100201002110022100231002410025100261002710028100291003010031100321003310034100351003610037100381003910040100411004210043100441004510046100471004810049100501005110052100531005410055100561005710058100591006010061100621006310064100651006610067100681006910070100711007210073100741007510076100771007810079100801008110082100831008410085100861008710088100891009010091100921009310094100951009610097100981009910100101011010210103101041010510106101071010810109101101011110112101131011410115101161011710118101191012010121101221012310124101251012610127101281012910130101311013210133101341013510136101371013810139101401014110142101431014410145101461014710148101491015010151101521015310154101551015610157101581015910160101611016210163101641016510166101671016810169101701017110172101731017410175101761017710178101791018010181101821018310184101851018610187101881018910190101911019210193101941019510196101971019810199102001020110202102031020410205102061020710208102091021010211102121021310214102151021610217102181021910220102211022210223102241022510226102271022810229102301023110232102331023410235102361023710238102391024010241102421024310244102451024610247102481024910250102511025210253102541025510256102571025810259102601026110262102631026410265102661026710268102691027010271102721027310274102751027610277102781027910280102811028210283102841028510286102871028810289102901029110292102931029410295102961029710298102991030010301103021030310304103051030610307103081030910310103111031210313103141031510316103171031810319103201032110322103231032410325103261032710328103291033010331103321033310334103351033610337103381033910340103411034210343103441034510346103471034810349103501035110352103531035410355103561035710358103591036010361103621036310364103651036610367103681036910370103711037210373103741037510376103771037810379103801038110382103831038410385103861038710388103891039010391103921039310394103951039610397103981039910400104011040210403104041040510406104071040810409104101041110412104131041410415104161041710418104191042010421104221042310424104251042610427104281042910430104311043210433104341043510436104371043810439104401044110442104431044410445104461044710448104491045010451104521045310454104551045610457104581045910460104611046210463104641046510466104671046810469104701047110472104731047410475104761047710478104791048010481104821048310484104851048610487104881048910490104911049210493104941049510496104971049810499105001050110502105031050410505105061050710508105091051010511105121051310514105151051610517105181051910520105211052210523105241052510526105271052810529105301053110532105331053410535105361053710538105391054010541105421054310544105451054610547105481054910550105511055210553105541055510556105571055810559105601056110562105631056410565105661056710568105691057010571105721057310574105751057610577105781057910580105811058210583105841058510586105871058810589105901059110592105931059410595105961059710598105991060010601106021060310604106051060610607106081060910610106111061210613106141061510616106171061810619106201062110622106231062410625106261062710628106291063010631106321063310634106351063610637// SPDX-License-Identifier: GPL-2.0
/*
 * Copyright (C) 2007 Oracle.  All rights reserved.
 */

#include <linux/kernel.h>
#include <linux/bio.h>
#include <linux/buffer_head.h>
#include <linux/file.h>
#include <linux/fs.h>
#include <linux/pagemap.h>
#include <linux/highmem.h>
#include <linux/time.h>
#include <linux/init.h>
#include <linux/string.h>
#include <linux/backing-dev.h>
#include <linux/mpage.h>
#include <linux/swap.h>
#include <linux/writeback.h>
#include <linux/compat.h>
#include <linux/bit_spinlock.h>
#include <linux/xattr.h>
#include <linux/posix_acl.h>
#include <linux/falloc.h>
#include <linux/slab.h>
#include <linux/ratelimit.h>
#include <linux/mount.h>
#include <linux/btrfs.h>
#include <linux/blkdev.h>
#include <linux/posix_acl_xattr.h>
#include <linux/uio.h>
#include <linux/magic.h>
#include <linux/iversion.h>
#include <asm/unaligned.h>
#include "ctree.h"
#include "disk-io.h"
#include "transaction.h"
#include "btrfs_inode.h"
#include "print-tree.h"
#include "ordered-data.h"
#include "xattr.h"
#include "tree-log.h"
#include "volumes.h"
#include "compression.h"
#include "locking.h"
#include "free-space-cache.h"
#include "inode-map.h"
#include "backref.h"
#include "props.h"
#include "qgroup.h"
#include "dedupe.h"

struct btrfs_iget_args {
	struct btrfs_key *location;
	struct btrfs_root *root;
};

struct btrfs_dio_data {
	u64 reserve;
	u64 unsubmitted_oe_range_start;
	u64 unsubmitted_oe_range_end;
	int overwrite;
};

static const struct inode_operations btrfs_dir_inode_operations;
static const struct inode_operations btrfs_symlink_inode_operations;
static const struct inode_operations btrfs_dir_ro_inode_operations;
static const struct inode_operations btrfs_special_inode_operations;
static const struct inode_operations btrfs_file_inode_operations;
static const struct address_space_operations btrfs_aops;
static const struct address_space_operations btrfs_symlink_aops;
static const struct file_operations btrfs_dir_file_operations;
static const struct extent_io_ops btrfs_extent_io_ops;

static struct kmem_cache *btrfs_inode_cachep;
struct kmem_cache *btrfs_trans_handle_cachep;
struct kmem_cache *btrfs_path_cachep;
struct kmem_cache *btrfs_free_space_cachep;

#define S_SHIFT 12
static const unsigned char btrfs_type_by_mode[S_IFMT >> S_SHIFT] = {
	[S_IFREG >> S_SHIFT]	= BTRFS_FT_REG_FILE,
	[S_IFDIR >> S_SHIFT]	= BTRFS_FT_DIR,
	[S_IFCHR >> S_SHIFT]	= BTRFS_FT_CHRDEV,
	[S_IFBLK >> S_SHIFT]	= BTRFS_FT_BLKDEV,
	[S_IFIFO >> S_SHIFT]	= BTRFS_FT_FIFO,
	[S_IFSOCK >> S_SHIFT]	= BTRFS_FT_SOCK,
	[S_IFLNK >> S_SHIFT]	= BTRFS_FT_SYMLINK,
};

static int btrfs_setsize(struct inode *inode, struct iattr *attr);
static int btrfs_truncate(struct inode *inode, bool skip_writeback);
static int btrfs_finish_ordered_io(struct btrfs_ordered_extent *ordered_extent);
static noinline int cow_file_range(struct inode *inode,
				   struct page *locked_page,
				   u64 start, u64 end, u64 delalloc_end,
				   int *page_started, unsigned long *nr_written,
				   int unlock, struct btrfs_dedupe_hash *hash);
static struct extent_map *create_io_em(struct inode *inode, u64 start, u64 len,
				       u64 orig_start, u64 block_start,
				       u64 block_len, u64 orig_block_len,
				       u64 ram_bytes, int compress_type,
				       int type);

static void __endio_write_update_ordered(struct inode *inode,
					 const u64 offset, const u64 bytes,
					 const bool uptodate);

/*
 * Cleanup all submitted ordered extents in specified range to handle errors
 * from the fill_dellaloc() callback.
 *
 * NOTE: caller must ensure that when an error happens, it can not call
 * extent_clear_unlock_delalloc() to clear both the bits EXTENT_DO_ACCOUNTING
 * and EXTENT_DELALLOC simultaneously, because that causes the reserved metadata
 * to be released, which we want to happen only when finishing the ordered
 * extent (btrfs_finish_ordered_io()). Also note that the caller of the
 * fill_delalloc() callback already does proper cleanup for the first page of
 * the range, that is, it invokes the callback writepage_end_io_hook() for the
 * range of the first page.
 */
static inline void btrfs_cleanup_ordered_extents(struct inode *inode,
						 const u64 offset,
						 const u64 bytes)
{
	unsigned long index = offset >> PAGE_SHIFT;
	unsigned long end_index = (offset + bytes - 1) >> PAGE_SHIFT;
	struct page *page;

	while (index <= end_index) {
		page = find_get_page(inode->i_mapping, index);
		index++;
		if (!page)
			continue;
		ClearPagePrivate2(page);
		put_page(page);
	}
	return __endio_write_update_ordered(inode, offset + PAGE_SIZE,
					    bytes - PAGE_SIZE, false);
}

static int btrfs_dirty_inode(struct inode *inode);

#ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS
void btrfs_test_inode_set_ops(struct inode *inode)
{
	BTRFS_I(inode)->io_tree.ops = &btrfs_extent_io_ops;
}
#endif

static int btrfs_init_inode_security(struct btrfs_trans_handle *trans,
				     struct inode *inode,  struct inode *dir,
				     const struct qstr *qstr)
{
	int err;

	err = btrfs_init_acl(trans, inode, dir);
	if (!err)
		err = btrfs_xattr_security_init(trans, inode, dir, qstr);
	return err;
}

/*
 * this does all the hard work for inserting an inline extent into
 * the btree.  The caller should have done a btrfs_drop_extents so that
 * no overlapping inline items exist in the btree
 */
static int insert_inline_extent(struct btrfs_trans_handle *trans,
				struct btrfs_path *path, int extent_inserted,
				struct btrfs_root *root, struct inode *inode,
				u64 start, size_t size, size_t compressed_size,
				int compress_type,
				struct page **compressed_pages)
{
	struct extent_buffer *leaf;
	struct page *page = NULL;
	char *kaddr;
	unsigned long ptr;
	struct btrfs_file_extent_item *ei;
	int ret;
	size_t cur_size = size;
	unsigned long offset;

	if (compressed_size && compressed_pages)
		cur_size = compressed_size;

	inode_add_bytes(inode, size);

	if (!extent_inserted) {
		struct btrfs_key key;
		size_t datasize;

		key.objectid = btrfs_ino(BTRFS_I(inode));
		key.offset = start;
		key.type = BTRFS_EXTENT_DATA_KEY;

		datasize = btrfs_file_extent_calc_inline_size(cur_size);
		path->leave_spinning = 1;
		ret = btrfs_insert_empty_item(trans, root, path, &key,
					      datasize);
		if (ret)
			goto fail;
	}
	leaf = path->nodes[0];
	ei = btrfs_item_ptr(leaf, path->slots[0],
			    struct btrfs_file_extent_item);
	btrfs_set_file_extent_generation(leaf, ei, trans->transid);
	btrfs_set_file_extent_type(leaf, ei, BTRFS_FILE_EXTENT_INLINE);
	btrfs_set_file_extent_encryption(leaf, ei, 0);
	btrfs_set_file_extent_other_encoding(leaf, ei, 0);
	btrfs_set_file_extent_ram_bytes(leaf, ei, size);
	ptr = btrfs_file_extent_inline_start(ei);

	if (compress_type != BTRFS_COMPRESS_NONE) {
		struct page *cpage;
		int i = 0;
		while (compressed_size > 0) {
			cpage = compressed_pages[i];
			cur_size = min_t(unsigned long, compressed_size,
				       PAGE_SIZE);

			kaddr = kmap_atomic(cpage);
			write_extent_buffer(leaf, kaddr, ptr, cur_size);
			kunmap_atomic(kaddr);

			i++;
			ptr += cur_size;
			compressed_size -= cur_size;
		}
		btrfs_set_file_extent_compression(leaf, ei,
						  compress_type);
	} else {
		page = find_get_page(inode->i_mapping,
				     start >> PAGE_SHIFT);
		btrfs_set_file_extent_compression(leaf, ei, 0);
		kaddr = kmap_atomic(page);
		offset = start & (PAGE_SIZE - 1);
		write_extent_buffer(leaf, kaddr + offset, ptr, size);
		kunmap_atomic(kaddr);
		put_page(page);
	}
	btrfs_mark_buffer_dirty(leaf);
	btrfs_release_path(path);

	/*
	 * we're an inline extent, so nobody can
	 * extend the file past i_size without locking
	 * a page we already have locked.
	 *
	 * We must do any isize and inode updates
	 * before we unlock the pages.  Otherwise we
	 * could end up racing with unlink.
	 */
	BTRFS_I(inode)->disk_i_size = inode->i_size;
	ret = btrfs_update_inode(trans, root, inode);

fail:
	return ret;
}


/*
 * conditionally insert an inline extent into the file.  This
 * does the checks required to make sure the data is small enough
 * to fit as an inline extent.
 */
static noinline int cow_file_range_inline(struct inode *inode, u64 start,
					  u64 end, size_t compressed_size,
					  int compress_type,
					  struct page **compressed_pages)
{
	struct btrfs_root *root = BTRFS_I(inode)->root;
	struct btrfs_fs_info *fs_info = root->fs_info;
	struct btrfs_trans_handle *trans;
	u64 isize = i_size_read(inode);
	u64 actual_end = min(end + 1, isize);
	u64 inline_len = actual_end - start;
	u64 aligned_end = ALIGN(end, fs_info->sectorsize);
	u64 data_len = inline_len;
	int ret;
	struct btrfs_path *path;
	int extent_inserted = 0;
	u32 extent_item_size;

	if (compressed_size)
		data_len = compressed_size;

	if (start > 0 ||
	    actual_end > fs_info->sectorsize ||
	    data_len > BTRFS_MAX_INLINE_DATA_SIZE(fs_info) ||
	    (!compressed_size &&
	    (actual_end & (fs_info->sectorsize - 1)) == 0) ||
	    end + 1 < isize ||
	    data_len > fs_info->max_inline) {
		return 1;
	}

	path = btrfs_alloc_path();
	if (!path)
		return -ENOMEM;

	trans = btrfs_join_transaction(root);
	if (IS_ERR(trans)) {
		btrfs_free_path(path);
		return PTR_ERR(trans);
	}
	trans->block_rsv = &BTRFS_I(inode)->block_rsv;

	if (compressed_size && compressed_pages)
		extent_item_size = btrfs_file_extent_calc_inline_size(
		   compressed_size);
	else
		extent_item_size = btrfs_file_extent_calc_inline_size(
		    inline_len);

	ret = __btrfs_drop_extents(trans, root, inode, path,
				   start, aligned_end, NULL,
				   1, 1, extent_item_size, &extent_inserted);
	if (ret) {
		btrfs_abort_transaction(trans, ret);
		goto out;
	}

	if (isize > actual_end)
		inline_len = min_t(u64, isize, actual_end);
	ret = insert_inline_extent(trans, path, extent_inserted,
				   root, inode, start,
				   inline_len, compressed_size,
				   compress_type, compressed_pages);
	if (ret && ret != -ENOSPC) {
		btrfs_abort_transaction(trans, ret);
		goto out;
	} else if (ret == -ENOSPC) {
		ret = 1;
		goto out;
	}

	set_bit(BTRFS_INODE_NEEDS_FULL_SYNC, &BTRFS_I(inode)->runtime_flags);
	btrfs_drop_extent_cache(BTRFS_I(inode), start, aligned_end - 1, 0);
out:
	/*
	 * Don't forget to free the reserved space, as for inlined extent
	 * it won't count as data extent, free them directly here.
	 * And at reserve time, it's always aligned to page size, so
	 * just free one page here.
	 */
	btrfs_qgroup_free_data(inode, NULL, 0, PAGE_SIZE);
	btrfs_free_path(path);
	btrfs_end_transaction(trans);
	return ret;
}

struct async_extent {
	u64 start;
	u64 ram_size;
	u64 compressed_size;
	struct page **pages;
	unsigned long nr_pages;
	int compress_type;
	struct list_head list;
};

struct async_cow {
	struct inode *inode;
	struct btrfs_root *root;
	struct page *locked_page;
	u64 start;
	u64 end;
	unsigned int write_flags;
	struct list_head extents;
	struct btrfs_work work;
};

static noinline int add_async_extent(struct async_cow *cow,
				     u64 start, u64 ram_size,
				     u64 compressed_size,
				     struct page **pages,
				     unsigned long nr_pages,
				     int compress_type)
{
	struct async_extent *async_extent;

	async_extent = kmalloc(sizeof(*async_extent), GFP_NOFS);
	BUG_ON(!async_extent); /* -ENOMEM */
	async_extent->start = start;
	async_extent->ram_size = ram_size;
	async_extent->compressed_size = compressed_size;
	async_extent->pages = pages;
	async_extent->nr_pages = nr_pages;
	async_extent->compress_type = compress_type;
	list_add_tail(&async_extent->list, &cow->extents);
	return 0;
}

static inline int inode_need_compress(struct inode *inode, u64 start, u64 end)
{
	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);

	/* force compress */
	if (btrfs_test_opt(fs_info, FORCE_COMPRESS))
		return 1;
	/* defrag ioctl */
	if (BTRFS_I(inode)->defrag_compress)
		return 1;
	/* bad compression ratios */
	if (BTRFS_I(inode)->flags & BTRFS_INODE_NOCOMPRESS)
		return 0;
	if (btrfs_test_opt(fs_info, COMPRESS) ||
	    BTRFS_I(inode)->flags & BTRFS_INODE_COMPRESS ||
	    BTRFS_I(inode)->prop_compress)
		return btrfs_compress_heuristic(inode, start, end);
	return 0;
}

static inline void inode_should_defrag(struct btrfs_inode *inode,
		u64 start, u64 end, u64 num_bytes, u64 small_write)
{
	/* If this is a small write inside eof, kick off a defrag */
	if (num_bytes < small_write &&
	    (start > 0 || end + 1 < inode->disk_i_size))
		btrfs_add_inode_defrag(NULL, inode);
}

/*
 * we create compressed extents in two phases.  The first
 * phase compresses a range of pages that have already been
 * locked (both pages and state bits are locked).
 *
 * This is done inside an ordered work queue, and the compression
 * is spread across many cpus.  The actual IO submission is step
 * two, and the ordered work queue takes care of making sure that
 * happens in the same order things were put onto the queue by
 * writepages and friends.
 *
 * If this code finds it can't get good compression, it puts an
 * entry onto the work queue to write the uncompressed bytes.  This
 * makes sure that both compressed inodes and uncompressed inodes
 * are written in the same order that the flusher thread sent them
 * down.
 */
static noinline void compress_file_range(struct inode *inode,
					struct page *locked_page,
					u64 start, u64 end,
					struct async_cow *async_cow,
					int *num_added)
{
	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
	u64 blocksize = fs_info->sectorsize;
	u64 actual_end;
	u64 isize = i_size_read(inode);
	int ret = 0;
	struct page **pages = NULL;
	unsigned long nr_pages;
	unsigned long total_compressed = 0;
	unsigned long total_in = 0;
	int i;
	int will_compress;
	int compress_type = fs_info->compress_type;
	int redirty = 0;

	inode_should_defrag(BTRFS_I(inode), start, end, end - start + 1,
			SZ_16K);

	actual_end = min_t(u64, isize, end + 1);
again:
	will_compress = 0;
	nr_pages = (end >> PAGE_SHIFT) - (start >> PAGE_SHIFT) + 1;
	BUILD_BUG_ON((BTRFS_MAX_COMPRESSED % PAGE_SIZE) != 0);
	nr_pages = min_t(unsigned long, nr_pages,
			BTRFS_MAX_COMPRESSED / PAGE_SIZE);

	/*
	 * we don't want to send crud past the end of i_size through
	 * compression, that's just a waste of CPU time.  So, if the
	 * end of the file is before the start of our current
	 * requested range of bytes, we bail out to the uncompressed
	 * cleanup code that can deal with all of this.
	 *
	 * It isn't really the fastest way to fix things, but this is a
	 * very uncommon corner.
	 */
	if (actual_end <= start)
		goto cleanup_and_bail_uncompressed;

	total_compressed = actual_end - start;

	/*
	 * skip compression for a small file range(<=blocksize) that
	 * isn't an inline extent, since it doesn't save disk space at all.
	 */
	if (total_compressed <= blocksize &&
	   (start > 0 || end + 1 < BTRFS_I(inode)->disk_i_size))
		goto cleanup_and_bail_uncompressed;

	total_compressed = min_t(unsigned long, total_compressed,
			BTRFS_MAX_UNCOMPRESSED);
	total_in = 0;
	ret = 0;

	/*
	 * we do compression for mount -o compress and when the
	 * inode has not been flagged as nocompress.  This flag can
	 * change at any time if we discover bad compression ratios.
	 */
	if (inode_need_compress(inode, start, end)) {
		WARN_ON(pages);
		pages = kcalloc(nr_pages, sizeof(struct page *), GFP_NOFS);
		if (!pages) {
			/* just bail out to the uncompressed code */
			goto cont;
		}

		if (BTRFS_I(inode)->defrag_compress)
			compress_type = BTRFS_I(inode)->defrag_compress;
		else if (BTRFS_I(inode)->prop_compress)
			compress_type = BTRFS_I(inode)->prop_compress;

		/*
		 * we need to call clear_page_dirty_for_io on each
		 * page in the range.  Otherwise applications with the file
		 * mmap'd can wander in and change the page contents while
		 * we are compressing them.
		 *
		 * If the compression fails for any reason, we set the pages
		 * dirty again later on.
		 *
		 * Note that the remaining part is redirtied, the start pointer
		 * has moved, the end is the original one.
		 */
		if (!redirty) {
			extent_range_clear_dirty_for_io(inode, start, end);
			redirty = 1;
		}

		/* Compression level is applied here and only here */
		ret = btrfs_compress_pages(
			compress_type | (fs_info->compress_level << 4),
					   inode->i_mapping, start,
					   pages,
					   &nr_pages,
					   &total_in,
					   &total_compressed);

		if (!ret) {
			unsigned long offset = total_compressed &
				(PAGE_SIZE - 1);
			struct page *page = pages[nr_pages - 1];
			char *kaddr;

			/* zero the tail end of the last page, we might be
			 * sending it down to disk
			 */
			if (offset) {
				kaddr = kmap_atomic(page);
				memset(kaddr + offset, 0,
				       PAGE_SIZE - offset);
				kunmap_atomic(kaddr);
			}
			will_compress = 1;
		}
	}
cont:
	if (start == 0) {
		/* lets try to make an inline extent */
		if (ret || total_in < actual_end) {
			/* we didn't compress the entire range, try
			 * to make an uncompressed inline extent.
			 */
			ret = cow_file_range_inline(inode, start, end, 0,
						    BTRFS_COMPRESS_NONE, NULL);
		} else {
			/* try making a compressed inline extent */
			ret = cow_file_range_inline(inode, start, end,
						    total_compressed,
						    compress_type, pages);
		}
		if (ret <= 0) {
			unsigned long clear_flags = EXTENT_DELALLOC |
				EXTENT_DELALLOC_NEW | EXTENT_DEFRAG |
				EXTENT_DO_ACCOUNTING;
			unsigned long page_error_op;

			page_error_op = ret < 0 ? PAGE_SET_ERROR : 0;

			/*
			 * inline extent creation worked or returned error,
			 * we don't need to create any more async work items.
			 * Unlock and free up our temp pages.
			 *
			 * We use DO_ACCOUNTING here because we need the
			 * delalloc_release_metadata to be done _after_ we drop
			 * our outstanding extent for clearing delalloc for this
			 * range.
			 */
			extent_clear_unlock_delalloc(inode, start, end, end,
						     NULL, clear_flags,
						     PAGE_UNLOCK |
						     PAGE_CLEAR_DIRTY |
						     PAGE_SET_WRITEBACK |
						     page_error_op |
						     PAGE_END_WRITEBACK);
			goto free_pages_out;
		}
	}

	if (will_compress) {
		/*
		 * we aren't doing an inline extent round the compressed size
		 * up to a block size boundary so the allocator does sane
		 * things
		 */
		total_compressed = ALIGN(total_compressed, blocksize);

		/*
		 * one last check to make sure the compression is really a
		 * win, compare the page count read with the blocks on disk,
		 * compression must free at least one sector size
		 */
		total_in = ALIGN(total_in, PAGE_SIZE);
		if (total_compressed + blocksize <= total_in) {
			*num_added += 1;

			/*
			 * The async work queues will take care of doing actual
			 * allocation on disk for these compressed pages, and
			 * will submit them to the elevator.
			 */
			add_async_extent(async_cow, start, total_in,
					total_compressed, pages, nr_pages,
					compress_type);

			if (start + total_in < end) {
				start += total_in;
				pages = NULL;
				cond_resched();
				goto again;
			}
			return;
		}
	}
	if (pages) {
		/*
		 * the compression code ran but failed to make things smaller,
		 * free any pages it allocated and our page pointer array
		 */
		for (i = 0; i < nr_pages; i++) {
			WARN_ON(pages[i]->mapping);
			put_page(pages[i]);
		}
		kfree(pages);
		pages = NULL;
		total_compressed = 0;
		nr_pages = 0;

		/* flag the file so we don't compress in the future */
		if (!btrfs_test_opt(fs_info, FORCE_COMPRESS) &&
		    !(BTRFS_I(inode)->prop_compress)) {
			BTRFS_I(inode)->flags |= BTRFS_INODE_NOCOMPRESS;
		}
	}
cleanup_and_bail_uncompressed:
	/*
	 * No compression, but we still need to write the pages in the file
	 * we've been given so far.  redirty the locked page if it corresponds
	 * to our extent and set things up for the async work queue to run
	 * cow_file_range to do the normal delalloc dance.
	 */
	if (page_offset(locked_page) >= start &&
	    page_offset(locked_page) <= end)
		__set_page_dirty_nobuffers(locked_page);
		/* unlocked later on in the async handlers */

	if (redirty)
		extent_range_redirty_for_io(inode, start, end);
	add_async_extent(async_cow, start, end - start + 1, 0, NULL, 0,
			 BTRFS_COMPRESS_NONE);
	*num_added += 1;

	return;

free_pages_out:
	for (i = 0; i < nr_pages; i++) {
		WARN_ON(pages[i]->mapping);
		put_page(pages[i]);
	}
	kfree(pages);
}

static void free_async_extent_pages(struct async_extent *async_extent)
{
	int i;

	if (!async_extent->pages)
		return;

	for (i = 0; i < async_extent->nr_pages; i++) {
		WARN_ON(async_extent->pages[i]->mapping);
		put_page(async_extent->pages[i]);
	}
	kfree(async_extent->pages);
	async_extent->nr_pages = 0;
	async_extent->pages = NULL;
}

/*
 * phase two of compressed writeback.  This is the ordered portion
 * of the code, which only gets called in the order the work was
 * queued.  We walk all the async extents created by compress_file_range
 * and send them down to the disk.
 */
static noinline void submit_compressed_extents(struct inode *inode,
					      struct async_cow *async_cow)
{
	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
	struct async_extent *async_extent;
	u64 alloc_hint = 0;
	struct btrfs_key ins;
	struct extent_map *em;
	struct btrfs_root *root = BTRFS_I(inode)->root;
	struct extent_io_tree *io_tree;
	int ret = 0;

again:
	while (!list_empty(&async_cow->extents)) {
		async_extent = list_entry(async_cow->extents.next,
					  struct async_extent, list);
		list_del(&async_extent->list);

		io_tree = &BTRFS_I(inode)->io_tree;

retry:
		/* did the compression code fall back to uncompressed IO? */
		if (!async_extent->pages) {
			int page_started = 0;
			unsigned long nr_written = 0;

			lock_extent(io_tree, async_extent->start,
					 async_extent->start +
					 async_extent->ram_size - 1);

			/* allocate blocks */
			ret = cow_file_range(inode, async_cow->locked_page,
					     async_extent->start,
					     async_extent->start +
					     async_extent->ram_size - 1,
					     async_extent->start +
					     async_extent->ram_size - 1,
					     &page_started, &nr_written, 0,
					     NULL);

			/* JDM XXX */

			/*
			 * if page_started, cow_file_range inserted an
			 * inline extent and took care of all the unlocking
			 * and IO for us.  Otherwise, we need to submit
			 * all those pages down to the drive.
			 */
			if (!page_started && !ret)
				extent_write_locked_range(inode,
						  async_extent->start,
						  async_extent->start +
						  async_extent->ram_size - 1,
						  WB_SYNC_ALL);
			else if (ret)
				unlock_page(async_cow->locked_page);
			kfree(async_extent);
			cond_resched();
			continue;
		}

		lock_extent(io_tree, async_extent->start,
			    async_extent->start + async_extent->ram_size - 1);

		ret = btrfs_reserve_extent(root, async_extent->ram_size,
					   async_extent->compressed_size,
					   async_extent->compressed_size,
					   0, alloc_hint, &ins, 1, 1);
		if (ret) {
			free_async_extent_pages(async_extent);

			if (ret == -ENOSPC) {
				unlock_extent(io_tree, async_extent->start,
					      async_extent->start +
					      async_extent->ram_size - 1);

				/*
				 * we need to redirty the pages if we decide to
				 * fallback to uncompressed IO, otherwise we
				 * will not submit these pages down to lower
				 * layers.
				 */
				extent_range_redirty_for_io(inode,
						async_extent->start,
						async_extent->start +
						async_extent->ram_size - 1);

				goto retry;
			}
			goto out_free;
		}
		/*
		 * here we're doing allocation and writeback of the
		 * compressed pages
		 */
		em = create_io_em(inode, async_extent->start,
				  async_extent->ram_size, /* len */
				  async_extent->start, /* orig_start */
				  ins.objectid, /* block_start */
				  ins.offset, /* block_len */
				  ins.offset, /* orig_block_len */
				  async_extent->ram_size, /* ram_bytes */
				  async_extent->compress_type,
				  BTRFS_ORDERED_COMPRESSED);
		if (IS_ERR(em))
			/* ret value is not necessary due to void function */
			goto out_free_reserve;
		free_extent_map(em);

		ret = btrfs_add_ordered_extent_compress(inode,
						async_extent->start,
						ins.objectid,
						async_extent->ram_size,
						ins.offset,
						BTRFS_ORDERED_COMPRESSED,
						async_extent->compress_type);
		if (ret) {
			btrfs_drop_extent_cache(BTRFS_I(inode),
						async_extent->start,
						async_extent->start +
						async_extent->ram_size - 1, 0);
			goto out_free_reserve;
		}
		btrfs_dec_block_group_reservations(fs_info, ins.objectid);

		/*
		 * clear dirty, set writeback and unlock the pages.
		 */
		extent_clear_unlock_delalloc(inode, async_extent->start,
				async_extent->start +
				async_extent->ram_size - 1,
				async_extent->start +
				async_extent->ram_size - 1,
				NULL, EXTENT_LOCKED | EXTENT_DELALLOC,
				PAGE_UNLOCK | PAGE_CLEAR_DIRTY |
				PAGE_SET_WRITEBACK);
		if (btrfs_submit_compressed_write(inode,
				    async_extent->start,
				    async_extent->ram_size,
				    ins.objectid,
				    ins.offset, async_extent->pages,
				    async_extent->nr_pages,
				    async_cow->write_flags)) {
			struct extent_io_tree *tree = &BTRFS_I(inode)->io_tree;
			struct page *p = async_extent->pages[0];
			const u64 start = async_extent->start;
			const u64 end = start + async_extent->ram_size - 1;

			p->mapping = inode->i_mapping;
			tree->ops->writepage_end_io_hook(p, start, end,
							 NULL, 0);
			p->mapping = NULL;
			extent_clear_unlock_delalloc(inode, start, end, end,
						     NULL, 0,
						     PAGE_END_WRITEBACK |
						     PAGE_SET_ERROR);
			free_async_extent_pages(async_extent);
		}
		alloc_hint = ins.objectid + ins.offset;
		kfree(async_extent);
		cond_resched();
	}
	return;
out_free_reserve:
	btrfs_dec_block_group_reservations(fs_info, ins.objectid);
	btrfs_free_reserved_extent(fs_info, ins.objectid, ins.offset, 1);
out_free:
	extent_clear_unlock_delalloc(inode, async_extent->start,
				     async_extent->start +
				     async_extent->ram_size - 1,
				     async_extent->start +
				     async_extent->ram_size - 1,
				     NULL, EXTENT_LOCKED | EXTENT_DELALLOC |
				     EXTENT_DELALLOC_NEW |
				     EXTENT_DEFRAG | EXTENT_DO_ACCOUNTING,
				     PAGE_UNLOCK | PAGE_CLEAR_DIRTY |
				     PAGE_SET_WRITEBACK | PAGE_END_WRITEBACK |
				     PAGE_SET_ERROR);
	free_async_extent_pages(async_extent);
	kfree(async_extent);
	goto again;
}

static u64 get_extent_allocation_hint(struct inode *inode, u64 start,
				      u64 num_bytes)
{
	struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
	struct extent_map *em;
	u64 alloc_hint = 0;

	read_lock(&em_tree->lock);
	em = search_extent_mapping(em_tree, start, num_bytes);
	if (em) {
		/*
		 * if block start isn't an actual block number then find the
		 * first block in this inode and use that as a hint.  If that
		 * block is also bogus then just don't worry about it.
		 */
		if (em->block_start >= EXTENT_MAP_LAST_BYTE) {
			free_extent_map(em);
			em = search_extent_mapping(em_tree, 0, 0);
			if (em && em->block_start < EXTENT_MAP_LAST_BYTE)
				alloc_hint = em->block_start;
			if (em)
				free_extent_map(em);
		} else {
			alloc_hint = em->block_start;
			free_extent_map(em);
		}
	}
	read_unlock(&em_tree->lock);

	return alloc_hint;
}

/*
 * when extent_io.c finds a delayed allocation range in the file,
 * the call backs end up in this code.  The basic idea is to
 * allocate extents on disk for the range, and create ordered data structs
 * in ram to track those extents.
 *
 * locked_page is the page that writepage had locked already.  We use
 * it to make sure we don't do extra locks or unlocks.
 *
 * *page_started is set to one if we unlock locked_page and do everything
 * required to start IO on it.  It may be clean and already done with
 * IO when we return.
 */
static noinline int cow_file_range(struct inode *inode,
				   struct page *locked_page,
				   u64 start, u64 end, u64 delalloc_end,
				   int *page_started, unsigned long *nr_written,
				   int unlock, struct btrfs_dedupe_hash *hash)
{
	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
	struct btrfs_root *root = BTRFS_I(inode)->root;
	u64 alloc_hint = 0;
	u64 num_bytes;
	unsigned long ram_size;
	u64 cur_alloc_size = 0;
	u64 blocksize = fs_info->sectorsize;
	struct btrfs_key ins;
	struct extent_map *em;
	unsigned clear_bits;
	unsigned long page_ops;
	bool extent_reserved = false;
	int ret = 0;

	if (btrfs_is_free_space_inode(BTRFS_I(inode))) {
		WARN_ON_ONCE(1);
		ret = -EINVAL;
		goto out_unlock;
	}

	num_bytes = ALIGN(end - start + 1, blocksize);
	num_bytes = max(blocksize,  num_bytes);
	ASSERT(num_bytes <= btrfs_super_total_bytes(fs_info->super_copy));

	inode_should_defrag(BTRFS_I(inode), start, end, num_bytes, SZ_64K);

	if (start == 0) {
		/* lets try to make an inline extent */
		ret = cow_file_range_inline(inode, start, end, 0,
					    BTRFS_COMPRESS_NONE, NULL);
		if (ret == 0) {
			/*
			 * We use DO_ACCOUNTING here because we need the
			 * delalloc_release_metadata to be run _after_ we drop
			 * our outstanding extent for clearing delalloc for this
			 * range.
			 */
			extent_clear_unlock_delalloc(inode, start, end,
				     delalloc_end, NULL,
				     EXTENT_LOCKED | EXTENT_DELALLOC |
				     EXTENT_DELALLOC_NEW | EXTENT_DEFRAG |
				     EXTENT_DO_ACCOUNTING, PAGE_UNLOCK |
				     PAGE_CLEAR_DIRTY | PAGE_SET_WRITEBACK |
				     PAGE_END_WRITEBACK);
			*nr_written = *nr_written +
			     (end - start + PAGE_SIZE) / PAGE_SIZE;
			*page_started = 1;
			goto out;
		} else if (ret < 0) {
			goto out_unlock;
		}
	}

	alloc_hint = get_extent_allocation_hint(inode, start, num_bytes);
	btrfs_drop_extent_cache(BTRFS_I(inode), start,
			start + num_bytes - 1, 0);

	while (num_bytes > 0) {
		cur_alloc_size = num_bytes;
		ret = btrfs_reserve_extent(root, cur_alloc_size, cur_alloc_size,
					   fs_info->sectorsize, 0, alloc_hint,
					   &ins, 1, 1);
		if (ret < 0)
			goto out_unlock;
		cur_alloc_size = ins.offset;
		extent_reserved = true;

		ram_size = ins.offset;
		em = create_io_em(inode, start, ins.offset, /* len */
				  start, /* orig_start */
				  ins.objectid, /* block_start */
				  ins.offset, /* block_len */
				  ins.offset, /* orig_block_len */
				  ram_size, /* ram_bytes */
				  BTRFS_COMPRESS_NONE, /* compress_type */
				  BTRFS_ORDERED_REGULAR /* type */);
		if (IS_ERR(em)) {
			ret = PTR_ERR(em);
			goto out_reserve;
		}
		free_extent_map(em);

		ret = btrfs_add_ordered_extent(inode, start, ins.objectid,
					       ram_size, cur_alloc_size, 0);
		if (ret)
			goto out_drop_extent_cache;

		if (root->root_key.objectid ==
		    BTRFS_DATA_RELOC_TREE_OBJECTID) {
			ret = btrfs_reloc_clone_csums(inode, start,
						      cur_alloc_size);
			/*
			 * Only drop cache here, and process as normal.
			 *
			 * We must not allow extent_clear_unlock_delalloc()
			 * at out_unlock label to free meta of this ordered
			 * extent, as its meta should be freed by
			 * btrfs_finish_ordered_io().
			 *
			 * So we must continue until @start is increased to
			 * skip current ordered extent.
			 */
			if (ret)
				btrfs_drop_extent_cache(BTRFS_I(inode), start,
						start + ram_size - 1, 0);
		}

		btrfs_dec_block_group_reservations(fs_info, ins.objectid);

		/* we're not doing compressed IO, don't unlock the first
		 * page (which the caller expects to stay locked), don't
		 * clear any dirty bits and don't set any writeback bits
		 *
		 * Do set the Private2 bit so we know this page was properly
		 * setup for writepage
		 */
		page_ops = unlock ? PAGE_UNLOCK : 0;
		page_ops |= PAGE_SET_PRIVATE2;

		extent_clear_unlock_delalloc(inode, start,
					     start + ram_size - 1,
					     delalloc_end, locked_page,
					     EXTENT_LOCKED | EXTENT_DELALLOC,
					     page_ops);
		if (num_bytes < cur_alloc_size)
			num_bytes = 0;
		else
			num_bytes -= cur_alloc_size;
		alloc_hint = ins.objectid + ins.offset;
		start += cur_alloc_size;
		extent_reserved = false;

		/*
		 * btrfs_reloc_clone_csums() error, since start is increased
		 * extent_clear_unlock_delalloc() at out_unlock label won't
		 * free metadata of current ordered extent, we're OK to exit.
		 */
		if (ret)
			goto out_unlock;
	}
out:
	return ret;

out_drop_extent_cache:
	btrfs_drop_extent_cache(BTRFS_I(inode), start, start + ram_size - 1, 0);
out_reserve:
	btrfs_dec_block_group_reservations(fs_info, ins.objectid);
	btrfs_free_reserved_extent(fs_info, ins.objectid, ins.offset, 1);
out_unlock:
	clear_bits = EXTENT_LOCKED | EXTENT_DELALLOC | EXTENT_DELALLOC_NEW |
		EXTENT_DEFRAG | EXTENT_CLEAR_META_RESV;
	page_ops = PAGE_UNLOCK | PAGE_CLEAR_DIRTY | PAGE_SET_WRITEBACK |
		PAGE_END_WRITEBACK;
	/*
	 * If we reserved an extent for our delalloc range (or a subrange) and
	 * failed to create the respective ordered extent, then it means that
	 * when we reserved the extent we decremented the extent's size from
	 * the data space_info's bytes_may_use counter and incremented the
	 * space_info's bytes_reserved counter by the same amount. We must make
	 * sure extent_clear_unlock_delalloc() does not try to decrement again
	 * the data space_info's bytes_may_use counter, therefore we do not pass
	 * it the flag EXTENT_CLEAR_DATA_RESV.
	 */
	if (extent_reserved) {
		extent_clear_unlock_delalloc(inode, start,
					     start + cur_alloc_size,
					     start + cur_alloc_size,
					     locked_page,
					     clear_bits,
					     page_ops);
		start += cur_alloc_size;
		if (start >= end)
			goto out;
	}
	extent_clear_unlock_delalloc(inode, start, end, delalloc_end,
				     locked_page,
				     clear_bits | EXTENT_CLEAR_DATA_RESV,
				     page_ops);
	goto out;
}

/*
 * work queue call back to started compression on a file and pages
 */
static noinline void async_cow_start(struct btrfs_work *work)
{
	struct async_cow *async_cow;
	int num_added = 0;
	async_cow = container_of(work, struct async_cow, work);

	compress_file_range(async_cow->inode, async_cow->locked_page,
			    async_cow->start, async_cow->end, async_cow,
			    &num_added);
	if (num_added == 0) {
		btrfs_add_delayed_iput(async_cow->inode);
		async_cow->inode = NULL;
	}
}

/*
 * work queue call back to submit previously compressed pages
 */
static noinline void async_cow_submit(struct btrfs_work *work)
{
	struct btrfs_fs_info *fs_info;
	struct async_cow *async_cow;
	struct btrfs_root *root;
	unsigned long nr_pages;

	async_cow = container_of(work, struct async_cow, work);

	root = async_cow->root;
	fs_info = root->fs_info;
	nr_pages = (async_cow->end - async_cow->start + PAGE_SIZE) >>
		PAGE_SHIFT;

	/* atomic_sub_return implies a barrier */
	if (atomic_sub_return(nr_pages, &fs_info->async_delalloc_pages) <
	    5 * SZ_1M)
		cond_wake_up_nomb(&fs_info->async_submit_wait);

	if (async_cow->inode)
		submit_compressed_extents(async_cow->inode, async_cow);
}

static noinline void async_cow_free(struct btrfs_work *work)
{
	struct async_cow *async_cow;
	async_cow = container_of(work, struct async_cow, work);
	if (async_cow->inode)
		btrfs_add_delayed_iput(async_cow->inode);
	kfree(async_cow);
}

static int cow_file_range_async(struct inode *inode, struct page *locked_page,
				u64 start, u64 end, int *page_started,
				unsigned long *nr_written,
				unsigned int write_flags)
{
	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
	struct async_cow *async_cow;
	struct btrfs_root *root = BTRFS_I(inode)->root;
	unsigned long nr_pages;
	u64 cur_end;

	clear_extent_bit(&BTRFS_I(inode)->io_tree, start, end, EXTENT_LOCKED,
			 1, 0, NULL);
	while (start < end) {
		async_cow = kmalloc(sizeof(*async_cow), GFP_NOFS);
		BUG_ON(!async_cow); /* -ENOMEM */
		async_cow->inode = igrab(inode);
		async_cow->root = root;
		async_cow->locked_page = locked_page;
		async_cow->start = start;
		async_cow->write_flags = write_flags;

		if (BTRFS_I(inode)->flags & BTRFS_INODE_NOCOMPRESS &&
		    !btrfs_test_opt(fs_info, FORCE_COMPRESS))
			cur_end = end;
		else
			cur_end = min(end, start + SZ_512K - 1);

		async_cow->end = cur_end;
		INIT_LIST_HEAD(&async_cow->extents);

		btrfs_init_work(&async_cow->work,
				btrfs_delalloc_helper,
				async_cow_start, async_cow_submit,
				async_cow_free);

		nr_pages = (cur_end - start + PAGE_SIZE) >>
			PAGE_SHIFT;
		atomic_add(nr_pages, &fs_info->async_delalloc_pages);

		btrfs_queue_work(fs_info->delalloc_workers, &async_cow->work);

		*nr_written += nr_pages;
		start = cur_end + 1;
	}
	*page_started = 1;
	return 0;
}

static noinline int csum_exist_in_range(struct btrfs_fs_info *fs_info,
					u64 bytenr, u64 num_bytes)
{
	int ret;
	struct btrfs_ordered_sum *sums;
	LIST_HEAD(list);

	ret = btrfs_lookup_csums_range(fs_info->csum_root, bytenr,
				       bytenr + num_bytes - 1, &list, 0);
	if (ret == 0 && list_empty(&list))
		return 0;

	while (!list_empty(&list)) {
		sums = list_entry(list.next, struct btrfs_ordered_sum, list);
		list_del(&sums->list);
		kfree(sums);
	}
	if (ret < 0)
		return ret;
	return 1;
}

/*
 * when nowcow writeback call back.  This checks for snapshots or COW copies
 * of the extents that exist in the file, and COWs the file as required.
 *
 * If no cow copies or snapshots exist, we write directly to the existing
 * blocks on disk
 */
static noinline int run_delalloc_nocow(struct inode *inode,
				       struct page *locked_page,
			      u64 start, u64 end, int *page_started, int force,
			      unsigned long *nr_written)
{
	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
	struct btrfs_root *root = BTRFS_I(inode)->root;
	struct extent_buffer *leaf;
	struct btrfs_path *path;
	struct btrfs_file_extent_item *fi;
	struct btrfs_key found_key;
	struct extent_map *em;
	u64 cow_start;
	u64 cur_offset;
	u64 extent_end;
	u64 extent_offset;
	u64 disk_bytenr;
	u64 num_bytes;
	u64 disk_num_bytes;
	u64 ram_bytes;
	int extent_type;
	int ret, err;
	int type;
	int nocow;
	int check_prev = 1;
	bool nolock;
	u64 ino = btrfs_ino(BTRFS_I(inode));

	path = btrfs_alloc_path();
	if (!path) {
		extent_clear_unlock_delalloc(inode, start, end, end,
					     locked_page,
					     EXTENT_LOCKED | EXTENT_DELALLOC |
					     EXTENT_DO_ACCOUNTING |
					     EXTENT_DEFRAG, PAGE_UNLOCK |
					     PAGE_CLEAR_DIRTY |
					     PAGE_SET_WRITEBACK |
					     PAGE_END_WRITEBACK);
		return -ENOMEM;
	}

	nolock = btrfs_is_free_space_inode(BTRFS_I(inode));

	cow_start = (u64)-1;
	cur_offset = start;
	while (1) {
		ret = btrfs_lookup_file_extent(NULL, root, path, ino,
					       cur_offset, 0);
		if (ret < 0)
			goto error;
		if (ret > 0 && path->slots[0] > 0 && check_prev) {
			leaf = path->nodes[0];
			btrfs_item_key_to_cpu(leaf, &found_key,
					      path->slots[0] - 1);
			if (found_key.objectid == ino &&
			    found_key.type == BTRFS_EXTENT_DATA_KEY)
				path->slots[0]--;
		}
		check_prev = 0;
next_slot:
		leaf = path->nodes[0];
		if (path->slots[0] >= btrfs_header_nritems(leaf)) {
			ret = btrfs_next_leaf(root, path);
			if (ret < 0) {
				if (cow_start != (u64)-1)
					cur_offset = cow_start;
				goto error;
			}
			if (ret > 0)
				break;
			leaf = path->nodes[0];
		}

		nocow = 0;
		disk_bytenr = 0;
		num_bytes = 0;
		btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);

		if (found_key.objectid > ino)
			break;
		if (WARN_ON_ONCE(found_key.objectid < ino) ||
		    found_key.type < BTRFS_EXTENT_DATA_KEY) {
			path->slots[0]++;
			goto next_slot;
		}
		if (found_key.type > BTRFS_EXTENT_DATA_KEY ||
		    found_key.offset > end)
			break;

		if (found_key.offset > cur_offset) {
			extent_end = found_key.offset;
			extent_type = 0;
			goto out_check;
		}

		fi = btrfs_item_ptr(leaf, path->slots[0],
				    struct btrfs_file_extent_item);
		extent_type = btrfs_file_extent_type(leaf, fi);

		ram_bytes = btrfs_file_extent_ram_bytes(leaf, fi);
		if (extent_type == BTRFS_FILE_EXTENT_REG ||
		    extent_type == BTRFS_FILE_EXTENT_PREALLOC) {
			disk_bytenr = btrfs_file_extent_disk_bytenr(leaf, fi);
			extent_offset = btrfs_file_extent_offset(leaf, fi);
			extent_end = found_key.offset +
				btrfs_file_extent_num_bytes(leaf, fi);
			disk_num_bytes =
				btrfs_file_extent_disk_num_bytes(leaf, fi);
			if (extent_end <= start) {
				path->slots[0]++;
				goto next_slot;
			}
			if (disk_bytenr == 0)
				goto out_check;
			if (btrfs_file_extent_compression(leaf, fi) ||
			    btrfs_file_extent_encryption(leaf, fi) ||
			    btrfs_file_extent_other_encoding(leaf, fi))
				goto out_check;
			/*
			 * Do the same check as in btrfs_cross_ref_exist but
			 * without the unnecessary search.
			 */
			if (btrfs_file_extent_generation(leaf, fi) <=
			    btrfs_root_last_snapshot(&root->root_item))
				goto out_check;
			if (extent_type == BTRFS_FILE_EXTENT_REG && !force)
				goto out_check;
			if (btrfs_extent_readonly(fs_info, disk_bytenr))
				goto out_check;
			ret = btrfs_cross_ref_exist(root, ino,
						    found_key.offset -
						    extent_offset, disk_bytenr);
			if (ret) {
				/*
				 * ret could be -EIO if the above fails to read
				 * metadata.
				 */
				if (ret < 0) {
					if (cow_start != (u64)-1)
						cur_offset = cow_start;
					goto error;
				}

				WARN_ON_ONCE(nolock);
				goto out_check;
			}
			disk_bytenr += extent_offset;
			disk_bytenr += cur_offset - found_key.offset;
			num_bytes = min(end + 1, extent_end) - cur_offset;
			/*
			 * if there are pending snapshots for this root,
			 * we fall into common COW way.
			 */
			if (!nolock) {
				err = btrfs_start_write_no_snapshotting(root);
				if (!err)
					goto out_check;
			}
			/*
			 * force cow if csum exists in the range.
			 * this ensure that csum for a given extent are
			 * either valid or do not exist.
			 */
			ret = csum_exist_in_range(fs_info, disk_bytenr,
						  num_bytes);
			if (ret) {
				if (!nolock)
					btrfs_end_write_no_snapshotting(root);

				/*
				 * ret could be -EIO if the above fails to read
				 * metadata.
				 */
				if (ret < 0) {
					if (cow_start != (u64)-1)
						cur_offset = cow_start;
					goto error;
				}
				WARN_ON_ONCE(nolock);
				goto out_check;
			}
			if (!btrfs_inc_nocow_writers(fs_info, disk_bytenr)) {
				if (!nolock)
					btrfs_end_write_no_snapshotting(root);
				goto out_check;
			}
			nocow = 1;
		} else if (extent_type == BTRFS_FILE_EXTENT_INLINE) {
			extent_end = found_key.offset +
				btrfs_file_extent_inline_len(leaf,
						     path->slots[0], fi);
			extent_end = ALIGN(extent_end,
					   fs_info->sectorsize);
		} else {
			BUG_ON(1);
		}
out_check:
		if (extent_end <= start) {
			path->slots[0]++;
			if (!nolock && nocow)
				btrfs_end_write_no_snapshotting(root);
			if (nocow)
				btrfs_dec_nocow_writers(fs_info, disk_bytenr);
			goto next_slot;
		}
		if (!nocow) {
			if (cow_start == (u64)-1)
				cow_start = cur_offset;
			cur_offset = extent_end;
			if (cur_offset > end)
				break;
			path->slots[0]++;
			goto next_slot;
		}

		btrfs_release_path(path);
		if (cow_start != (u64)-1) {
			ret = cow_file_range(inode, locked_page,
					     cow_start, found_key.offset - 1,
					     end, page_started, nr_written, 1,
					     NULL);
			if (ret) {
				if (!nolock && nocow)
					btrfs_end_write_no_snapshotting(root);
				if (nocow)
					btrfs_dec_nocow_writers(fs_info,
								disk_bytenr);
				goto error;
			}
			cow_start = (u64)-1;
		}

		if (extent_type == BTRFS_FILE_EXTENT_PREALLOC) {
			u64 orig_start = found_key.offset - extent_offset;

			em = create_io_em(inode, cur_offset, num_bytes,
					  orig_start,
					  disk_bytenr, /* block_start */
					  num_bytes, /* block_len */
					  disk_num_bytes, /* orig_block_len */
					  ram_bytes, BTRFS_COMPRESS_NONE,
					  BTRFS_ORDERED_PREALLOC);
			if (IS_ERR(em)) {
				if (!nolock && nocow)
					btrfs_end_write_no_snapshotting(root);
				if (nocow)
					btrfs_dec_nocow_writers(fs_info,
								disk_bytenr);
				ret = PTR_ERR(em);
				goto error;
			}
			free_extent_map(em);
		}

		if (extent_type == BTRFS_FILE_EXTENT_PREALLOC) {
			type = BTRFS_ORDERED_PREALLOC;
		} else {
			type = BTRFS_ORDERED_NOCOW;
		}

		ret = btrfs_add_ordered_extent(inode, cur_offset, disk_bytenr,
					       num_bytes, num_bytes, type);
		if (nocow)
			btrfs_dec_nocow_writers(fs_info, disk_bytenr);
		BUG_ON(ret); /* -ENOMEM */

		if (root->root_key.objectid ==
		    BTRFS_DATA_RELOC_TREE_OBJECTID)
			/*
			 * Error handled later, as we must prevent
			 * extent_clear_unlock_delalloc() in error handler
			 * from freeing metadata of created ordered extent.
			 */
			ret = btrfs_reloc_clone_csums(inode, cur_offset,
						      num_bytes);

		extent_clear_unlock_delalloc(inode, cur_offset,
					     cur_offset + num_bytes - 1, end,
					     locked_page, EXTENT_LOCKED |
					     EXTENT_DELALLOC |
					     EXTENT_CLEAR_DATA_RESV,
					     PAGE_UNLOCK | PAGE_SET_PRIVATE2);

		if (!nolock && nocow)
			btrfs_end_write_no_snapshotting(root);
		cur_offset = extent_end;

		/*
		 * btrfs_reloc_clone_csums() error, now we're OK to call error
		 * handler, as metadata for created ordered extent will only
		 * be freed by btrfs_finish_ordered_io().
		 */
		if (ret)
			goto error;
		if (cur_offset > end)
			break;
	}
	btrfs_release_path(path);

	if (cur_offset <= end && cow_start == (u64)-1) {
		cow_start = cur_offset;
		cur_offset = end;
	}

	if (cow_start != (u64)-1) {
		ret = cow_file_range(inode, locked_page, cow_start, end, end,
				     page_started, nr_written, 1, NULL);
		if (ret)
			goto error;
	}

error:
	if (ret && cur_offset < end)
		extent_clear_unlock_delalloc(inode, cur_offset, end, end,
					     locked_page, EXTENT_LOCKED |
					     EXTENT_DELALLOC | EXTENT_DEFRAG |
					     EXTENT_DO_ACCOUNTING, PAGE_UNLOCK |
					     PAGE_CLEAR_DIRTY |
					     PAGE_SET_WRITEBACK |
					     PAGE_END_WRITEBACK);
	btrfs_free_path(path);
	return ret;
}

static inline int need_force_cow(struct inode *inode, u64 start, u64 end)
{

	if (!(BTRFS_I(inode)->flags & BTRFS_INODE_NODATACOW) &&
	    !(BTRFS_I(inode)->flags & BTRFS_INODE_PREALLOC))
		return 0;

	/*
	 * @defrag_bytes is a hint value, no spinlock held here,
	 * if is not zero, it means the file is defragging.
	 * Force cow if given extent needs to be defragged.
	 */
	if (BTRFS_I(inode)->defrag_bytes &&
	    test_range_bit(&BTRFS_I(inode)->io_tree, start, end,
			   EXTENT_DEFRAG, 0, NULL))
		return 1;

	return 0;
}

/*
 * extent_io.c call back to do delayed allocation processing
 */
static int run_delalloc_range(void *private_data, struct page *locked_page,
			      u64 start, u64 end, int *page_started,
			      unsigned long *nr_written,
			      struct writeback_control *wbc)
{
	struct inode *inode = private_data;
	int ret;
	int force_cow = need_force_cow(inode, start, end);
	unsigned int write_flags = wbc_to_write_flags(wbc);

	if (BTRFS_I(inode)->flags & BTRFS_INODE_NODATACOW && !force_cow) {
		ret = run_delalloc_nocow(inode, locked_page, start, end,
					 page_started, 1, nr_written);
	} else if (BTRFS_I(inode)->flags & BTRFS_INODE_PREALLOC && !force_cow) {
		ret = run_delalloc_nocow(inode, locked_page, start, end,
					 page_started, 0, nr_written);
	} else if (!inode_need_compress(inode, start, end)) {
		ret = cow_file_range(inode, locked_page, start, end, end,
				      page_started, nr_written, 1, NULL);
	} else {
		set_bit(BTRFS_INODE_HAS_ASYNC_EXTENT,
			&BTRFS_I(inode)->runtime_flags);
		ret = cow_file_range_async(inode, locked_page, start, end,
					   page_started, nr_written,
					   write_flags);
	}
	if (ret)
		btrfs_cleanup_ordered_extents(inode, start, end - start + 1);
	return ret;
}

static void btrfs_split_extent_hook(void *private_data,
				    struct extent_state *orig, u64 split)
{
	struct inode *inode = private_data;
	u64 size;

	/* not delalloc, ignore it */
	if (!(orig->state & EXTENT_DELALLOC))
		return;

	size = orig->end - orig->start + 1;
	if (size > BTRFS_MAX_EXTENT_SIZE) {
		u32 num_extents;
		u64 new_size;

		/*
		 * See the explanation in btrfs_merge_extent_hook, the same
		 * applies here, just in reverse.
		 */
		new_size = orig->end - split + 1;
		num_extents = count_max_extents(new_size);
		new_size = split - orig->start;
		num_extents += count_max_extents(new_size);
		if (count_max_extents(size) >= num_extents)
			return;
	}

	spin_lock(&BTRFS_I(inode)->lock);
	btrfs_mod_outstanding_extents(BTRFS_I(inode), 1);
	spin_unlock(&BTRFS_I(inode)->lock);
}

/*
 * extent_io.c merge_extent_hook, used to track merged delayed allocation
 * extents so we can keep track of new extents that are just merged onto old
 * extents, such as when we are doing sequential writes, so we can properly
 * account for the metadata space we'll need.
 */
static void btrfs_merge_extent_hook(void *private_data,
				    struct extent_state *new,
				    struct extent_state *other)
{
	struct inode *inode = private_data;
	u64 new_size, old_size;
	u32 num_extents;

	/* not delalloc, ignore it */
	if (!(other->state & EXTENT_DELALLOC))
		return;

	if (new->start > other->start)
		new_size = new->end - other->start + 1;
	else
		new_size = other->end - new->start + 1;

	/* we're not bigger than the max, unreserve the space and go */
	if (new_size <= BTRFS_MAX_EXTENT_SIZE) {
		spin_lock(&BTRFS_I(inode)->lock);
		btrfs_mod_outstanding_extents(BTRFS_I(inode), -1);
		spin_unlock(&BTRFS_I(inode)->lock);
		return;
	}

	/*
	 * We have to add up either side to figure out how many extents were
	 * accounted for before we merged into one big extent.  If the number of
	 * extents we accounted for is <= the amount we need for the new range
	 * then we can return, otherwise drop.  Think of it like this
	 *
	 * [ 4k][MAX_SIZE]
	 *
	 * So we've grown the extent by a MAX_SIZE extent, this would mean we
	 * need 2 outstanding extents, on one side we have 1 and the other side
	 * we have 1 so they are == and we can return.  But in this case
	 *
	 * [MAX_SIZE+4k][MAX_SIZE+4k]
	 *
	 * Each range on their own accounts for 2 extents, but merged together
	 * they are only 3 extents worth of accounting, so we need to drop in
	 * this case.
	 */
	old_size = other->end - other->start + 1;
	num_extents = count_max_extents(old_size);
	old_size = new->end - new->start + 1;
	num_extents += count_max_extents(old_size);
	if (count_max_extents(new_size) >= num_extents)
		return;

	spin_lock(&BTRFS_I(inode)->lock);
	btrfs_mod_outstanding_extents(BTRFS_I(inode), -1);
	spin_unlock(&BTRFS_I(inode)->lock);
}

static void btrfs_add_delalloc_inodes(struct btrfs_root *root,
				      struct inode *inode)
{
	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);

	spin_lock(&root->delalloc_lock);
	if (list_empty(&BTRFS_I(inode)->delalloc_inodes)) {
		list_add_tail(&BTRFS_I(inode)->delalloc_inodes,
			      &root->delalloc_inodes);
		set_bit(BTRFS_INODE_IN_DELALLOC_LIST,
			&BTRFS_I(inode)->runtime_flags);
		root->nr_delalloc_inodes++;
		if (root->nr_delalloc_inodes == 1) {
			spin_lock(&fs_info->delalloc_root_lock);
			BUG_ON(!list_empty(&root->delalloc_root));
			list_add_tail(&root->delalloc_root,
				      &fs_info->delalloc_roots);
			spin_unlock(&fs_info->delalloc_root_lock);
		}
	}
	spin_unlock(&root->delalloc_lock);
}


void __btrfs_del_delalloc_inode(struct btrfs_root *root,
				struct btrfs_inode *inode)
{
	struct btrfs_fs_info *fs_info = btrfs_sb(inode->vfs_inode.i_sb);

	if (!list_empty(&inode->delalloc_inodes)) {
		list_del_init(&inode->delalloc_inodes);
		clear_bit(BTRFS_INODE_IN_DELALLOC_LIST,
			  &inode->runtime_flags);
		root->nr_delalloc_inodes--;
		if (!root->nr_delalloc_inodes) {
			ASSERT(list_empty(&root->delalloc_inodes));
			spin_lock(&fs_info->delalloc_root_lock);
			BUG_ON(list_empty(&root->delalloc_root));
			list_del_init(&root->delalloc_root);
			spin_unlock(&fs_info->delalloc_root_lock);
		}
	}
}

static void btrfs_del_delalloc_inode(struct btrfs_root *root,
				     struct btrfs_inode *inode)
{
	spin_lock(&root->delalloc_lock);
	__btrfs_del_delalloc_inode(root, inode);
	spin_unlock(&root->delalloc_lock);
}

/*
 * extent_io.c set_bit_hook, used to track delayed allocation
 * bytes in this file, and to maintain the list of inodes that
 * have pending delalloc work to be done.
 */
static void btrfs_set_bit_hook(void *private_data,
			       struct extent_state *state, unsigned *bits)
{
	struct inode *inode = private_data;

	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);

	if ((*bits & EXTENT_DEFRAG) && !(*bits & EXTENT_DELALLOC))
		WARN_ON(1);
	/*
	 * set_bit and clear bit hooks normally require _irqsave/restore
	 * but in this case, we are only testing for the DELALLOC
	 * bit, which is only set or cleared with irqs on
	 */
	if (!(state->state & EXTENT_DELALLOC) && (*bits & EXTENT_DELALLOC)) {
		struct btrfs_root *root = BTRFS_I(inode)->root;
		u64 len = state->end + 1 - state->start;
		u32 num_extents = count_max_extents(len);
		bool do_list = !btrfs_is_free_space_inode(BTRFS_I(inode));

		spin_lock(&BTRFS_I(inode)->lock);
		btrfs_mod_outstanding_extents(BTRFS_I(inode), num_extents);
		spin_unlock(&BTRFS_I(inode)->lock);

		/* For sanity tests */
		if (btrfs_is_testing(fs_info))
			return;

		percpu_counter_add_batch(&fs_info->delalloc_bytes, len,
					 fs_info->delalloc_batch);
		spin_lock(&BTRFS_I(inode)->lock);
		BTRFS_I(inode)->delalloc_bytes += len;
		if (*bits & EXTENT_DEFRAG)
			BTRFS_I(inode)->defrag_bytes += len;
		if (do_list && !test_bit(BTRFS_INODE_IN_DELALLOC_LIST,
					 &BTRFS_I(inode)->runtime_flags))
			btrfs_add_delalloc_inodes(root, inode);
		spin_unlock(&BTRFS_I(inode)->lock);
	}

	if (!(state->state & EXTENT_DELALLOC_NEW) &&
	    (*bits & EXTENT_DELALLOC_NEW)) {
		spin_lock(&BTRFS_I(inode)->lock);
		BTRFS_I(inode)->new_delalloc_bytes += state->end + 1 -
			state->start;
		spin_unlock(&BTRFS_I(inode)->lock);
	}
}

/*
 * extent_io.c clear_bit_hook, see set_bit_hook for why
 */
static void btrfs_clear_bit_hook(void *private_data,
				 struct extent_state *state,
				 unsigned *bits)
{
	struct btrfs_inode *inode = BTRFS_I((struct inode *)private_data);
	struct btrfs_fs_info *fs_info = btrfs_sb(inode->vfs_inode.i_sb);
	u64 len = state->end + 1 - state->start;
	u32 num_extents = count_max_extents(len);

	if ((state->state & EXTENT_DEFRAG) && (*bits & EXTENT_DEFRAG)) {
		spin_lock(&inode->lock);
		inode->defrag_bytes -= len;
		spin_unlock(&inode->lock);
	}

	/*
	 * set_bit and clear bit hooks normally require _irqsave/restore
	 * but in this case, we are only testing for the DELALLOC
	 * bit, which is only set or cleared with irqs on
	 */
	if ((state->state & EXTENT_DELALLOC) && (*bits & EXTENT_DELALLOC)) {
		struct btrfs_root *root = inode->root;
		bool do_list = !btrfs_is_free_space_inode(inode);

		spin_lock(&inode->lock);
		btrfs_mod_outstanding_extents(inode, -num_extents);
		spin_unlock(&inode->lock);

		/*
		 * We don't reserve metadata space for space cache inodes so we
		 * don't need to call dellalloc_release_metadata if there is an
		 * error.
		 */
		if (*bits & EXTENT_CLEAR_META_RESV &&
		    root != fs_info->tree_root)
			btrfs_delalloc_release_metadata(inode, len, false);

		/* For sanity tests. */
		if (btrfs_is_testing(fs_info))
			return;

		if (root->root_key.objectid != BTRFS_DATA_RELOC_TREE_OBJECTID &&
		    do_list && !(state->state & EXTENT_NORESERVE) &&
		    (*bits & EXTENT_CLEAR_DATA_RESV))
			btrfs_free_reserved_data_space_noquota(
					&inode->vfs_inode,
					state->start, len);

		percpu_counter_add_batch(&fs_info->delalloc_bytes, -len,
					 fs_info->delalloc_batch);
		spin_lock(&inode->lock);
		inode->delalloc_bytes -= len;
		if (do_list && inode->delalloc_bytes == 0 &&
		    test_bit(BTRFS_INODE_IN_DELALLOC_LIST,
					&inode->runtime_flags))
			btrfs_del_delalloc_inode(root, inode);
		spin_unlock(&inode->lock);
	}

	if ((state->state & EXTENT_DELALLOC_NEW) &&
	    (*bits & EXTENT_DELALLOC_NEW)) {
		spin_lock(&inode->lock);
		ASSERT(inode->new_delalloc_bytes >= len);
		inode->new_delalloc_bytes -= len;
		spin_unlock(&inode->lock);
	}
}

/*
 * extent_io.c merge_bio_hook, this must check the chunk tree to make sure
 * we don't create bios that span stripes or chunks
 *
 * return 1 if page cannot be merged to bio
 * return 0 if page can be merged to bio
 * return error otherwise
 */
int btrfs_merge_bio_hook(struct page *page, unsigned long offset,
			 size_t size, struct bio *bio,
			 unsigned long bio_flags)
{
	struct inode *inode = page->mapping->host;
	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
	u64 logical = (u64)bio->bi_iter.bi_sector << 9;
	u64 length = 0;
	u64 map_length;
	int ret;

	if (bio_flags & EXTENT_BIO_COMPRESSED)
		return 0;

	length = bio->bi_iter.bi_size;
	map_length = length;
	ret = btrfs_map_block(fs_info, btrfs_op(bio), logical, &map_length,
			      NULL, 0);
	if (ret < 0)
		return ret;
	if (map_length < length + size)
		return 1;
	return 0;
}

/*
 * in order to insert checksums into the metadata in large chunks,
 * we wait until bio submission time.   All the pages in the bio are
 * checksummed and sums are attached onto the ordered extent record.
 *
 * At IO completion time the cums attached on the ordered extent record
 * are inserted into the btree
 */
static blk_status_t btrfs_submit_bio_start(void *private_data, struct bio *bio,
				    u64 bio_offset)
{
	struct inode *inode = private_data;
	blk_status_t ret = 0;

	ret = btrfs_csum_one_bio(inode, bio, 0, 0);
	BUG_ON(ret); /* -ENOMEM */
	return 0;
}

/*
 * in order to insert checksums into the metadata in large chunks,
 * we wait until bio submission time.   All the pages in the bio are
 * checksummed and sums are attached onto the ordered extent record.
 *
 * At IO completion time the cums attached on the ordered extent record
 * are inserted into the btree
 */
static blk_status_t btrfs_submit_bio_done(void *private_data, struct bio *bio,
			  int mirror_num)
{
	struct inode *inode = private_data;
	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
	blk_status_t ret;

	ret = btrfs_map_bio(fs_info, bio, mirror_num, 1);
	if (ret) {
		bio->bi_status = ret;
		bio_endio(bio);
	}
	return ret;
}

/*
 * extent_io.c submission hook. This does the right thing for csum calculation
 * on write, or reading the csums from the tree before a read.
 *
 * Rules about async/sync submit,
 * a) read:				sync submit
 *
 * b) write without checksum:		sync submit
 *
 * c) write with checksum:
 *    c-1) if bio is issued by fsync:	sync submit
 *         (sync_writers != 0)
 *
 *    c-2) if root is reloc root:	sync submit
 *         (only in case of buffered IO)
 *
 *    c-3) otherwise:			async submit
 */
static blk_status_t btrfs_submit_bio_hook(void *private_data, struct bio *bio,
				 int mirror_num, unsigned long bio_flags,
				 u64 bio_offset)
{
	struct inode *inode = private_data;
	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
	struct btrfs_root *root = BTRFS_I(inode)->root;
	enum btrfs_wq_endio_type metadata = BTRFS_WQ_ENDIO_DATA;
	blk_status_t ret = 0;
	int skip_sum;
	int async = !atomic_read(&BTRFS_I(inode)->sync_writers);

	skip_sum = BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM;

	if (btrfs_is_free_space_inode(BTRFS_I(inode)))
		metadata = BTRFS_WQ_ENDIO_FREE_SPACE;

	if (bio_op(bio) != REQ_OP_WRITE) {
		ret = btrfs_bio_wq_end_io(fs_info, bio, metadata);
		if (ret)
			goto out;

		if (bio_flags & EXTENT_BIO_COMPRESSED) {
			ret = btrfs_submit_compressed_read(inode, bio,
							   mirror_num,
							   bio_flags);
			goto out;
		} else if (!skip_sum) {
			ret = btrfs_lookup_bio_sums(inode, bio, NULL);
			if (ret)
				goto out;
		}
		goto mapit;
	} else if (async && !skip_sum) {
		/* csum items have already been cloned */
		if (root->root_key.objectid == BTRFS_DATA_RELOC_TREE_OBJECTID)
			goto mapit;
		/* we're doing a write, do the async checksumming */
		ret = btrfs_wq_submit_bio(fs_info, bio, mirror_num, bio_flags,
					  bio_offset, inode,
					  btrfs_submit_bio_start,
					  btrfs_submit_bio_done);
		goto out;
	} else if (!skip_sum) {
		ret = btrfs_csum_one_bio(inode, bio, 0, 0);
		if (ret)
			goto out;
	}

mapit:
	ret = btrfs_map_bio(fs_info, bio, mirror_num, 0);

out:
	if (ret) {
		bio->bi_status = ret;
		bio_endio(bio);
	}
	return ret;
}

/*
 * given a list of ordered sums record them in the inode.  This happens
 * at IO completion time based on sums calculated at bio submission time.
 */
static noinline int add_pending_csums(struct btrfs_trans_handle *trans,
			     struct inode *inode, struct list_head *list)
{
	struct btrfs_ordered_sum *sum;
	int ret;

	list_for_each_entry(sum, list, list) {
		trans->adding_csums = true;
		ret = btrfs_csum_file_blocks(trans,
		       BTRFS_I(inode)->root->fs_info->csum_root, sum);
		trans->adding_csums = false;
		if (ret)
			return ret;
	}
	return 0;
}

int btrfs_set_extent_delalloc(struct inode *inode, u64 start, u64 end,
			      unsigned int extra_bits,
			      struct extent_state **cached_state, int dedupe)
{
	WARN_ON((end & (PAGE_SIZE - 1)) == 0);
	return set_extent_delalloc(&BTRFS_I(inode)->io_tree, start, end,
				   extra_bits, cached_state);
}

/* see btrfs_writepage_start_hook for details on why this is required */
struct btrfs_writepage_fixup {
	struct page *page;
	struct btrfs_work work;
};

static void btrfs_writepage_fixup_worker(struct btrfs_work *work)
{
	struct btrfs_writepage_fixup *fixup;
	struct btrfs_ordered_extent *ordered;
	struct extent_state *cached_state = NULL;
	struct extent_changeset *data_reserved = NULL;
	struct page *page;
	struct inode *inode;
	u64 page_start;
	u64 page_end;
	int ret;

	fixup = container_of(work, struct btrfs_writepage_fixup, work);
	page = fixup->page;
again:
	lock_page(page);
	if (!page->mapping || !PageDirty(page) || !PageChecked(page)) {
		ClearPageChecked(page);
		goto out_page;
	}

	inode = page->mapping->host;
	page_start = page_offset(page);
	page_end = page_offset(page) + PAGE_SIZE - 1;

	lock_extent_bits(&BTRFS_I(inode)->io_tree, page_start, page_end,
			 &cached_state);

	/* already ordered? We're done */
	if (PagePrivate2(page))
		goto out;

	ordered = btrfs_lookup_ordered_range(BTRFS_I(inode), page_start,
					PAGE_SIZE);
	if (ordered) {
		unlock_extent_cached(&BTRFS_I(inode)->io_tree, page_start,
				     page_end, &cached_state);
		unlock_page(page);
		btrfs_start_ordered_extent(inode, ordered, 1);
		btrfs_put_ordered_extent(ordered);
		goto again;
	}

	ret = btrfs_delalloc_reserve_space(inode, &data_reserved, page_start,
					   PAGE_SIZE);
	if (ret) {
		mapping_set_error(page->mapping, ret);
		end_extent_writepage(page, ret, page_start, page_end);
		ClearPageChecked(page);
		goto out;
	 }

	ret = btrfs_set_extent_delalloc(inode, page_start, page_end, 0,
					&cached_state, 0);
	if (ret) {
		mapping_set_error(page->mapping, ret);
		end_extent_writepage(page, ret, page_start, page_end);
		ClearPageChecked(page);
		goto out;
	}

	ClearPageChecked(page);
	set_page_dirty(page);
	btrfs_delalloc_release_extents(BTRFS_I(inode), PAGE_SIZE, false);
out:
	unlock_extent_cached(&BTRFS_I(inode)->io_tree, page_start, page_end,
			     &cached_state);
out_page:
	unlock_page(page);
	put_page(page);
	kfree(fixup);
	extent_changeset_free(data_reserved);
}

/*
 * There are a few paths in the higher layers of the kernel that directly
 * set the page dirty bit without asking the filesystem if it is a
 * good idea.  This causes problems because we want to make sure COW
 * properly happens and the data=ordered rules are followed.
 *
 * In our case any range that doesn't have the ORDERED bit set
 * hasn't been properly setup for IO.  We kick off an async process
 * to fix it up.  The async helper will wait for ordered extents, set
 * the delalloc bit and make it safe to write the page.
 */
static int btrfs_writepage_start_hook(struct page *page, u64 start, u64 end)
{
	struct inode *inode = page->mapping->host;
	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
	struct btrfs_writepage_fixup *fixup;

	/* this page is properly in the ordered list */
	if (TestClearPagePrivate2(page))
		return 0;

	if (PageChecked(page))
		return -EAGAIN;

	fixup = kzalloc(sizeof(*fixup), GFP_NOFS);
	if (!fixup)
		return -EAGAIN;

	SetPageChecked(page);
	get_page(page);
	btrfs_init_work(&fixup->work, btrfs_fixup_helper,
			btrfs_writepage_fixup_worker, NULL, NULL);
	fixup->page = page;
	btrfs_queue_work(fs_info->fixup_workers, &fixup->work);
	return -EBUSY;
}

static int insert_reserved_file_extent(struct btrfs_trans_handle *trans,
				       struct inode *inode, u64 file_pos,
				       u64 disk_bytenr, u64 disk_num_bytes,
				       u64 num_bytes, u64 ram_bytes,
				       u8 compression, u8 encryption,
				       u16 other_encoding, int extent_type)
{
	struct btrfs_root *root = BTRFS_I(inode)->root;
	struct btrfs_file_extent_item *fi;
	struct btrfs_path *path;
	struct extent_buffer *leaf;
	struct btrfs_key ins;
	u64 qg_released;
	int extent_inserted = 0;
	int ret;

	path = btrfs_alloc_path();
	if (!path)
		return -ENOMEM;

	/*
	 * we may be replacing one extent in the tree with another.
	 * The new extent is pinned in the extent map, and we don't want
	 * to drop it from the cache until it is completely in the btree.
	 *
	 * So, tell btrfs_drop_extents to leave this extent in the cache.
	 * the caller is expected to unpin it and allow it to be merged
	 * with the others.
	 */
	ret = __btrfs_drop_extents(trans, root, inode, path, file_pos,
				   file_pos + num_bytes, NULL, 0,
				   1, sizeof(*fi), &extent_inserted);
	if (ret)
		goto out;

	if (!extent_inserted) {
		ins.objectid = btrfs_ino(BTRFS_I(inode));
		ins.offset = file_pos;
		ins.type = BTRFS_EXTENT_DATA_KEY;

		path->leave_spinning = 1;
		ret = btrfs_insert_empty_item(trans, root, path, &ins,
					      sizeof(*fi));
		if (ret)
			goto out;
	}
	leaf = path->nodes[0];
	fi = btrfs_item_ptr(leaf, path->slots[0],
			    struct btrfs_file_extent_item);
	btrfs_set_file_extent_generation(leaf, fi, trans->transid);
	btrfs_set_file_extent_type(leaf, fi, extent_type);
	btrfs_set_file_extent_disk_bytenr(leaf, fi, disk_bytenr);
	btrfs_set_file_extent_disk_num_bytes(leaf, fi, disk_num_bytes);
	btrfs_set_file_extent_offset(leaf, fi, 0);
	btrfs_set_file_extent_num_bytes(leaf, fi, num_bytes);
	btrfs_set_file_extent_ram_bytes(leaf, fi, ram_bytes);
	btrfs_set_file_extent_compression(leaf, fi, compression);
	btrfs_set_file_extent_encryption(leaf, fi, encryption);
	btrfs_set_file_extent_other_encoding(leaf, fi, other_encoding);

	btrfs_mark_buffer_dirty(leaf);
	btrfs_release_path(path);

	inode_add_bytes(inode, num_bytes);

	ins.objectid = disk_bytenr;
	ins.offset = disk_num_bytes;
	ins.type = BTRFS_EXTENT_ITEM_KEY;

	/*
	 * Release the reserved range from inode dirty range map, as it is
	 * already moved into delayed_ref_head
	 */
	ret = btrfs_qgroup_release_data(inode, file_pos, ram_bytes);
	if (ret < 0)
		goto out;
	qg_released = ret;
	ret = btrfs_alloc_reserved_file_extent(trans, root,
					       btrfs_ino(BTRFS_I(inode)),
					       file_pos, qg_released, &ins);
out:
	btrfs_free_path(path);

	return ret;
}

/* snapshot-aware defrag */
struct sa_defrag_extent_backref {
	struct rb_node node;
	struct old_sa_defrag_extent *old;
	u64 root_id;
	u64 inum;
	u64 file_pos;
	u64 extent_offset;
	u64 num_bytes;
	u64 generation;
};

struct old_sa_defrag_extent {
	struct list_head list;
	struct new_sa_defrag_extent *new;

	u64 extent_offset;
	u64 bytenr;
	u64 offset;
	u64 len;
	int count;
};

struct new_sa_defrag_extent {
	struct rb_root root;
	struct list_head head;
	struct btrfs_path *path;
	struct inode *inode;
	u64 file_pos;
	u64 len;
	u64 bytenr;
	u64 disk_len;
	u8 compress_type;
};

static int backref_comp(struct sa_defrag_extent_backref *b1,
			struct sa_defrag_extent_backref *b2)
{
	if (b1->root_id < b2->root_id)
		return -1;
	else if (b1->root_id > b2->root_id)
		return 1;

	if (b1->inum < b2->inum)
		return -1;
	else if (b1->inum > b2->inum)
		return 1;

	if (b1->file_pos < b2->file_pos)
		return -1;
	else if (b1->file_pos > b2->file_pos)
		return 1;

	/*
	 * [------------------------------] ===> (a range of space)
	 *     |<--->|   |<---->| =============> (fs/file tree A)
	 * |<---------------------------->| ===> (fs/file tree B)
	 *
	 * A range of space can refer to two file extents in one tree while
	 * refer to only one file extent in another tree.
	 *
	 * So we may process a disk offset more than one time(two extents in A)
	 * and locate at the same extent(one extent in B), then insert two same
	 * backrefs(both refer to the extent in B).
	 */
	return 0;
}

static void backref_insert(struct rb_root *root,
			   struct sa_defrag_extent_backref *backref)
{
	struct rb_node **p = &root->rb_node;
	struct rb_node *parent = NULL;
	struct sa_defrag_extent_backref *entry;
	int ret;

	while (*p) {
		parent = *p;
		entry = rb_entry(parent, struct sa_defrag_extent_backref, node);

		ret = backref_comp(backref, entry);
		if (ret < 0)
			p = &(*p)->rb_left;
		else
			p = &(*p)->rb_right;
	}

	rb_link_node(&backref->node, parent, p);
	rb_insert_color(&backref->node, root);
}

/*
 * Note the backref might has changed, and in this case we just return 0.
 */
static noinline int record_one_backref(u64 inum, u64 offset, u64 root_id,
				       void *ctx)
{
	struct btrfs_file_extent_item *extent;
	struct old_sa_defrag_extent *old = ctx;
	struct new_sa_defrag_extent *new = old->new;
	struct btrfs_path *path = new->path;
	struct btrfs_key key;
	struct btrfs_root *root;
	struct sa_defrag_extent_backref *backref;
	struct extent_buffer *leaf;
	struct inode *inode = new->inode;
	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
	int slot;
	int ret;
	u64 extent_offset;
	u64 num_bytes;

	if (BTRFS_I(inode)->root->root_key.objectid == root_id &&
	    inum == btrfs_ino(BTRFS_I(inode)))
		return 0;

	key.objectid = root_id;
	key.type = BTRFS_ROOT_ITEM_KEY;
	key.offset = (u64)-1;

	root = btrfs_read_fs_root_no_name(fs_info, &key);
	if (IS_ERR(root)) {
		if (PTR_ERR(root) == -ENOENT)
			return 0;
		WARN_ON(1);
		btrfs_debug(fs_info, "inum=%llu, offset=%llu, root_id=%llu",
			 inum, offset, root_id);
		return PTR_ERR(root);
	}

	key.objectid = inum;
	key.type = BTRFS_EXTENT_DATA_KEY;
	if (offset > (u64)-1 << 32)
		key.offset = 0;
	else
		key.offset = offset;

	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
	if (WARN_ON(ret < 0))
		return ret;
	ret = 0;

	while (1) {
		cond_resched();

		leaf = path->nodes[0];
		slot = path->slots[0];

		if (slot >= btrfs_header_nritems(leaf)) {
			ret = btrfs_next_leaf(root, path);
			if (ret < 0) {
				goto out;
			} else if (ret > 0) {
				ret = 0;
				goto out;
			}
			continue;
		}

		path->slots[0]++;

		btrfs_item_key_to_cpu(leaf, &key, slot);

		if (key.objectid > inum)
			goto out;

		if (key.objectid < inum || key.type != BTRFS_EXTENT_DATA_KEY)
			continue;

		extent = btrfs_item_ptr(leaf, slot,
					struct btrfs_file_extent_item);

		if (btrfs_file_extent_disk_bytenr(leaf, extent) != old->bytenr)
			continue;

		/*
		 * 'offset' refers to the exact key.offset,
		 * NOT the 'offset' field in btrfs_extent_data_ref, ie.
		 * (key.offset - extent_offset).
		 */
		if (key.offset != offset)
			continue;

		extent_offset = btrfs_file_extent_offset(leaf, extent);
		num_bytes = btrfs_file_extent_num_bytes(leaf, extent);

		if (extent_offset >= old->extent_offset + old->offset +
		    old->len || extent_offset + num_bytes <=
		    old->extent_offset + old->offset)
			continue;
		break;
	}

	backref = kmalloc(sizeof(*backref), GFP_NOFS);
	if (!backref) {
		ret = -ENOENT;
		goto out;
	}

	backref->root_id = root_id;
	backref->inum = inum;
	backref->file_pos = offset;
	backref->num_bytes = num_bytes;
	backref->extent_offset = extent_offset;
	backref->generation = btrfs_file_extent_generation(leaf, extent);
	backref->old = old;
	backref_insert(&new->root, backref);
	old->count++;
out:
	btrfs_release_path(path);
	WARN_ON(ret);
	return ret;
}

static noinline bool record_extent_backrefs(struct btrfs_path *path,
				   struct new_sa_defrag_extent *new)
{
	struct btrfs_fs_info *fs_info = btrfs_sb(new->inode->i_sb);
	struct old_sa_defrag_extent *old, *tmp;
	int ret;

	new->path = path;

	list_for_each_entry_safe(old, tmp, &new->head, list) {
		ret = iterate_inodes_from_logical(old->bytenr +
						  old->extent_offset, fs_info,
						  path, record_one_backref,
						  old, false);
		if (ret < 0 && ret != -ENOENT)
			return false;

		/* no backref to be processed for this extent */
		if (!old->count) {
			list_del(&old->list);
			kfree(old);
		}
	}

	if (list_empty(&new->head))
		return false;

	return true;
}

static int relink_is_mergable(struct extent_buffer *leaf,
			      struct btrfs_file_extent_item *fi,
			      struct new_sa_defrag_extent *new)
{
	if (btrfs_file_extent_disk_bytenr(leaf, fi) != new->bytenr)
		return 0;

	if (btrfs_file_extent_type(leaf, fi) != BTRFS_FILE_EXTENT_REG)
		return 0;

	if (btrfs_file_extent_compression(leaf, fi) != new->compress_type)
		return 0;

	if (btrfs_file_extent_encryption(leaf, fi) ||
	    btrfs_file_extent_other_encoding(leaf, fi))
		return 0;

	return 1;
}

/*
 * Note the backref might has changed, and in this case we just return 0.
 */
static noinline int relink_extent_backref(struct btrfs_path *path,
				 struct sa_defrag_extent_backref *prev,
				 struct sa_defrag_extent_backref *backref)
{
	struct btrfs_file_extent_item *extent;
	struct btrfs_file_extent_item *item;
	struct btrfs_ordered_extent *ordered;
	struct btrfs_trans_handle *trans;
	struct btrfs_root *root;
	struct btrfs_key key;
	struct extent_buffer *leaf;
	struct old_sa_defrag_extent *old = backref->old;
	struct new_sa_defrag_extent *new = old->new;
	struct btrfs_fs_info *fs_info = btrfs_sb(new->inode->i_sb);
	struct inode *inode;
	struct extent_state *cached = NULL;
	int ret = 0;
	u64 start;
	u64 len;
	u64 lock_start;
	u64 lock_end;
	bool merge = false;
	int index;

	if (prev && prev->root_id == backref->root_id &&
	    prev->inum == backref->inum &&
	    prev->file_pos + prev->num_bytes == backref->file_pos)
		merge = true;

	/* step 1: get root */
	key.objectid = backref->root_id;
	key.type = BTRFS_ROOT_ITEM_KEY;
	key.offset = (u64)-1;

	index = srcu_read_lock(&fs_info->subvol_srcu);

	root = btrfs_read_fs_root_no_name(fs_info, &key);
	if (IS_ERR(root)) {
		srcu_read_unlock(&fs_info->subvol_srcu, index);
		if (PTR_ERR(root) == -ENOENT)
			return 0;
		return PTR_ERR(root);
	}

	if (btrfs_root_readonly(root)) {
		srcu_read_unlock(&fs_info->subvol_srcu, index);
		return 0;
	}

	/* step 2: get inode */
	key.objectid = backref->inum;
	key.type = BTRFS_INODE_ITEM_KEY;
	key.offset = 0;

	inode = btrfs_iget(fs_info->sb, &key, root, NULL);
	if (IS_ERR(inode)) {
		srcu_read_unlock(&fs_info->subvol_srcu, index);
		return 0;
	}

	srcu_read_unlock(&fs_info->subvol_srcu, index);

	/* step 3: relink backref */
	lock_start = backref->file_pos;
	lock_end = backref->file_pos + backref->num_bytes - 1;
	lock_extent_bits(&BTRFS_I(inode)->io_tree, lock_start, lock_end,
			 &cached);

	ordered = btrfs_lookup_first_ordered_extent(inode, lock_end);
	if (ordered) {
		btrfs_put_ordered_extent(ordered);
		goto out_unlock;
	}

	trans = btrfs_join_transaction(root);
	if (IS_ERR(trans)) {
		ret = PTR_ERR(trans);
		goto out_unlock;
	}

	key.objectid = backref->inum;
	key.type = BTRFS_EXTENT_DATA_KEY;
	key.offset = backref->file_pos;

	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
	if (ret < 0) {
		goto out_free_path;
	} else if (ret > 0) {
		ret = 0;
		goto out_free_path;
	}

	extent = btrfs_item_ptr(path->nodes[0], path->slots[0],
				struct btrfs_file_extent_item);

	if (btrfs_file_extent_generation(path->nodes[0], extent) !=
	    backref->generation)
		goto out_free_path;

	btrfs_release_path(path);

	start = backref->file_pos;
	if (backref->extent_offset < old->extent_offset + old->offset)
		start += old->extent_offset + old->offset -
			 backref->extent_offset;

	len = min(backref->extent_offset + backref->num_bytes,
		  old->extent_offset + old->offset + old->len);
	len -= max(backref->extent_offset, old->extent_offset + old->offset);

	ret = btrfs_drop_extents(trans, root, inode, start,
				 start + len, 1);
	if (ret)
		goto out_free_path;
again:
	key.objectid = btrfs_ino(BTRFS_I(inode));
	key.type = BTRFS_EXTENT_DATA_KEY;
	key.offset = start;

	path->leave_spinning = 1;
	if (merge) {
		struct btrfs_file_extent_item *fi;
		u64 extent_len;
		struct btrfs_key found_key;

		ret = btrfs_search_slot(trans, root, &key, path, 0, 1);
		if (ret < 0)
			goto out_free_path;

		path->slots[0]--;
		leaf = path->nodes[0];
		btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);

		fi = btrfs_item_ptr(leaf, path->slots[0],
				    struct btrfs_file_extent_item);
		extent_len = btrfs_file_extent_num_bytes(leaf, fi);

		if (extent_len + found_key.offset == start &&
		    relink_is_mergable(leaf, fi, new)) {
			btrfs_set_file_extent_num_bytes(leaf, fi,
							extent_len + len);
			btrfs_mark_buffer_dirty(leaf);
			inode_add_bytes(inode, len);

			ret = 1;
			goto out_free_path;
		} else {
			merge = false;
			btrfs_release_path(path);
			goto again;
		}
	}

	ret = btrfs_insert_empty_item(trans, root, path, &key,
					sizeof(*extent));
	if (ret) {
		btrfs_abort_transaction(trans, ret);
		goto out_free_path;
	}

	leaf = path->nodes[0];
	item = btrfs_item_ptr(leaf, path->slots[0],
				struct btrfs_file_extent_item);
	btrfs_set_file_extent_disk_bytenr(leaf, item, new->bytenr);
	btrfs_set_file_extent_disk_num_bytes(leaf, item, new->disk_len);
	btrfs_set_file_extent_offset(leaf, item, start - new->file_pos);
	btrfs_set_file_extent_num_bytes(leaf, item, len);
	btrfs_set_file_extent_ram_bytes(leaf, item, new->len);
	btrfs_set_file_extent_generation(leaf, item, trans->transid);
	btrfs_set_file_extent_type(leaf, item, BTRFS_FILE_EXTENT_REG);
	btrfs_set_file_extent_compression(leaf, item, new->compress_type);
	btrfs_set_file_extent_encryption(leaf, item, 0);
	btrfs_set_file_extent_other_encoding(leaf, item, 0);

	btrfs_mark_buffer_dirty(leaf);
	inode_add_bytes(inode, len);
	btrfs_release_path(path);

	ret = btrfs_inc_extent_ref(trans, root, new->bytenr,
			new->disk_len, 0,
			backref->root_id, backref->inum,
			new->file_pos);	/* start - extent_offset */
	if (ret) {
		btrfs_abort_transaction(trans, ret);
		goto out_free_path;
	}

	ret = 1;
out_free_path:
	btrfs_release_path(path);
	path->leave_spinning = 0;
	btrfs_end_transaction(trans);
out_unlock:
	unlock_extent_cached(&BTRFS_I(inode)->io_tree, lock_start, lock_end,
			     &cached);
	iput(inode);
	return ret;
}

static void free_sa_defrag_extent(struct new_sa_defrag_extent *new)
{
	struct old_sa_defrag_extent *old, *tmp;

	if (!new)
		return;

	list_for_each_entry_safe(old, tmp, &new->head, list) {
		kfree(old);
	}
	kfree(new);
}

static void relink_file_extents(struct new_sa_defrag_extent *new)
{
	struct btrfs_fs_info *fs_info = btrfs_sb(new->inode->i_sb);
	struct btrfs_path *path;
	struct sa_defrag_extent_backref *backref;
	struct sa_defrag_extent_backref *prev = NULL;
	struct inode *inode;
	struct rb_node *node;
	int ret;

	inode = new->inode;

	path = btrfs_alloc_path();
	if (!path)
		return;

	if (!record_extent_backrefs(path, new)) {
		btrfs_free_path(path);
		goto out;
	}
	btrfs_release_path(path);

	while (1) {
		node = rb_first(&new->root);
		if (!node)
			break;
		rb_erase(node, &new->root);

		backref = rb_entry(node, struct sa_defrag_extent_backref, node);

		ret = relink_extent_backref(path, prev, backref);
		WARN_ON(ret < 0);

		kfree(prev);

		if (ret == 1)
			prev = backref;
		else
			prev = NULL;
		cond_resched();
	}
	kfree(prev);

	btrfs_free_path(path);
out:
	free_sa_defrag_extent(new);

	atomic_dec(&fs_info->defrag_running);
	wake_up(&fs_info->transaction_wait);
}

static struct new_sa_defrag_extent *
record_old_file_extents(struct inode *inode,
			struct btrfs_ordered_extent *ordered)
{
	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
	struct btrfs_root *root = BTRFS_I(inode)->root;
	struct btrfs_path *path;
	struct btrfs_key key;
	struct old_sa_defrag_extent *old;
	struct new_sa_defrag_extent *new;
	int ret;

	new = kmalloc(sizeof(*new), GFP_NOFS);
	if (!new)
		return NULL;

	new->inode = inode;
	new->file_pos = ordered->file_offset;
	new->len = ordered->len;
	new->bytenr = ordered->start;
	new->disk_len = ordered->disk_len;
	new->compress_type = ordered->compress_type;
	new->root = RB_ROOT;
	INIT_LIST_HEAD(&new->head);

	path = btrfs_alloc_path();
	if (!path)
		goto out_kfree;

	key.objectid = btrfs_ino(BTRFS_I(inode));
	key.type = BTRFS_EXTENT_DATA_KEY;
	key.offset = new->file_pos;

	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
	if (ret < 0)
		goto out_free_path;
	if (ret > 0 && path->slots[0] > 0)
		path->slots[0]--;

	/* find out all the old extents for the file range */
	while (1) {
		struct btrfs_file_extent_item *extent;
		struct extent_buffer *l;
		int slot;
		u64 num_bytes;
		u64 offset;
		u64 end;
		u64 disk_bytenr;
		u64 extent_offset;

		l = path->nodes[0];
		slot = path->slots[0];

		if (slot >= btrfs_header_nritems(l)) {
			ret = btrfs_next_leaf(root, path);
			if (ret < 0)
				goto out_free_path;
			else if (ret > 0)
				break;
			continue;
		}

		btrfs_item_key_to_cpu(l, &key, slot);

		if (key.objectid != btrfs_ino(BTRFS_I(inode)))
			break;
		if (key.type != BTRFS_EXTENT_DATA_KEY)
			break;
		if (key.offset >= new->file_pos + new->len)
			break;

		extent = btrfs_item_ptr(l, slot, struct btrfs_file_extent_item);

		num_bytes = btrfs_file_extent_num_bytes(l, extent);
		if (key.offset + num_bytes < new->file_pos)
			goto next;

		disk_bytenr = btrfs_file_extent_disk_bytenr(l, extent);
		if (!disk_bytenr)
			goto next;

		extent_offset = btrfs_file_extent_offset(l, extent);

		old = kmalloc(sizeof(*old), GFP_NOFS);
		if (!old)
			goto out_free_path;

		offset = max(new->file_pos, key.offset);
		end = min(new->file_pos + new->len, key.offset + num_bytes);

		old->bytenr = disk_bytenr;
		old->extent_offset = extent_offset;
		old->offset = offset - key.offset;
		old->len = end - offset;
		old->new = new;
		old->count = 0;
		list_add_tail(&old->list, &new->head);
next:
		path->slots[0]++;
		cond_resched();
	}

	btrfs_free_path(path);
	atomic_inc(&fs_info->defrag_running);

	return new;

out_free_path:
	btrfs_free_path(path);
out_kfree:
	free_sa_defrag_extent(new);
	return NULL;
}

static void btrfs_release_delalloc_bytes(struct btrfs_fs_info *fs_info,
					 u64 start, u64 len)
{
	struct btrfs_block_group_cache *cache;

	cache = btrfs_lookup_block_group(fs_info, start);
	ASSERT(cache);

	spin_lock(&cache->lock);
	cache->delalloc_bytes -= len;
	spin_unlock(&cache->lock);

	btrfs_put_block_group(cache);
}

/* as ordered data IO finishes, this gets called so we can finish
 * an ordered extent if the range of bytes in the file it covers are
 * fully written.
 */
static int btrfs_finish_ordered_io(struct btrfs_ordered_extent *ordered_extent)
{
	struct inode *inode = ordered_extent->inode;
	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
	struct btrfs_root *root = BTRFS_I(inode)->root;
	struct btrfs_trans_handle *trans = NULL;
	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
	struct extent_state *cached_state = NULL;
	struct new_sa_defrag_extent *new = NULL;
	int compress_type = 0;
	int ret = 0;
	u64 logical_len = ordered_extent->len;
	bool nolock;
	bool truncated = false;
	bool range_locked = false;
	bool clear_new_delalloc_bytes = false;

	if (!test_bit(BTRFS_ORDERED_NOCOW, &ordered_extent->flags) &&
	    !test_bit(BTRFS_ORDERED_PREALLOC, &ordered_extent->flags) &&
	    !test_bit(BTRFS_ORDERED_DIRECT, &ordered_extent->flags))
		clear_new_delalloc_bytes = true;

	nolock = btrfs_is_free_space_inode(BTRFS_I(inode));

	if (test_bit(BTRFS_ORDERED_IOERR, &ordered_extent->flags)) {
		ret = -EIO;
		goto out;
	}

	btrfs_free_io_failure_record(BTRFS_I(inode),
			ordered_extent->file_offset,
			ordered_extent->file_offset +
			ordered_extent->len - 1);

	if (test_bit(BTRFS_ORDERED_TRUNCATED, &ordered_extent->flags)) {
		truncated = true;
		logical_len = ordered_extent->truncated_len;
		/* Truncated the entire extent, don't bother adding */
		if (!logical_len)
			goto out;
	}

	if (test_bit(BTRFS_ORDERED_NOCOW, &ordered_extent->flags)) {
		BUG_ON(!list_empty(&ordered_extent->list)); /* Logic error */

		/*
		 * For mwrite(mmap + memset to write) case, we still reserve
		 * space for NOCOW range.
		 * As NOCOW won't cause a new delayed ref, just free the space
		 */
		btrfs_qgroup_free_data(inode, NULL, ordered_extent->file_offset,
				       ordered_extent->len);
		btrfs_ordered_update_i_size(inode, 0, ordered_extent);
		if (nolock)
			trans = btrfs_join_transaction_nolock(root);
		else
			trans = btrfs_join_transaction(root);
		if (IS_ERR(trans)) {
			ret = PTR_ERR(trans);
			trans = NULL;
			goto out;
		}
		trans->block_rsv = &BTRFS_I(inode)->block_rsv;
		ret = btrfs_update_inode_fallback(trans, root, inode);
		if (ret) /* -ENOMEM or corruption */
			btrfs_abort_transaction(trans, ret);
		goto out;
	}

	range_locked = true;
	lock_extent_bits(io_tree, ordered_extent->file_offset,
			 ordered_extent->file_offset + ordered_extent->len - 1,
			 &cached_state);

	ret = test_range_bit(io_tree, ordered_extent->file_offset,
			ordered_extent->file_offset + ordered_extent->len - 1,
			EXTENT_DEFRAG, 0, cached_state);
	if (ret) {
		u64 last_snapshot = btrfs_root_last_snapshot(&root->root_item);
		if (0 && last_snapshot >= BTRFS_I(inode)->generation)
			/* the inode is shared */
			new = record_old_file_extents(inode, ordered_extent);

		clear_extent_bit(io_tree, ordered_extent->file_offset,
			ordered_extent->file_offset + ordered_extent->len - 1,
			EXTENT_DEFRAG, 0, 0, &cached_state);
	}

	if (nolock)
		trans = btrfs_join_transaction_nolock(root);
	else
		trans = btrfs_join_transaction(root);
	if (IS_ERR(trans)) {
		ret = PTR_ERR(trans);
		trans = NULL;
		goto out;
	}

	trans->block_rsv = &BTRFS_I(inode)->block_rsv;

	if (test_bit(BTRFS_ORDERED_COMPRESSED, &ordered_extent->flags))
		compress_type = ordered_extent->compress_type;
	if (test_bit(BTRFS_ORDERED_PREALLOC, &ordered_extent->flags)) {
		BUG_ON(compress_type);
		btrfs_qgroup_free_data(inode, NULL, ordered_extent->file_offset,
				       ordered_extent->len);
		ret = btrfs_mark_extent_written(trans, BTRFS_I(inode),
						ordered_extent->file_offset,
						ordered_extent->file_offset +
						logical_len);
	} else {
		BUG_ON(root == fs_info->tree_root);
		ret = insert_reserved_file_extent(trans, inode,
						ordered_extent->file_offset,
						ordered_extent->start,
						ordered_extent->disk_len,
						logical_len, logical_len,
						compress_type, 0, 0,
						BTRFS_FILE_EXTENT_REG);
		if (!ret)
			btrfs_release_delalloc_bytes(fs_info,
						     ordered_extent->start,
						     ordered_extent->disk_len);
	}
	unpin_extent_cache(&BTRFS_I(inode)->extent_tree,
			   ordered_extent->file_offset, ordered_extent->len,
			   trans->transid);
	if (ret < 0) {
		btrfs_abort_transaction(trans, ret);
		goto out;
	}

	ret = add_pending_csums(trans, inode, &ordered_extent->list);
	if (ret) {
		btrfs_abort_transaction(trans, ret);
		goto out;
	}

	btrfs_ordered_update_i_size(inode, 0, ordered_extent);
	ret = btrfs_update_inode_fallback(trans, root, inode);
	if (ret) { /* -ENOMEM or corruption */
		btrfs_abort_transaction(trans, ret);
		goto out;
	}
	ret = 0;
out:
	if (range_locked || clear_new_delalloc_bytes) {
		unsigned int clear_bits = 0;

		if (range_locked)
			clear_bits |= EXTENT_LOCKED;
		if (clear_new_delalloc_bytes)
			clear_bits |= EXTENT_DELALLOC_NEW;
		clear_extent_bit(&BTRFS_I(inode)->io_tree,
				 ordered_extent->file_offset,
				 ordered_extent->file_offset +
				 ordered_extent->len - 1,
				 clear_bits,
				 (clear_bits & EXTENT_LOCKED) ? 1 : 0,
				 0, &cached_state);
	}

	if (trans)
		btrfs_end_transaction(trans);

	if (ret || truncated) {
		u64 start, end;

		if (truncated)
			start = ordered_extent->file_offset + logical_len;
		else
			start = ordered_extent->file_offset;
		end = ordered_extent->file_offset + ordered_extent->len - 1;
		clear_extent_uptodate(io_tree, start, end, NULL);

		/* Drop the cache for the part of the extent we didn't write. */
		btrfs_drop_extent_cache(BTRFS_I(inode), start, end, 0);

		/*
		 * If the ordered extent had an IOERR or something else went
		 * wrong we need to return the space for this ordered extent
		 * back to the allocator.  We only free the extent in the
		 * truncated case if we didn't write out the extent at all.
		 */
		if ((ret || !logical_len) &&
		    !test_bit(BTRFS_ORDERED_NOCOW, &ordered_extent->flags) &&
		    !test_bit(BTRFS_ORDERED_PREALLOC, &ordered_extent->flags))
			btrfs_free_reserved_extent(fs_info,
						   ordered_extent->start,
						   ordered_extent->disk_len, 1);
	}


	/*
	 * This needs to be done to make sure anybody waiting knows we are done
	 * updating everything for this ordered extent.
	 */
	btrfs_remove_ordered_extent(inode, ordered_extent);

	/* for snapshot-aware defrag */
	if (new) {
		if (ret) {
			free_sa_defrag_extent(new);
			atomic_dec(&fs_info->defrag_running);
		} else {
			relink_file_extents(new);
		}
	}

	/* once for us */
	btrfs_put_ordered_extent(ordered_extent);
	/* once for the tree */
	btrfs_put_ordered_extent(ordered_extent);

	/* Try to release some metadata so we don't get an OOM but don't wait */
	btrfs_btree_balance_dirty_nodelay(fs_info);

	return ret;
}

static void finish_ordered_fn(struct btrfs_work *work)
{
	struct btrfs_ordered_extent *ordered_extent;
	ordered_extent = container_of(work, struct btrfs_ordered_extent, work);
	btrfs_finish_ordered_io(ordered_extent);
}

static void btrfs_writepage_end_io_hook(struct page *page, u64 start, u64 end,
				struct extent_state *state, int uptodate)
{
	struct inode *inode = page->mapping->host;
	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
	struct btrfs_ordered_extent *ordered_extent = NULL;
	struct btrfs_workqueue *wq;
	btrfs_work_func_t func;

	trace_btrfs_writepage_end_io_hook(page, start, end, uptodate);

	ClearPagePrivate2(page);
	if (!btrfs_dec_test_ordered_pending(inode, &ordered_extent, start,
					    end - start + 1, uptodate))
		return;

	if (btrfs_is_free_space_inode(BTRFS_I(inode))) {
		wq = fs_info->endio_freespace_worker;
		func = btrfs_freespace_write_helper;
	} else {
		wq = fs_info->endio_write_workers;
		func = btrfs_endio_write_helper;
	}

	btrfs_init_work(&ordered_extent->work, func, finish_ordered_fn, NULL,
			NULL);
	btrfs_queue_work(wq, &ordered_extent->work);
}

static int __readpage_endio_check(struct inode *inode,
				  struct btrfs_io_bio *io_bio,
				  int icsum, struct page *page,
				  int pgoff, u64 start, size_t len)
{
	char *kaddr;
	u32 csum_expected;
	u32 csum = ~(u32)0;

	csum_expected = *(((u32 *)io_bio->csum) + icsum);

	kaddr = kmap_atomic(page);
	csum = btrfs_csum_data(kaddr + pgoff, csum,  len);
	btrfs_csum_final(csum, (u8 *)&csum);
	if (csum != csum_expected)
		goto zeroit;

	kunmap_atomic(kaddr);
	return 0;
zeroit:
	btrfs_print_data_csum_error(BTRFS_I(inode), start, csum, csum_expected,
				    io_bio->mirror_num);
	memset(kaddr + pgoff, 1, len);
	flush_dcache_page(page);
	kunmap_atomic(kaddr);
	return -EIO;
}

/*
 * when reads are done, we need to check csums to verify the data is correct
 * if there's a match, we allow the bio to finish.  If not, the code in
 * extent_io.c will try to find good copies for us.
 */
static int btrfs_readpage_end_io_hook(struct btrfs_io_bio *io_bio,
				      u64 phy_offset, struct page *page,
				      u64 start, u64 end, int mirror)
{
	size_t offset = start - page_offset(page);
	struct inode *inode = page->mapping->host;
	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
	struct btrfs_root *root = BTRFS_I(inode)->root;

	if (PageChecked(page)) {
		ClearPageChecked(page);
		return 0;
	}

	if (BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM)
		return 0;

	if (root->root_key.objectid == BTRFS_DATA_RELOC_TREE_OBJECTID &&
	    test_range_bit(io_tree, start, end, EXTENT_NODATASUM, 1, NULL)) {
		clear_extent_bits(io_tree, start, end, EXTENT_NODATASUM);
		return 0;
	}

	phy_offset >>= inode->i_sb->s_blocksize_bits;
	return __readpage_endio_check(inode, io_bio, phy_offset, page, offset,
				      start, (size_t)(end - start + 1));
}

/*
 * btrfs_add_delayed_iput - perform a delayed iput on @inode
 *
 * @inode: The inode we want to perform iput on
 *
 * This function uses the generic vfs_inode::i_count to track whether we should
 * just decrement it (in case it's > 1) or if this is the last iput then link
 * the inode to the delayed iput machinery. Delayed iputs are processed at
 * transaction commit time/superblock commit/cleaner kthread.
 */
void btrfs_add_delayed_iput(struct inode *inode)
{
	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
	struct btrfs_inode *binode = BTRFS_I(inode);

	if (atomic_add_unless(&inode->i_count, -1, 1))
		return;

	spin_lock(&fs_info->delayed_iput_lock);
	ASSERT(list_empty(&binode->delayed_iput));
	list_add_tail(&binode->delayed_iput, &fs_info->delayed_iputs);
	spin_unlock(&fs_info->delayed_iput_lock);
}

void btrfs_run_delayed_iputs(struct btrfs_fs_info *fs_info)
{

	spin_lock(&fs_info->delayed_iput_lock);
	while (!list_empty(&fs_info->delayed_iputs)) {
		struct btrfs_inode *inode;

		inode = list_first_entry(&fs_info->delayed_iputs,
				struct btrfs_inode, delayed_iput);
		list_del_init(&inode->delayed_iput);
		spin_unlock(&fs_info->delayed_iput_lock);
		iput(&inode->vfs_inode);
		spin_lock(&fs_info->delayed_iput_lock);
	}
	spin_unlock(&fs_info->delayed_iput_lock);
}

/*
 * This creates an orphan entry for the given inode in case something goes wrong
 * in the middle of an unlink.
 */
int btrfs_orphan_add(struct btrfs_trans_handle *trans,
		     struct btrfs_inode *inode)
{
	int ret;

	ret = btrfs_insert_orphan_item(trans, inode->root, btrfs_ino(inode));
	if (ret && ret != -EEXIST) {
		btrfs_abort_transaction(trans, ret);
		return ret;
	}

	return 0;
}

/*
 * We have done the delete so we can go ahead and remove the orphan item for
 * this particular inode.
 */
static int btrfs_orphan_del(struct btrfs_trans_handle *trans,
			    struct btrfs_inode *inode)
{
	return btrfs_del_orphan_item(trans, inode->root, btrfs_ino(inode));
}

/*
 * this cleans up any orphans that may be left on the list from the last use
 * of this root.
 */
int btrfs_orphan_cleanup(struct btrfs_root *root)
{
	struct btrfs_fs_info *fs_info = root->fs_info;
	struct btrfs_path *path;
	struct extent_buffer *leaf;
	struct btrfs_key key, found_key;
	struct btrfs_trans_handle *trans;
	struct inode *inode;
	u64 last_objectid = 0;
	int ret = 0, nr_unlink = 0;

	if (cmpxchg(&root->orphan_cleanup_state, 0, ORPHAN_CLEANUP_STARTED))
		return 0;

	path = btrfs_alloc_path();
	if (!path) {
		ret = -ENOMEM;
		goto out;
	}
	path->reada = READA_BACK;

	key.objectid = BTRFS_ORPHAN_OBJECTID;
	key.type = BTRFS_ORPHAN_ITEM_KEY;
	key.offset = (u64)-1;

	while (1) {
		ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
		if (ret < 0)
			goto out;

		/*
		 * if ret == 0 means we found what we were searching for, which
		 * is weird, but possible, so only screw with path if we didn't
		 * find the key and see if we have stuff that matches
		 */
		if (ret > 0) {
			ret = 0;
			if (path->slots[0] == 0)
				break;
			path->slots[0]--;
		}

		/* pull out the item */
		leaf = path->nodes[0];
		btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);

		/* make sure the item matches what we want */
		if (found_key.objectid != BTRFS_ORPHAN_OBJECTID)
			break;
		if (found_key.type != BTRFS_ORPHAN_ITEM_KEY)
			break;

		/* release the path since we're done with it */
		btrfs_release_path(path);

		/*
		 * this is where we are basically btrfs_lookup, without the
		 * crossing root thing.  we store the inode number in the
		 * offset of the orphan item.
		 */

		if (found_key.offset == last_objectid) {
			btrfs_err(fs_info,
				  "Error removing orphan entry, stopping orphan cleanup");
			ret = -EINVAL;
			goto out;
		}

		last_objectid = found_key.offset;

		found_key.objectid = found_key.offset;
		found_key.type = BTRFS_INODE_ITEM_KEY;
		found_key.offset = 0;
		inode = btrfs_iget(fs_info->sb, &found_key, root, NULL);
		ret = PTR_ERR_OR_ZERO(inode);
		if (ret && ret != -ENOENT)
			goto out;

		if (ret == -ENOENT && root == fs_info->tree_root) {
			struct btrfs_root *dead_root;
			struct btrfs_fs_info *fs_info = root->fs_info;
			int is_dead_root = 0;

			/*
			 * this is an orphan in the tree root. Currently these
			 * could come from 2 sources:
			 *  a) a snapshot deletion in progress
			 *  b) a free space cache inode
			 * We need to distinguish those two, as the snapshot
			 * orphan must not get deleted.
			 * find_dead_roots already ran before us, so if this
			 * is a snapshot deletion, we should find the root
			 * in the dead_roots list
			 */
			spin_lock(&fs_info->trans_lock);
			list_for_each_entry(dead_root, &fs_info->dead_roots,
					    root_list) {
				if (dead_root->root_key.objectid ==
				    found_key.objectid) {
					is_dead_root = 1;
					break;
				}
			}
			spin_unlock(&fs_info->trans_lock);
			if (is_dead_root) {
				/* prevent this orphan from being found again */
				key.offset = found_key.objectid - 1;
				continue;
			}

		}

		/*
		 * If we have an inode with links, there are a couple of
		 * possibilities. Old kernels (before v3.12) used to create an
		 * orphan item for truncate indicating that there were possibly
		 * extent items past i_size that needed to be deleted. In v3.12,
		 * truncate was changed to update i_size in sync with the extent
		 * items, but the (useless) orphan item was still created. Since
		 * v4.18, we don't create the orphan item for truncate at all.
		 *
		 * So, this item could mean that we need to do a truncate, but
		 * only if this filesystem was last used on a pre-v3.12 kernel
		 * and was not cleanly unmounted. The odds of that are quite
		 * slim, and it's a pain to do the truncate now, so just delete
		 * the orphan item.
		 *
		 * It's also possible that this orphan item was supposed to be
		 * deleted but wasn't. The inode number may have been reused,
		 * but either way, we can delete the orphan item.
		 */
		if (ret == -ENOENT || inode->i_nlink) {
			if (!ret)
				iput(inode);
			trans = btrfs_start_transaction(root, 1);
			if (IS_ERR(trans)) {
				ret = PTR_ERR(trans);
				goto out;
			}
			btrfs_debug(fs_info, "auto deleting %Lu",
				    found_key.objectid);
			ret = btrfs_del_orphan_item(trans, root,
						    found_key.objectid);
			btrfs_end_transaction(trans);
			if (ret)
				goto out;
			continue;
		}

		nr_unlink++;

		/* this will do delete_inode and everything for us */
		iput(inode);
		if (ret)
			goto out;
	}
	/* release the path since we're done with it */
	btrfs_release_path(path);

	root->orphan_cleanup_state = ORPHAN_CLEANUP_DONE;

	if (test_bit(BTRFS_ROOT_ORPHAN_ITEM_INSERTED, &root->state)) {
		trans = btrfs_join_transaction(root);
		if (!IS_ERR(trans))
			btrfs_end_transaction(trans);
	}

	if (nr_unlink)
		btrfs_debug(fs_info, "unlinked %d orphans", nr_unlink);

out:
	if (ret)
		btrfs_err(fs_info, "could not do orphan cleanup %d", ret);
	btrfs_free_path(path);
	return ret;
}

/*
 * very simple check to peek ahead in the leaf looking for xattrs.  If we
 * don't find any xattrs, we know there can't be any acls.
 *
 * slot is the slot the inode is in, objectid is the objectid of the inode
 */
static noinline int acls_after_inode_item(struct extent_buffer *leaf,
					  int slot, u64 objectid,
					  int *first_xattr_slot)
{
	u32 nritems = btrfs_header_nritems(leaf);
	struct btrfs_key found_key;
	static u64 xattr_access = 0;
	static u64 xattr_default = 0;
	int scanned = 0;

	if (!xattr_access) {
		xattr_access = btrfs_name_hash(XATTR_NAME_POSIX_ACL_ACCESS,
					strlen(XATTR_NAME_POSIX_ACL_ACCESS));
		xattr_default = btrfs_name_hash(XATTR_NAME_POSIX_ACL_DEFAULT,
					strlen(XATTR_NAME_POSIX_ACL_DEFAULT));
	}

	slot++;
	*first_xattr_slot = -1;
	while (slot < nritems) {
		btrfs_item_key_to_cpu(leaf, &found_key, slot);

		/* we found a different objectid, there must not be acls */
		if (found_key.objectid != objectid)
			return 0;

		/* we found an xattr, assume we've got an acl */
		if (found_key.type == BTRFS_XATTR_ITEM_KEY) {
			if (*first_xattr_slot == -1)
				*first_xattr_slot = slot;
			if (found_key.offset == xattr_access ||
			    found_key.offset == xattr_default)
				return 1;
		}

		/*
		 * we found a key greater than an xattr key, there can't
		 * be any acls later on
		 */
		if (found_key.type > BTRFS_XATTR_ITEM_KEY)
			return 0;

		slot++;
		scanned++;

		/*
		 * it goes inode, inode backrefs, xattrs, extents,
		 * so if there are a ton of hard links to an inode there can
		 * be a lot of backrefs.  Don't waste time searching too hard,
		 * this is just an optimization
		 */
		if (scanned >= 8)
			break;
	}
	/* we hit the end of the leaf before we found an xattr or
	 * something larger than an xattr.  We have to assume the inode
	 * has acls
	 */
	if (*first_xattr_slot == -1)
		*first_xattr_slot = slot;
	return 1;
}

/*
 * read an inode from the btree into the in-memory inode
 */
static int btrfs_read_locked_inode(struct inode *inode)
{
	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
	struct btrfs_path *path;
	struct extent_buffer *leaf;
	struct btrfs_inode_item *inode_item;
	struct btrfs_root *root = BTRFS_I(inode)->root;
	struct btrfs_key location;
	unsigned long ptr;
	int maybe_acls;
	u32 rdev;
	int ret;
	bool filled = false;
	int first_xattr_slot;

	ret = btrfs_fill_inode(inode, &rdev);
	if (!ret)
		filled = true;

	path = btrfs_alloc_path();
	if (!path) {
		ret = -ENOMEM;
		goto make_bad;
	}

	memcpy(&location, &BTRFS_I(inode)->location, sizeof(location));

	ret = btrfs_lookup_inode(NULL, root, path, &location, 0);
	if (ret) {
		if (ret > 0)
			ret = -ENOENT;
		goto make_bad;
	}

	leaf = path->nodes[0];

	if (filled)
		goto cache_index;

	inode_item = btrfs_item_ptr(leaf, path->slots[0],
				    struct btrfs_inode_item);
	inode->i_mode = btrfs_inode_mode(leaf, inode_item);
	set_nlink(inode, btrfs_inode_nlink(leaf, inode_item));
	i_uid_write(inode, btrfs_inode_uid(leaf, inode_item));
	i_gid_write(inode, btrfs_inode_gid(leaf, inode_item));
	btrfs_i_size_write(BTRFS_I(inode), btrfs_inode_size(leaf, inode_item));

	inode->i_atime.tv_sec = btrfs_timespec_sec(leaf, &inode_item->atime);
	inode->i_atime.tv_nsec = btrfs_timespec_nsec(leaf, &inode_item->atime);

	inode->i_mtime.tv_sec = btrfs_timespec_sec(leaf, &inode_item->mtime);
	inode->i_mtime.tv_nsec = btrfs_timespec_nsec(leaf, &inode_item->mtime);

	inode->i_ctime.tv_sec = btrfs_timespec_sec(leaf, &inode_item->ctime);
	inode->i_ctime.tv_nsec = btrfs_timespec_nsec(leaf, &inode_item->ctime);

	BTRFS_I(inode)->i_otime.tv_sec =
		btrfs_timespec_sec(leaf, &inode_item->otime);
	BTRFS_I(inode)->i_otime.tv_nsec =
		btrfs_timespec_nsec(leaf, &inode_item->otime);

	inode_set_bytes(inode, btrfs_inode_nbytes(leaf, inode_item));
	BTRFS_I(inode)->generation = btrfs_inode_generation(leaf, inode_item);
	BTRFS_I(inode)->last_trans = btrfs_inode_transid(leaf, inode_item);

	inode_set_iversion_queried(inode,
				   btrfs_inode_sequence(leaf, inode_item));
	inode->i_generation = BTRFS_I(inode)->generation;
	inode->i_rdev = 0;
	rdev = btrfs_inode_rdev(leaf, inode_item);

	BTRFS_I(inode)->index_cnt = (u64)-1;
	BTRFS_I(inode)->flags = btrfs_inode_flags(leaf, inode_item);

cache_index:
	/*
	 * If we were modified in the current generation and evicted from memory
	 * and then re-read we need to do a full sync since we don't have any
	 * idea about which extents were modified before we were evicted from
	 * cache.
	 *
	 * This is required for both inode re-read from disk and delayed inode
	 * in delayed_nodes_tree.
	 */
	if (BTRFS_I(inode)->last_trans == fs_info->generation)
		set_bit(BTRFS_INODE_NEEDS_FULL_SYNC,
			&BTRFS_I(inode)->runtime_flags);

	/*
	 * We don't persist the id of the transaction where an unlink operation
	 * against the inode was last made. So here we assume the inode might
	 * have been evicted, and therefore the exact value of last_unlink_trans
	 * lost, and set it to last_trans to avoid metadata inconsistencies
	 * between the inode and its parent if the inode is fsync'ed and the log
	 * replayed. For example, in the scenario:
	 *
	 * touch mydir/foo
	 * ln mydir/foo mydir/bar
	 * sync
	 * unlink mydir/bar
	 * echo 2 > /proc/sys/vm/drop_caches   # evicts inode
	 * xfs_io -c fsync mydir/foo
	 * <power failure>
	 * mount fs, triggers fsync log replay
	 *
	 * We must make sure that when we fsync our inode foo we also log its
	 * parent inode, otherwise after log replay the parent still has the
	 * dentry with the "bar" name but our inode foo has a link count of 1
	 * and doesn't have an inode ref with the name "bar" anymore.
	 *
	 * Setting last_unlink_trans to last_trans is a pessimistic approach,
	 * but it guarantees correctness at the expense of occasional full
	 * transaction commits on fsync if our inode is a directory, or if our
	 * inode is not a directory, logging its parent unnecessarily.
	 */
	BTRFS_I(inode)->last_unlink_trans = BTRFS_I(inode)->last_trans;

	path->slots[0]++;
	if (inode->i_nlink != 1 ||
	    path->slots[0] >= btrfs_header_nritems(leaf))
		goto cache_acl;

	btrfs_item_key_to_cpu(leaf, &location, path->slots[0]);
	if (location.objectid != btrfs_ino(BTRFS_I(inode)))
		goto cache_acl;

	ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
	if (location.type == BTRFS_INODE_REF_KEY) {
		struct btrfs_inode_ref *ref;

		ref = (struct btrfs_inode_ref *)ptr;
		BTRFS_I(inode)->dir_index = btrfs_inode_ref_index(leaf, ref);
	} else if (location.type == BTRFS_INODE_EXTREF_KEY) {
		struct btrfs_inode_extref *extref;

		extref = (struct btrfs_inode_extref *)ptr;
		BTRFS_I(inode)->dir_index = btrfs_inode_extref_index(leaf,
								     extref);
	}
cache_acl:
	/*
	 * try to precache a NULL acl entry for files that don't have
	 * any xattrs or acls
	 */
	maybe_acls = acls_after_inode_item(leaf, path->slots[0],
			btrfs_ino(BTRFS_I(inode)), &first_xattr_slot);
	if (first_xattr_slot != -1) {
		path->slots[0] = first_xattr_slot;
		ret = btrfs_load_inode_props(inode, path);
		if (ret)
			btrfs_err(fs_info,
				  "error loading props for ino %llu (root %llu): %d",
				  btrfs_ino(BTRFS_I(inode)),
				  root->root_key.objectid, ret);
	}
	btrfs_free_path(path);

	if (!maybe_acls)
		cache_no_acl(inode);

	switch (inode->i_mode & S_IFMT) {
	case S_IFREG:
		inode->i_mapping->a_ops = &btrfs_aops;
		BTRFS_I(inode)->io_tree.ops = &btrfs_extent_io_ops;
		inode->i_fop = &btrfs_file_operations;
		inode->i_op = &btrfs_file_inode_operations;
		break;
	case S_IFDIR:
		inode->i_fop = &btrfs_dir_file_operations;
		inode->i_op = &btrfs_dir_inode_operations;
		break;
	case S_IFLNK:
		inode->i_op = &btrfs_symlink_inode_operations;
		inode_nohighmem(inode);
		inode->i_mapping->a_ops = &btrfs_symlink_aops;
		break;
	default:
		inode->i_op = &btrfs_special_inode_operations;
		init_special_inode(inode, inode->i_mode, rdev);
		break;
	}

	btrfs_sync_inode_flags_to_i_flags(inode);
	return 0;

make_bad:
	btrfs_free_path(path);
	make_bad_inode(inode);
	return ret;
}

/*
 * given a leaf and an inode, copy the inode fields into the leaf
 */
static void fill_inode_item(struct btrfs_trans_handle *trans,
			    struct extent_buffer *leaf,
			    struct btrfs_inode_item *item,
			    struct inode *inode)
{
	struct btrfs_map_token token;

	btrfs_init_map_token(&token);

	btrfs_set_token_inode_uid(leaf, item, i_uid_read(inode), &token);
	btrfs_set_token_inode_gid(leaf, item, i_gid_read(inode), &token);
	btrfs_set_token_inode_size(leaf, item, BTRFS_I(inode)->disk_i_size,
				   &token);
	btrfs_set_token_inode_mode(leaf, item, inode->i_mode, &token);
	btrfs_set_token_inode_nlink(leaf, item, inode->i_nlink, &token);

	btrfs_set_token_timespec_sec(leaf, &item->atime,
				     inode->i_atime.tv_sec, &token);
	btrfs_set_token_timespec_nsec(leaf, &item->atime,
				      inode->i_atime.tv_nsec, &token);

	btrfs_set_token_timespec_sec(leaf, &item->mtime,
				     inode->i_mtime.tv_sec, &token);
	btrfs_set_token_timespec_nsec(leaf, &item->mtime,
				      inode->i_mtime.tv_nsec, &token);

	btrfs_set_token_timespec_sec(leaf, &item->ctime,
				     inode->i_ctime.tv_sec, &token);
	btrfs_set_token_timespec_nsec(leaf, &item->ctime,
				      inode->i_ctime.tv_nsec, &token);

	btrfs_set_token_timespec_sec(leaf, &item->otime,
				     BTRFS_I(inode)->i_otime.tv_sec, &token);
	btrfs_set_token_timespec_nsec(leaf, &item->otime,
				      BTRFS_I(inode)->i_otime.tv_nsec, &token);

	btrfs_set_token_inode_nbytes(leaf, item, inode_get_bytes(inode),
				     &token);
	btrfs_set_token_inode_generation(leaf, item, BTRFS_I(inode)->generation,
					 &token);
	btrfs_set_token_inode_sequence(leaf, item, inode_peek_iversion(inode),
				       &token);
	btrfs_set_token_inode_transid(leaf, item, trans->transid, &token);
	btrfs_set_token_inode_rdev(leaf, item, inode->i_rdev, &token);
	btrfs_set_token_inode_flags(leaf, item, BTRFS_I(inode)->flags, &token);
	btrfs_set_token_inode_block_group(leaf, item, 0, &token);
}

/*
 * copy everything in the in-memory inode into the btree.
 */
static noinline int btrfs_update_inode_item(struct btrfs_trans_handle *trans,
				struct btrfs_root *root, struct inode *inode)
{
	struct btrfs_inode_item *inode_item;
	struct btrfs_path *path;
	struct extent_buffer *leaf;
	int ret;

	path = btrfs_alloc_path();
	if (!path)
		return -ENOMEM;

	path->leave_spinning = 1;
	ret = btrfs_lookup_inode(trans, root, path, &BTRFS_I(inode)->location,
				 1);
	if (ret) {
		if (ret > 0)
			ret = -ENOENT;
		goto failed;
	}

	leaf = path->nodes[0];
	inode_item = btrfs_item_ptr(leaf, path->slots[0],
				    struct btrfs_inode_item);

	fill_inode_item(trans, leaf, inode_item, inode);
	btrfs_mark_buffer_dirty(leaf);
	btrfs_set_inode_last_trans(trans, inode);
	ret = 0;
failed:
	btrfs_free_path(path);
	return ret;
}

/*
 * copy everything in the in-memory inode into the btree.
 */
noinline int btrfs_update_inode(struct btrfs_trans_handle *trans,
				struct btrfs_root *root, struct inode *inode)
{
	struct btrfs_fs_info *fs_info = root->fs_info;
	int ret;

	/*
	 * If the inode is a free space inode, we can deadlock during commit
	 * if we put it into the delayed code.
	 *
	 * The data relocation inode should also be directly updated
	 * without delay
	 */
	if (!btrfs_is_free_space_inode(BTRFS_I(inode))
	    && root->root_key.objectid != BTRFS_DATA_RELOC_TREE_OBJECTID
	    && !test_bit(BTRFS_FS_LOG_RECOVERING, &fs_info->flags)) {
		btrfs_update_root_times(trans, root);

		ret = btrfs_delayed_update_inode(trans, root, inode);
		if (!ret)
			btrfs_set_inode_last_trans(trans, inode);
		return ret;
	}

	return btrfs_update_inode_item(trans, root, inode);
}

noinline int btrfs_update_inode_fallback(struct btrfs_trans_handle *trans,
					 struct btrfs_root *root,
					 struct inode *inode)
{
	int ret;

	ret = btrfs_update_inode(trans, root, inode);
	if (ret == -ENOSPC)
		return btrfs_update_inode_item(trans, root, inode);
	return ret;
}

/*
 * unlink helper that gets used here in inode.c and in the tree logging
 * recovery code.  It remove a link in a directory with a given name, and
 * also drops the back refs in the inode to the directory
 */
static int __btrfs_unlink_inode(struct btrfs_trans_handle *trans,
				struct btrfs_root *root,
				struct btrfs_inode *dir,
				struct btrfs_inode *inode,
				const char *name, int name_len)
{
	struct btrfs_fs_info *fs_info = root->fs_info;
	struct btrfs_path *path;
	int ret = 0;
	struct extent_buffer *leaf;
	struct btrfs_dir_item *di;
	struct btrfs_key key;
	u64 index;
	u64 ino = btrfs_ino(inode);
	u64 dir_ino = btrfs_ino(dir);

	path = btrfs_alloc_path();
	if (!path) {
		ret = -ENOMEM;
		goto out;
	}

	path->leave_spinning = 1;
	di = btrfs_lookup_dir_item(trans, root, path, dir_ino,
				    name, name_len, -1);
	if (IS_ERR(di)) {
		ret = PTR_ERR(di);
		goto err;
	}
	if (!di) {
		ret = -ENOENT;
		goto err;
	}
	leaf = path->nodes[0];
	btrfs_dir_item_key_to_cpu(leaf, di, &key);
	ret = btrfs_delete_one_dir_name(trans, root, path, di);
	if (ret)
		goto err;
	btrfs_release_path(path);

	/*
	 * If we don't have dir index, we have to get it by looking up
	 * the inode ref, since we get the inode ref, remove it directly,
	 * it is unnecessary to do delayed deletion.
	 *
	 * But if we have dir index, needn't search inode ref to get it.
	 * Since the inode ref is close to the inode item, it is better
	 * that we delay to delete it, and just do this deletion when
	 * we update the inode item.
	 */
	if (inode->dir_index) {
		ret = btrfs_delayed_delete_inode_ref(inode);
		if (!ret) {
			index = inode->dir_index;
			goto skip_backref;
		}
	}

	ret = btrfs_del_inode_ref(trans, root, name, name_len, ino,
				  dir_ino, &index);
	if (ret) {
		btrfs_info(fs_info,
			"failed to delete reference to %.*s, inode %llu parent %llu",
			name_len, name, ino, dir_ino);
		btrfs_abort_transaction(trans, ret);
		goto err;
	}
skip_backref:
	ret = btrfs_delete_delayed_dir_index(trans, fs_info, dir, index);
	if (ret) {
		btrfs_abort_transaction(trans, ret);
		goto err;
	}

	ret = btrfs_del_inode_ref_in_log(trans, root, name, name_len, inode,
			dir_ino);
	if (ret != 0 && ret != -ENOENT) {
		btrfs_abort_transaction(trans, ret);
		goto err;
	}

	ret = btrfs_del_dir_entries_in_log(trans, root, name, name_len, dir,
			index);
	if (ret == -ENOENT)
		ret = 0;
	else if (ret)
		btrfs_abort_transaction(trans, ret);
err:
	btrfs_free_path(path);
	if (ret)
		goto out;

	btrfs_i_size_write(dir, dir->vfs_inode.i_size - name_len * 2);
	inode_inc_iversion(&inode->vfs_inode);
	inode_inc_iversion(&dir->vfs_inode);
	inode->vfs_inode.i_ctime = dir->vfs_inode.i_mtime =
		dir->vfs_inode.i_ctime = current_time(&inode->vfs_inode);
	ret = btrfs_update_inode(trans, root, &dir->vfs_inode);
out:
	return ret;
}

int btrfs_unlink_inode(struct btrfs_trans_handle *trans,
		       struct btrfs_root *root,
		       struct btrfs_inode *dir, struct btrfs_inode *inode,
		       const char *name, int name_len)
{
	int ret;
	ret = __btrfs_unlink_inode(trans, root, dir, inode, name, name_len);
	if (!ret) {
		drop_nlink(&inode->vfs_inode);
		ret = btrfs_update_inode(trans, root, &inode->vfs_inode);
	}
	return ret;
}

/*
 * helper to start transaction for unlink and rmdir.
 *
 * unlink and rmdir are special in btrfs, they do not always free space, so
 * if we cannot make our reservations the normal way try and see if there is
 * plenty of slack room in the global reserve to migrate, otherwise we cannot
 * allow the unlink to occur.
 */
static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir)
{
	struct btrfs_root *root = BTRFS_I(dir)->root;

	/*
	 * 1 for the possible orphan item
	 * 1 for the dir item
	 * 1 for the dir index
	 * 1 for the inode ref
	 * 1 for the inode
	 */
	return btrfs_start_transaction_fallback_global_rsv(root, 5, 5);
}

static int btrfs_unlink(struct inode *dir, struct dentry *dentry)
{
	struct btrfs_root *root = BTRFS_I(dir)->root;
	struct btrfs_trans_handle *trans;
	struct inode *inode = d_inode(dentry);
	int ret;

	trans = __unlink_start_trans(dir);
	if (IS_ERR(trans))
		return PTR_ERR(trans);

	btrfs_record_unlink_dir(trans, BTRFS_I(dir), BTRFS_I(d_inode(dentry)),
			0);

	ret = btrfs_unlink_inode(trans, root, BTRFS_I(dir),
			BTRFS_I(d_inode(dentry)), dentry->d_name.name,
			dentry->d_name.len);
	if (ret)
		goto out;

	if (inode->i_nlink == 0) {
		ret = btrfs_orphan_add(trans, BTRFS_I(inode));
		if (ret)
			goto out;
	}

out:
	btrfs_end_transaction(trans);
	btrfs_btree_balance_dirty(root->fs_info);
	return ret;
}

static int btrfs_unlink_subvol(struct btrfs_trans_handle *trans,
			struct btrfs_root *root,
			struct inode *dir, u64 objectid,
			const char *name, int name_len)
{
	struct btrfs_fs_info *fs_info = root->fs_info;
	struct btrfs_path *path;
	struct extent_buffer *leaf;
	struct btrfs_dir_item *di;
	struct btrfs_key key;
	u64 index;
	int ret;
	u64 dir_ino = btrfs_ino(BTRFS_I(dir));

	path = btrfs_alloc_path();
	if (!path)
		return -ENOMEM;

	di = btrfs_lookup_dir_item(trans, root, path, dir_ino,
				   name, name_len, -1);
	if (IS_ERR_OR_NULL(di)) {
		if (!di)
			ret = -ENOENT;
		else
			ret = PTR_ERR(di);
		goto out;
	}

	leaf = path->nodes[0];
	btrfs_dir_item_key_to_cpu(leaf, di, &key);
	WARN_ON(key.type != BTRFS_ROOT_ITEM_KEY || key.objectid != objectid);
	ret = btrfs_delete_one_dir_name(trans, root, path, di);
	if (ret) {
		btrfs_abort_transaction(trans, ret);
		goto out;
	}
	btrfs_release_path(path);

	ret = btrfs_del_root_ref(trans, fs_info, objectid,
				 root->root_key.objectid, dir_ino,
				 &index, name, name_len);
	if (ret < 0) {
		if (ret != -ENOENT) {
			btrfs_abort_transaction(trans, ret);
			goto out;
		}
		di = btrfs_search_dir_index_item(root, path, dir_ino,
						 name, name_len);
		if (IS_ERR_OR_NULL(di)) {
			if (!di)
				ret = -ENOENT;
			else
				ret = PTR_ERR(di);
			btrfs_abort_transaction(trans, ret);
			goto out;
		}

		leaf = path->nodes[0];
		btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
		btrfs_release_path(path);
		index = key.offset;
	}
	btrfs_release_path(path);

	ret = btrfs_delete_delayed_dir_index(trans, fs_info, BTRFS_I(dir), index);
	if (ret) {
		btrfs_abort_transaction(trans, ret);
		goto out;
	}

	btrfs_i_size_write(BTRFS_I(dir), dir->i_size - name_len * 2);
	inode_inc_iversion(dir);
	dir->i_mtime = dir->i_ctime = current_time(dir);
	ret = btrfs_update_inode_fallback(trans, root, dir);
	if (ret)
		btrfs_abort_transaction(trans, ret);
out:
	btrfs_free_path(path);
	return ret;
}

/*
 * Helper to check if the subvolume references other subvolumes or if it's
 * default.
 */
static noinline int may_destroy_subvol(struct btrfs_root *root)
{
	struct btrfs_fs_info *fs_info = root->fs_info;
	struct btrfs_path *path;
	struct btrfs_dir_item *di;
	struct btrfs_key key;
	u64 dir_id;
	int ret;

	path = btrfs_alloc_path();
	if (!path)
		return -ENOMEM;

	/* Make sure this root isn't set as the default subvol */
	dir_id = btrfs_super_root_dir(fs_info->super_copy);
	di = btrfs_lookup_dir_item(NULL, fs_info->tree_root, path,
				   dir_id, "default", 7, 0);
	if (di && !IS_ERR(di)) {
		btrfs_dir_item_key_to_cpu(path->nodes[0], di, &key);
		if (key.objectid == root->root_key.objectid) {
			ret = -EPERM;
			btrfs_err(fs_info,
				  "deleting default subvolume %llu is not allowed",
				  key.objectid);
			goto out;
		}
		btrfs_release_path(path);
	}

	key.objectid = root->root_key.objectid;
	key.type = BTRFS_ROOT_REF_KEY;
	key.offset = (u64)-1;

	ret = btrfs_search_slot(NULL, fs_info->tree_root, &key, path, 0, 0);
	if (ret < 0)
		goto out;
	BUG_ON(ret == 0);

	ret = 0;
	if (path->slots[0] > 0) {
		path->slots[0]--;
		btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
		if (key.objectid == root->root_key.objectid &&
		    key.type == BTRFS_ROOT_REF_KEY)
			ret = -ENOTEMPTY;
	}
out:
	btrfs_free_path(path);
	return ret;
}

/* Delete all dentries for inodes belonging to the root */
static void btrfs_prune_dentries(struct btrfs_root *root)
{
	struct btrfs_fs_info *fs_info = root->fs_info;
	struct rb_node *node;
	struct rb_node *prev;
	struct btrfs_inode *entry;
	struct inode *inode;
	u64 objectid = 0;

	if (!test_bit(BTRFS_FS_STATE_ERROR, &fs_info->fs_state))
		WARN_ON(btrfs_root_refs(&root->root_item) != 0);

	spin_lock(&root->inode_lock);
again:
	node = root->inode_tree.rb_node;
	prev = NULL;
	while (node) {
		prev = node;
		entry = rb_entry(node, struct btrfs_inode, rb_node);

		if (objectid < btrfs_ino(BTRFS_I(&entry->vfs_inode)))
			node = node->rb_left;
		else if (objectid > btrfs_ino(BTRFS_I(&entry->vfs_inode)))
			node = node->rb_right;
		else
			break;
	}
	if (!node) {
		while (prev) {
			entry = rb_entry(prev, struct btrfs_inode, rb_node);
			if (objectid <= btrfs_ino(BTRFS_I(&entry->vfs_inode))) {
				node = prev;
				break;
			}
			prev = rb_next(prev);
		}
	}
	while (node) {
		entry = rb_entry(node, struct btrfs_inode, rb_node);
		objectid = btrfs_ino(BTRFS_I(&entry->vfs_inode)) + 1;
		inode = igrab(&entry->vfs_inode);
		if (inode) {
			spin_unlock(&root->inode_lock);
			if (atomic_read(&inode->i_count) > 1)
				d_prune_aliases(inode);
			/*
			 * btrfs_drop_inode will have it removed from the inode
			 * cache when its usage count hits zero.
			 */
			iput(inode);
			cond_resched();
			spin_lock(&root->inode_lock);
			goto again;
		}

		if (cond_resched_lock(&root->inode_lock))
			goto again;

		node = rb_next(node);
	}
	spin_unlock(&root->inode_lock);
}

int btrfs_delete_subvolume(struct inode *dir, struct dentry *dentry)
{
	struct btrfs_fs_info *fs_info = btrfs_sb(dentry->d_sb);
	struct btrfs_root *root = BTRFS_I(dir)->root;
	struct inode *inode = d_inode(dentry);
	struct btrfs_root *dest = BTRFS_I(inode)->root;
	struct btrfs_trans_handle *trans;
	struct btrfs_block_rsv block_rsv;
	u64 root_flags;
	int ret;
	int err;

	/*
	 * Don't allow to delete a subvolume with send in progress. This is
	 * inside the inode lock so the error handling that has to drop the bit
	 * again is not run concurrently.
	 */
	spin_lock(&dest->root_item_lock);
	root_flags = btrfs_root_flags(&dest->root_item);
	if (dest->send_in_progress == 0) {
		btrfs_set_root_flags(&dest->root_item,
				root_flags | BTRFS_ROOT_SUBVOL_DEAD);
		spin_unlock(&dest->root_item_lock);
	} else {
		spin_unlock(&dest->root_item_lock);
		btrfs_warn(fs_info,
			   "attempt to delete subvolume %llu during send",
			   dest->root_key.objectid);
		return -EPERM;
	}

	down_write(&fs_info->subvol_sem);

	err = may_destroy_subvol(dest);
	if (err)
		goto out_up_write;

	btrfs_init_block_rsv(&block_rsv, BTRFS_BLOCK_RSV_TEMP);
	/*
	 * One for dir inode,
	 * two for dir entries,
	 * two for root ref/backref.
	 */
	err = btrfs_subvolume_reserve_metadata(root, &block_rsv, 5, true);
	if (err)
		goto out_up_write;

	trans = btrfs_start_transaction(root, 0);
	if (IS_ERR(trans)) {
		err = PTR_ERR(trans);
		goto out_release;
	}
	trans->block_rsv = &block_rsv;
	trans->bytes_reserved = block_rsv.size;

	btrfs_record_snapshot_destroy(trans, BTRFS_I(dir));

	ret = btrfs_unlink_subvol(trans, root, dir,
				dest->root_key.objectid,
				dentry->d_name.name,
				dentry->d_name.len);
	if (ret) {
		err = ret;
		btrfs_abort_transaction(trans, ret);
		goto out_end_trans;
	}

	btrfs_record_root_in_trans(trans, dest);

	memset(&dest->root_item.drop_progress, 0,
		sizeof(dest->root_item.drop_progress));
	dest->root_item.drop_level = 0;
	btrfs_set_root_refs(&dest->root_item, 0);

	if (!test_and_set_bit(BTRFS_ROOT_ORPHAN_ITEM_INSERTED, &dest->state)) {
		ret = btrfs_insert_orphan_item(trans,
					fs_info->tree_root,
					dest->root_key.objectid);
		if (ret) {
			btrfs_abort_transaction(trans, ret);
			err = ret;
			goto out_end_trans;
		}
	}

	ret = btrfs_uuid_tree_remove(trans, dest->root_item.uuid,
				  BTRFS_UUID_KEY_SUBVOL,
				  dest->root_key.objectid);
	if (ret && ret != -ENOENT) {
		btrfs_abort_transaction(trans, ret);
		err = ret;
		goto out_end_trans;
	}
	if (!btrfs_is_empty_uuid(dest->root_item.received_uuid)) {
		ret = btrfs_uuid_tree_remove(trans,
					  dest->root_item.received_uuid,
					  BTRFS_UUID_KEY_RECEIVED_SUBVOL,
					  dest->root_key.objectid);
		if (ret && ret != -ENOENT) {
			btrfs_abort_transaction(trans, ret);
			err = ret;
			goto out_end_trans;
		}
	}

out_end_trans:
	trans->block_rsv = NULL;
	trans->bytes_reserved = 0;
	ret = btrfs_end_transaction(trans);
	if (ret && !err)
		err = ret;
	inode->i_flags |= S_DEAD;
out_release:
	btrfs_subvolume_release_metadata(fs_info, &block_rsv);
out_up_write:
	up_write(&fs_info->subvol_sem);
	if (err) {
		spin_lock(&dest->root_item_lock);
		root_flags = btrfs_root_flags(&dest->root_item);
		btrfs_set_root_flags(&dest->root_item,
				root_flags & ~BTRFS_ROOT_SUBVOL_DEAD);
		spin_unlock(&dest->root_item_lock);
	} else {
		d_invalidate(dentry);
		btrfs_prune_dentries(dest);
		ASSERT(dest->send_in_progress == 0);

		/* the last ref */
		if (dest->ino_cache_inode) {
			iput(dest->ino_cache_inode);
			dest->ino_cache_inode = NULL;
		}
	}

	return err;
}

static int btrfs_rmdir(struct inode *dir, struct dentry *dentry)
{
	struct inode *inode = d_inode(dentry);
	int err = 0;
	struct btrfs_root *root = BTRFS_I(dir)->root;
	struct btrfs_trans_handle *trans;
	u64 last_unlink_trans;

	if (inode->i_size > BTRFS_EMPTY_DIR_SIZE)
		return -ENOTEMPTY;
	if (btrfs_ino(BTRFS_I(inode)) == BTRFS_FIRST_FREE_OBJECTID)
		return btrfs_delete_subvolume(dir, dentry);

	trans = __unlink_start_trans(dir);
	if (IS_ERR(trans))
		return PTR_ERR(trans);

	if (unlikely(btrfs_ino(BTRFS_I(inode)) == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID)) {
		err = btrfs_unlink_subvol(trans, root, dir,
					  BTRFS_I(inode)->location.objectid,
					  dentry->d_name.name,
					  dentry->d_name.len);
		goto out;
	}

	err = btrfs_orphan_add(trans, BTRFS_I(inode));
	if (err)
		goto out;

	last_unlink_trans = BTRFS_I(inode)->last_unlink_trans;

	/* now the directory is empty */
	err = btrfs_unlink_inode(trans, root, BTRFS_I(dir),
			BTRFS_I(d_inode(dentry)), dentry->d_name.name,
			dentry->d_name.len);
	if (!err) {
		btrfs_i_size_write(BTRFS_I(inode), 0);
		/*
		 * Propagate the last_unlink_trans value of the deleted dir to
		 * its parent directory. This is to prevent an unrecoverable
		 * log tree in the case we do something like this:
		 * 1) create dir foo
		 * 2) create snapshot under dir foo
		 * 3) delete the snapshot
		 * 4) rmdir foo
		 * 5) mkdir foo
		 * 6) fsync foo or some file inside foo
		 */
		if (last_unlink_trans >= trans->transid)
			BTRFS_I(dir)->last_unlink_trans = last_unlink_trans;
	}
out:
	btrfs_end_transaction(trans);
	btrfs_btree_balance_dirty(root->fs_info);

	return err;
}

static int truncate_space_check(struct btrfs_trans_handle *trans,
				struct btrfs_root *root,
				u64 bytes_deleted)
{
	struct btrfs_fs_info *fs_info = root->fs_info;
	int ret;

	/*
	 * This is only used to apply pressure to the enospc system, we don't
	 * intend to use this reservation at all.
	 */
	bytes_deleted = btrfs_csum_bytes_to_leaves(fs_info, bytes_deleted);
	bytes_deleted *= fs_info->nodesize;
	ret = btrfs_block_rsv_add(root, &fs_info->trans_block_rsv,
				  bytes_deleted, BTRFS_RESERVE_NO_FLUSH);
	if (!ret) {
		trace_btrfs_space_reservation(fs_info, "transaction",
					      trans->transid,
					      bytes_deleted, 1);
		trans->bytes_reserved += bytes_deleted;
	}
	return ret;

}

/*
 * Return this if we need to call truncate_block for the last bit of the
 * truncate.
 */
#define NEED_TRUNCATE_BLOCK 1

/*
 * this can truncate away extent items, csum items and directory items.
 * It starts at a high offset and removes keys until it can't find
 * any higher than new_size
 *
 * csum items that cross the new i_size are truncated to the new size
 * as well.
 *
 * min_type is the minimum key type to truncate down to.  If set to 0, this
 * will kill all the items on this inode, including the INODE_ITEM_KEY.
 */
int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans,
			       struct btrfs_root *root,
			       struct inode *inode,
			       u64 new_size, u32 min_type)
{
	struct btrfs_fs_info *fs_info = root->fs_info;
	struct btrfs_path *path;
	struct extent_buffer *leaf;
	struct btrfs_file_extent_item *fi;
	struct btrfs_key key;
	struct btrfs_key found_key;
	u64 extent_start = 0;
	u64 extent_num_bytes = 0;
	u64 extent_offset = 0;
	u64 item_end = 0;
	u64 last_size = new_size;
	u32 found_type = (u8)-1;
	int found_extent;
	int del_item;
	int pending_del_nr = 0;
	int pending_del_slot = 0;
	int extent_type = -1;
	int ret;
	u64 ino = btrfs_ino(BTRFS_I(inode));
	u64 bytes_deleted = 0;
	bool be_nice = false;
	bool should_throttle = false;
	bool should_end = false;

	BUG_ON(new_size > 0 && min_type != BTRFS_EXTENT_DATA_KEY);

	/*
	 * for non-free space inodes and ref cows, we want to back off from
	 * time to time
	 */
	if (!btrfs_is_free_space_inode(BTRFS_I(inode)) &&
	    test_bit(BTRFS_ROOT_REF_COWS, &root->state))
		be_nice = true;

	path = btrfs_alloc_path();
	if (!path)
		return -ENOMEM;
	path->reada = READA_BACK;

	/*
	 * We want to drop from the next block forward in case this new size is
	 * not block aligned since we will be keeping the last block of the
	 * extent just the way it is.
	 */
	if (test_bit(BTRFS_ROOT_REF_COWS, &root->state) ||
	    root == fs_info->tree_root)
		btrfs_drop_extent_cache(BTRFS_I(inode), ALIGN(new_size,
					fs_info->sectorsize),
					(u64)-1, 0);

	/*
	 * This function is also used to drop the items in the log tree before
	 * we relog the inode, so if root != BTRFS_I(inode)->root, it means
	 * it is used to drop the loged items. So we shouldn't kill the delayed
	 * items.
	 */
	if (min_type == 0 && root == BTRFS_I(inode)->root)
		btrfs_kill_delayed_inode_items(BTRFS_I(inode));

	key.objectid = ino;
	key.offset = (u64)-1;
	key.type = (u8)-1;

search_again:
	/*
	 * with a 16K leaf size and 128MB extents, you can actually queue
	 * up a huge file in a single leaf.  Most of the time that
	 * bytes_deleted is > 0, it will be huge by the time we get here
	 */
	if (be_nice && bytes_deleted > SZ_32M &&
	    btrfs_should_end_transaction(trans)) {
		ret = -EAGAIN;
		goto out;
	}

	path->leave_spinning = 1;
	ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
	if (ret < 0)
		goto out;

	if (ret > 0) {
		ret = 0;
		/* there are no items in the tree for us to truncate, we're
		 * done
		 */
		if (path->slots[0] == 0)
			goto out;
		path->slots[0]--;
	}

	while (1) {
		fi = NULL;
		leaf = path->nodes[0];
		btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
		found_type = found_key.type;

		if (found_key.objectid != ino)
			break;

		if (found_type < min_type)
			break;

		item_end = found_key.offset;
		if (found_type == BTRFS_EXTENT_DATA_KEY) {
			fi = btrfs_item_ptr(leaf, path->slots[0],
					    struct btrfs_file_extent_item);
			extent_type = btrfs_file_extent_type(leaf, fi);
			if (extent_type != BTRFS_FILE_EXTENT_INLINE) {
				item_end +=
				    btrfs_file_extent_num_bytes(leaf, fi);

				trace_btrfs_truncate_show_fi_regular(
					BTRFS_I(inode), leaf, fi,
					found_key.offset);
			} else if (extent_type == BTRFS_FILE_EXTENT_INLINE) {
				item_end += btrfs_file_extent_inline_len(leaf,
							 path->slots[0], fi);

				trace_btrfs_truncate_show_fi_inline(
					BTRFS_I(inode), leaf, fi, path->slots[0],
					found_key.offset);
			}
			item_end--;
		}
		if (found_type > min_type) {
			del_item = 1;
		} else {
			if (item_end < new_size)
				break;
			if (found_key.offset >= new_size)
				del_item = 1;
			else
				del_item = 0;
		}
		found_extent = 0;
		/* FIXME, shrink the extent if the ref count is only 1 */
		if (found_type != BTRFS_EXTENT_DATA_KEY)
			goto delete;

		if (extent_type != BTRFS_FILE_EXTENT_INLINE) {
			u64 num_dec;
			extent_start = btrfs_file_extent_disk_bytenr(leaf, fi);
			if (!del_item) {
				u64 orig_num_bytes =
					btrfs_file_extent_num_bytes(leaf, fi);
				extent_num_bytes = ALIGN(new_size -
						found_key.offset,
						fs_info->sectorsize);
				btrfs_set_file_extent_num_bytes(leaf, fi,
							 extent_num_bytes);
				num_dec = (orig_num_bytes -
					   extent_num_bytes);
				if (test_bit(BTRFS_ROOT_REF_COWS,
					     &root->state) &&
				    extent_start != 0)
					inode_sub_bytes(inode, num_dec);
				btrfs_mark_buffer_dirty(leaf);
			} else {
				extent_num_bytes =
					btrfs_file_extent_disk_num_bytes(leaf,
									 fi);
				extent_offset = found_key.offset -
					btrfs_file_extent_offset(leaf, fi);

				/* FIXME blocksize != 4096 */
				num_dec = btrfs_file_extent_num_bytes(leaf, fi);
				if (extent_start != 0) {
					found_extent = 1;
					if (test_bit(BTRFS_ROOT_REF_COWS,
						     &root->state))
						inode_sub_bytes(inode, num_dec);
				}
			}
		} else if (extent_type == BTRFS_FILE_EXTENT_INLINE) {
			/*
			 * we can't truncate inline items that have had
			 * special encodings
			 */
			if (!del_item &&
			    btrfs_file_extent_encryption(leaf, fi) == 0 &&
			    btrfs_file_extent_other_encoding(leaf, fi) == 0 &&
			    btrfs_file_extent_compression(leaf, fi) == 0) {
				u32 size = (u32)(new_size - found_key.offset);

				btrfs_set_file_extent_ram_bytes(leaf, fi, size);
				size = btrfs_file_extent_calc_inline_size(size);
				btrfs_truncate_item(root->fs_info, path, size, 1);
			} else if (!del_item) {
				/*
				 * We have to bail so the last_size is set to
				 * just before this extent.
				 */
				ret = NEED_TRUNCATE_BLOCK;
				break;
			}

			if (test_bit(BTRFS_ROOT_REF_COWS, &root->state))
				inode_sub_bytes(inode, item_end + 1 - new_size);
		}
delete:
		if (del_item)
			last_size = found_key.offset;
		else
			last_size = new_size;
		if (del_item) {
			if (!pending_del_nr) {
				/* no pending yet, add ourselves */
				pending_del_slot = path->slots[0];
				pending_del_nr = 1;
			} else if (pending_del_nr &&
				   path->slots[0] + 1 == pending_del_slot) {
				/* hop on the pending chunk */
				pending_del_nr++;
				pending_del_slot = path->slots[0];
			} else {
				BUG();
			}
		} else {
			break;
		}
		should_throttle = false;

		if (found_extent &&
		    (test_bit(BTRFS_ROOT_REF_COWS, &root->state) ||
		     root == fs_info->tree_root)) {
			btrfs_set_path_blocking(path);
			bytes_deleted += extent_num_bytes;
			ret = btrfs_free_extent(trans, root, extent_start,
						extent_num_bytes, 0,
						btrfs_header_owner(leaf),
						ino, extent_offset);
			if (ret) {
				btrfs_abort_transaction(trans, ret);
				break;
			}
			if (btrfs_should_throttle_delayed_refs(trans, fs_info))
				btrfs_async_run_delayed_refs(fs_info,
					trans->delayed_ref_updates * 2,
					trans->transid, 0);
			if (be_nice) {
				if (truncate_space_check(trans, root,
							 extent_num_bytes)) {
					should_end = true;
				}
				if (btrfs_should_throttle_delayed_refs(trans,
								       fs_info))
					should_throttle = true;
			}
		}

		if (found_type == BTRFS_INODE_ITEM_KEY)
			break;

		if (path->slots[0] == 0 ||
		    path->slots[0] != pending_del_slot ||
		    should_throttle || should_end) {
			if (pending_del_nr) {
				ret = btrfs_del_items(trans, root, path,
						pending_del_slot,
						pending_del_nr);
				if (ret) {
					btrfs_abort_transaction(trans, ret);
					break;
				}
				pending_del_nr = 0;
			}
			btrfs_release_path(path);
			if (should_throttle) {
				unsigned long updates = trans->delayed_ref_updates;
				if (updates) {
					trans->delayed_ref_updates = 0;
					ret = btrfs_run_delayed_refs(trans,
								   updates * 2);
					if (ret)
						break;
				}
			}
			/*
			 * if we failed to refill our space rsv, bail out
			 * and let the transaction restart
			 */
			if (should_end) {
				ret = -EAGAIN;
				break;
			}
			goto search_again;
		} else {
			path->slots[0]--;
		}
	}
out:
	if (ret >= 0 && pending_del_nr) {
		int err;

		err = btrfs_del_items(trans, root, path, pending_del_slot,
				      pending_del_nr);
		if (err) {
			btrfs_abort_transaction(trans, err);
			ret = err;
		}
	}
	if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID) {
		ASSERT(last_size >= new_size);
		if (!ret && last_size > new_size)
			last_size = new_size;
		btrfs_ordered_update_i_size(inode, last_size, NULL);
	}

	btrfs_free_path(path);

	if (be_nice && bytes_deleted > SZ_32M && (ret >= 0 || ret == -EAGAIN)) {
		unsigned long updates = trans->delayed_ref_updates;
		int err;

		if (updates) {
			trans->delayed_ref_updates = 0;
			err = btrfs_run_delayed_refs(trans, updates * 2);
			if (err)
				ret = err;
		}
	}
	return ret;
}

/*
 * btrfs_truncate_block - read, zero a chunk and write a block
 * @inode - inode that we're zeroing
 * @from - the offset to start zeroing
 * @len - the length to zero, 0 to zero the entire range respective to the
 *	offset
 * @front - zero up to the offset instead of from the offset on
 *
 * This will find the block for the "from" offset and cow the block and zero the
 * part we want to zero.  This is used with truncate and hole punching.
 */
int btrfs_truncate_block(struct inode *inode, loff_t from, loff_t len,
			int front)
{
	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
	struct address_space *mapping = inode->i_mapping;
	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
	struct btrfs_ordered_extent *ordered;
	struct extent_state *cached_state = NULL;
	struct extent_changeset *data_reserved = NULL;
	char *kaddr;
	u32 blocksize = fs_info->sectorsize;
	pgoff_t index = from >> PAGE_SHIFT;
	unsigned offset = from & (blocksize - 1);
	struct page *page;
	gfp_t mask = btrfs_alloc_write_mask(mapping);
	int ret = 0;
	u64 block_start;
	u64 block_end;

	if (IS_ALIGNED(offset, blocksize) &&
	    (!len || IS_ALIGNED(len, blocksize)))
		goto out;

	block_start = round_down(from, blocksize);
	block_end = block_start + blocksize - 1;

	ret = btrfs_delalloc_reserve_space(inode, &data_reserved,
					   block_start, blocksize);
	if (ret)
		goto out;

again:
	page = find_or_create_page(mapping, index, mask);
	if (!page) {
		btrfs_delalloc_release_space(inode, data_reserved,
					     block_start, blocksize, true);
		btrfs_delalloc_release_extents(BTRFS_I(inode), blocksize, true);
		ret = -ENOMEM;
		goto out;
	}

	if (!PageUptodate(page)) {
		ret = btrfs_readpage(NULL, page);
		lock_page(page);
		if (page->mapping != mapping) {
			unlock_page(page);
			put_page(page);
			goto again;
		}
		if (!PageUptodate(page)) {
			ret = -EIO;
			goto out_unlock;
		}
	}
	wait_on_page_writeback(page);

	lock_extent_bits(io_tree, block_start, block_end, &cached_state);
	set_page_extent_mapped(page);

	ordered = btrfs_lookup_ordered_extent(inode, block_start);
	if (ordered) {
		unlock_extent_cached(io_tree, block_start, block_end,
				     &cached_state);
		unlock_page(page);
		put_page(page);
		btrfs_start_ordered_extent(inode, ordered, 1);
		btrfs_put_ordered_extent(ordered);
		goto again;
	}

	clear_extent_bit(&BTRFS_I(inode)->io_tree, block_start, block_end,
			  EXTENT_DIRTY | EXTENT_DELALLOC |
			  EXTENT_DO_ACCOUNTING | EXTENT_DEFRAG,
			  0, 0, &cached_state);

	ret = btrfs_set_extent_delalloc(inode, block_start, block_end, 0,
					&cached_state, 0);
	if (ret) {
		unlock_extent_cached(io_tree, block_start, block_end,
				     &cached_state);
		goto out_unlock;
	}

	if (offset != blocksize) {
		if (!len)
			len = blocksize - offset;
		kaddr = kmap(page);
		if (front)
			memset(kaddr + (block_start - page_offset(page)),
				0, offset);
		else
			memset(kaddr + (block_start - page_offset(page)) +  offset,
				0, len);
		flush_dcache_page(page);
		kunmap(page);
	}
	ClearPageChecked(page);
	set_page_dirty(page);
	unlock_extent_cached(io_tree, block_start, block_end, &cached_state);

out_unlock:
	if (ret)
		btrfs_delalloc_release_space(inode, data_reserved, block_start,
					     blocksize, true);
	btrfs_delalloc_release_extents(BTRFS_I(inode), blocksize, (ret != 0));
	unlock_page(page);
	put_page(page);
out:
	extent_changeset_free(data_reserved);
	return ret;
}

static int maybe_insert_hole(struct btrfs_root *root, struct inode *inode,
			     u64 offset, u64 len)
{
	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
	struct btrfs_trans_handle *trans;
	int ret;

	/*
	 * Still need to make sure the inode looks like it's been updated so
	 * that any holes get logged if we fsync.
	 */
	if (btrfs_fs_incompat(fs_info, NO_HOLES)) {
		BTRFS_I(inode)->last_trans = fs_info->generation;
		BTRFS_I(inode)->last_sub_trans = root->log_transid;
		BTRFS_I(inode)->last_log_commit = root->last_log_commit;
		return 0;
	}

	/*
	 * 1 - for the one we're dropping
	 * 1 - for the one we're adding
	 * 1 - for updating the inode.
	 */
	trans = btrfs_start_transaction(root, 3);
	if (IS_ERR(trans))
		return PTR_ERR(trans);

	ret = btrfs_drop_extents(trans, root, inode, offset, offset + len, 1);
	if (ret) {
		btrfs_abort_transaction(trans, ret);
		btrfs_end_transaction(trans);
		return ret;
	}

	ret = btrfs_insert_file_extent(trans, root, btrfs_ino(BTRFS_I(inode)),
			offset, 0, 0, len, 0, len, 0, 0, 0);
	if (ret)
		btrfs_abort_transaction(trans, ret);
	else
		btrfs_update_inode(trans, root, inode);
	btrfs_end_transaction(trans);
	return ret;
}

/*
 * This function puts in dummy file extents for the area we're creating a hole
 * for.  So if we are truncating this file to a larger size we need to insert
 * these file extents so that btrfs_get_extent will return a EXTENT_MAP_HOLE for
 * the range between oldsize and size
 */
int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size)
{
	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
	struct btrfs_root *root = BTRFS_I(inode)->root;
	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
	struct extent_map *em = NULL;
	struct extent_state *cached_state = NULL;
	struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
	u64 hole_start = ALIGN(oldsize, fs_info->sectorsize);
	u64 block_end = ALIGN(size, fs_info->sectorsize);
	u64 last_byte;
	u64 cur_offset;
	u64 hole_size;
	int err = 0;

	/*
	 * If our size started in the middle of a block we need to zero out the
	 * rest of the block before we expand the i_size, otherwise we could
	 * expose stale data.
	 */
	err = btrfs_truncate_block(inode, oldsize, 0, 0);
	if (err)
		return err;

	if (size <= hole_start)
		return 0;

	while (1) {
		struct btrfs_ordered_extent *ordered;

		lock_extent_bits(io_tree, hole_start, block_end - 1,
				 &cached_state);
		ordered = btrfs_lookup_ordered_range(BTRFS_I(inode), hole_start,
						     block_end - hole_start);
		if (!ordered)
			break;
		unlock_extent_cached(io_tree, hole_start, block_end - 1,
				     &cached_state);
		btrfs_start_ordered_extent(inode, ordered, 1);
		btrfs_put_ordered_extent(ordered);
	}

	cur_offset = hole_start;
	while (1) {
		em = btrfs_get_extent(BTRFS_I(inode), NULL, 0, cur_offset,
				block_end - cur_offset, 0);
		if (IS_ERR(em)) {
			err = PTR_ERR(em);
			em = NULL;
			break;
		}
		last_byte = min(extent_map_end(em), block_end);
		last_byte = ALIGN(last_byte, fs_info->sectorsize);
		if (!test_bit(EXTENT_FLAG_PREALLOC, &em->flags)) {
			struct extent_map *hole_em;
			hole_size = last_byte - cur_offset;

			err = maybe_insert_hole(root, inode, cur_offset,
						hole_size);
			if (err)
				break;
			btrfs_drop_extent_cache(BTRFS_I(inode), cur_offset,
						cur_offset + hole_size - 1, 0);
			hole_em = alloc_extent_map();
			if (!hole_em) {
				set_bit(BTRFS_INODE_NEEDS_FULL_SYNC,
					&BTRFS_I(inode)->runtime_flags);
				goto next;
			}
			hole_em->start = cur_offset;
			hole_em->len = hole_size;
			hole_em->orig_start = cur_offset;

			hole_em->block_start = EXTENT_MAP_HOLE;
			hole_em->block_len = 0;
			hole_em->orig_block_len = 0;
			hole_em->ram_bytes = hole_size;
			hole_em->bdev = fs_info->fs_devices->latest_bdev;
			hole_em->compress_type = BTRFS_COMPRESS_NONE;
			hole_em->generation = fs_info->generation;

			while (1) {
				write_lock(&em_tree->lock);
				err = add_extent_mapping(em_tree, hole_em, 1);
				write_unlock(&em_tree->lock);
				if (err != -EEXIST)
					break;
				btrfs_drop_extent_cache(BTRFS_I(inode),
							cur_offset,
							cur_offset +
							hole_size - 1, 0);
			}
			free_extent_map(hole_em);
		}
next:
		free_extent_map(em);
		em = NULL;
		cur_offset = last_byte;
		if (cur_offset >= block_end)
			break;
	}
	free_extent_map(em);
	unlock_extent_cached(io_tree, hole_start, block_end - 1, &cached_state);
	return err;
}

static int btrfs_setsize(struct inode *inode, struct iattr *attr)
{
	struct btrfs_root *root = BTRFS_I(inode)->root;
	struct btrfs_trans_handle *trans;
	loff_t oldsize = i_size_read(inode);
	loff_t newsize = attr->ia_size;
	int mask = attr->ia_valid;
	int ret;

	/*
	 * The regular truncate() case without ATTR_CTIME and ATTR_MTIME is a
	 * special case where we need to update the times despite not having
	 * these flags set.  For all other operations the VFS set these flags
	 * explicitly if it wants a timestamp update.
	 */
	if (newsize != oldsize) {
		inode_inc_iversion(inode);
		if (!(mask & (ATTR_CTIME | ATTR_MTIME)))
			inode->i_ctime = inode->i_mtime =
				current_time(inode);
	}

	if (newsize > oldsize) {
		/*
		 * Don't do an expanding truncate while snapshotting is ongoing.
		 * This is to ensure the snapshot captures a fully consistent
		 * state of this file - if the snapshot captures this expanding
		 * truncation, it must capture all writes that happened before
		 * this truncation.
		 */
		btrfs_wait_for_snapshot_creation(root);
		ret = btrfs_cont_expand(inode, oldsize, newsize);
		if (ret) {
			btrfs_end_write_no_snapshotting(root);
			return ret;
		}

		trans = btrfs_start_transaction(root, 1);
		if (IS_ERR(trans)) {
			btrfs_end_write_no_snapshotting(root);
			return PTR_ERR(trans);
		}

		i_size_write(inode, newsize);
		btrfs_ordered_update_i_size(inode, i_size_read(inode), NULL);
		pagecache_isize_extended(inode, oldsize, newsize);
		ret = btrfs_update_inode(trans, root, inode);
		btrfs_end_write_no_snapshotting(root);
		btrfs_end_transaction(trans);
	} else {

		/*
		 * We're truncating a file that used to have good data down to
		 * zero. Make sure it gets into the ordered flush list so that
		 * any new writes get down to disk quickly.
		 */
		if (newsize == 0)
			set_bit(BTRFS_INODE_ORDERED_DATA_CLOSE,
				&BTRFS_I(inode)->runtime_flags);

		truncate_setsize(inode, newsize);

		/* Disable nonlocked read DIO to avoid the end less truncate */
		btrfs_inode_block_unlocked_dio(BTRFS_I(inode));
		inode_dio_wait(inode);
		btrfs_inode_resume_unlocked_dio(BTRFS_I(inode));

		ret = btrfs_truncate(inode, newsize == oldsize);
		if (ret && inode->i_nlink) {
			int err;

			/*
			 * Truncate failed, so fix up the in-memory size. We
			 * adjusted disk_i_size down as we removed extents, so
			 * wait for disk_i_size to be stable and then update the
			 * in-memory size to match.
			 */
			err = btrfs_wait_ordered_range(inode, 0, (u64)-1);
			if (err)
				return err;
			i_size_write(inode, BTRFS_I(inode)->disk_i_size);
		}
	}

	return ret;
}

static int btrfs_setattr(struct dentry *dentry, struct iattr *attr)
{
	struct inode *inode = d_inode(dentry);
	struct btrfs_root *root = BTRFS_I(inode)->root;
	int err;

	if (btrfs_root_readonly(root))
		return -EROFS;

	err = setattr_prepare(dentry, attr);
	if (err)
		return err;

	if (S_ISREG(inode->i_mode) && (attr->ia_valid & ATTR_SIZE)) {
		err = btrfs_setsize(inode, attr);
		if (err)
			return err;
	}

	if (attr->ia_valid) {
		setattr_copy(inode, attr);
		inode_inc_iversion(inode);
		err = btrfs_dirty_inode(inode);

		if (!err && attr->ia_valid & ATTR_MODE)
			err = posix_acl_chmod(inode, inode->i_mode);
	}

	return err;
}

/*
 * While truncating the inode pages during eviction, we get the VFS calling
 * btrfs_invalidatepage() against each page of the inode. This is slow because
 * the calls to btrfs_invalidatepage() result in a huge amount of calls to
 * lock_extent_bits() and clear_extent_bit(), which keep merging and splitting
 * extent_state structures over and over, wasting lots of time.
 *
 * Therefore if the inode is being evicted, let btrfs_invalidatepage() skip all
 * those expensive operations on a per page basis and do only the ordered io
 * finishing, while we release here the extent_map and extent_state structures,
 * without the excessive merging and splitting.
 */
static void evict_inode_truncate_pages(struct inode *inode)
{
	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
	struct extent_map_tree *map_tree = &BTRFS_I(inode)->extent_tree;
	struct rb_node *node;

	ASSERT(inode->i_state & I_FREEING);
	truncate_inode_pages_final(&inode->i_data);

	write_lock(&map_tree->lock);
	while (!RB_EMPTY_ROOT(&map_tree->map)) {
		struct extent_map *em;

		node = rb_first(&map_tree->map);
		em = rb_entry(node, struct extent_map, rb_node);
		clear_bit(EXTENT_FLAG_PINNED, &em->flags);
		clear_bit(EXTENT_FLAG_LOGGING, &em->flags);
		remove_extent_mapping(map_tree, em);
		free_extent_map(em);
		if (need_resched()) {
			write_unlock(&map_tree->lock);
			cond_resched();
			write_lock(&map_tree->lock);
		}
	}
	write_unlock(&map_tree->lock);

	/*
	 * Keep looping until we have no more ranges in the io tree.
	 * We can have ongoing bios started by readpages (called from readahead)
	 * that have their endio callback (extent_io.c:end_bio_extent_readpage)
	 * still in progress (unlocked the pages in the bio but did not yet
	 * unlocked the ranges in the io tree). Therefore this means some
	 * ranges can still be locked and eviction started because before
	 * submitting those bios, which are executed by a separate task (work
	 * queue kthread), inode references (inode->i_count) were not taken
	 * (which would be dropped in the end io callback of each bio).
	 * Therefore here we effectively end up waiting for those bios and
	 * anyone else holding locked ranges without having bumped the inode's
	 * reference count - if we don't do it, when they access the inode's
	 * io_tree to unlock a range it may be too late, leading to an
	 * use-after-free issue.
	 */
	spin_lock(&io_tree->lock);
	while (!RB_EMPTY_ROOT(&io_tree->state)) {
		struct extent_state *state;
		struct extent_state *cached_state = NULL;
		u64 start;
		u64 end;

		node = rb_first(&io_tree->state);
		state = rb_entry(node, struct extent_state, rb_node);
		start = state->start;
		end = state->end;
		spin_unlock(&io_tree->lock);

		lock_extent_bits(io_tree, start, end, &cached_state);

		/*
		 * If still has DELALLOC flag, the extent didn't reach disk,
		 * and its reserved space won't be freed by delayed_ref.
		 * So we need to free its reserved space here.
		 * (Refer to comment in btrfs_invalidatepage, case 2)
		 *
		 * Note, end is the bytenr of last byte, so we need + 1 here.
		 */
		if (state->state & EXTENT_DELALLOC)
			btrfs_qgroup_free_data(inode, NULL, start, end - start + 1);

		clear_extent_bit(io_tree, start, end,
				 EXTENT_LOCKED | EXTENT_DIRTY |
				 EXTENT_DELALLOC | EXTENT_DO_ACCOUNTING |
				 EXTENT_DEFRAG, 1, 1, &cached_state);

		cond_resched();
		spin_lock(&io_tree->lock);
	}
	spin_unlock(&io_tree->lock);
}

static struct btrfs_trans_handle *evict_refill_and_join(struct btrfs_root *root,
							struct btrfs_block_rsv *rsv,
							u64 min_size)
{
	struct btrfs_fs_info *fs_info = root->fs_info;
	struct btrfs_block_rsv *global_rsv = &fs_info->global_block_rsv;
	int failures = 0;

	for (;;) {
		struct btrfs_trans_handle *trans;
		int ret;

		ret = btrfs_block_rsv_refill(root, rsv, min_size,
					     BTRFS_RESERVE_FLUSH_LIMIT);

		if (ret && ++failures > 2) {
			btrfs_warn(fs_info,
				   "could not allocate space for a delete; will truncate on mount");
			return ERR_PTR(-ENOSPC);
		}

		trans = btrfs_join_transaction(root);
		if (IS_ERR(trans) || !ret)
			return trans;

		/*
		 * Try to steal from the global reserve if there is space for
		 * it.
		 */
		if (!btrfs_check_space_for_delayed_refs(trans, fs_info) &&
		    !btrfs_block_rsv_migrate(global_rsv, rsv, min_size, 0))
			return trans;

		/* If not, commit and try again. */
		ret = btrfs_commit_transaction(trans);
		if (ret)
			return ERR_PTR(ret);
	}
}

void btrfs_evict_inode(struct inode *inode)
{
	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
	struct btrfs_trans_handle *trans;
	struct btrfs_root *root = BTRFS_I(inode)->root;
	struct btrfs_block_rsv *rsv;
	u64 min_size;
	int ret;

	trace_btrfs_inode_evict(inode);

	if (!root) {
		clear_inode(inode);
		return;
	}

	min_size = btrfs_calc_trunc_metadata_size(fs_info, 1);

	evict_inode_truncate_pages(inode);

	if (inode->i_nlink &&
	    ((btrfs_root_refs(&root->root_item) != 0 &&
	      root->root_key.objectid != BTRFS_ROOT_TREE_OBJECTID) ||
	     btrfs_is_free_space_inode(BTRFS_I(inode))))
		goto no_delete;

	if (is_bad_inode(inode))
		goto no_delete;
	/* do we really want it for ->i_nlink > 0 and zero btrfs_root_refs? */
	if (!special_file(inode->i_mode))
		btrfs_wait_ordered_range(inode, 0, (u64)-1);

	btrfs_free_io_failure_record(BTRFS_I(inode), 0, (u64)-1);

	if (test_bit(BTRFS_FS_LOG_RECOVERING, &fs_info->flags))
		goto no_delete;

	if (inode->i_nlink > 0) {
		BUG_ON(btrfs_root_refs(&root->root_item) != 0 &&
		       root->root_key.objectid != BTRFS_ROOT_TREE_OBJECTID);
		goto no_delete;
	}

	ret = btrfs_commit_inode_delayed_inode(BTRFS_I(inode));
	if (ret)
		goto no_delete;

	rsv = btrfs_alloc_block_rsv(fs_info, BTRFS_BLOCK_RSV_TEMP);
	if (!rsv)
		goto no_delete;
	rsv->size = min_size;
	rsv->failfast = 1;

	btrfs_i_size_write(BTRFS_I(inode), 0);

	while (1) {
		trans = evict_refill_and_join(root, rsv, min_size);
		if (IS_ERR(trans))
			goto free_rsv;

		trans->block_rsv = rsv;

		ret = btrfs_truncate_inode_items(trans, root, inode, 0, 0);
		trans->block_rsv = &fs_info->trans_block_rsv;
		btrfs_end_transaction(trans);
		btrfs_btree_balance_dirty(fs_info);
		if (ret && ret != -ENOSPC && ret != -EAGAIN)
			goto free_rsv;
		else if (!ret)
			break;
	}

	/*
	 * Errors here aren't a big deal, it just means we leave orphan items in
	 * the tree. They will be cleaned up on the next mount. If the inode
	 * number gets reused, cleanup deletes the orphan item without doing
	 * anything, and unlink reuses the existing orphan item.
	 *
	 * If it turns out that we are dropping too many of these, we might want
	 * to add a mechanism for retrying these after a commit.
	 */
	trans = evict_refill_and_join(root, rsv, min_size);
	if (!IS_ERR(trans)) {
		trans->block_rsv = rsv;
		btrfs_orphan_del(trans, BTRFS_I(inode));
		trans->block_rsv = &fs_info->trans_block_rsv;
		btrfs_end_transaction(trans);
	}

	if (!(root == fs_info->tree_root ||
	      root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID))
		btrfs_return_ino(root, btrfs_ino(BTRFS_I(inode)));

free_rsv:
	btrfs_free_block_rsv(fs_info, rsv);
no_delete:
	/*
	 * If we didn't successfully delete, the orphan item will still be in
	 * the tree and we'll retry on the next mount. Again, we might also want
	 * to retry these periodically in the future.
	 */
	btrfs_remove_delayed_node(BTRFS_I(inode));
	clear_inode(inode);
}

/*
 * this returns the key found in the dir entry in the location pointer.
 * If no dir entries were found, returns -ENOENT.
 * If found a corrupted location in dir entry, returns -EUCLEAN.
 */
static int btrfs_inode_by_name(struct inode *dir, struct dentry *dentry,
			       struct btrfs_key *location)
{
	const char *name = dentry->d_name.name;
	int namelen = dentry->d_name.len;
	struct btrfs_dir_item *di;
	struct btrfs_path *path;
	struct btrfs_root *root = BTRFS_I(dir)->root;
	int ret = 0;

	path = btrfs_alloc_path();
	if (!path)
		return -ENOMEM;

	di = btrfs_lookup_dir_item(NULL, root, path, btrfs_ino(BTRFS_I(dir)),
			name, namelen, 0);
	if (!di) {
		ret = -ENOENT;
		goto out;
	}
	if (IS_ERR(di)) {
		ret = PTR_ERR(di);
		goto out;
	}

	btrfs_dir_item_key_to_cpu(path->nodes[0], di, location);
	if (location->type != BTRFS_INODE_ITEM_KEY &&
	    location->type != BTRFS_ROOT_ITEM_KEY) {
		ret = -EUCLEAN;
		btrfs_warn(root->fs_info,
"%s gets something invalid in DIR_ITEM (name %s, directory ino %llu, location(%llu %u %llu))",
			   __func__, name, btrfs_ino(BTRFS_I(dir)),
			   location->objectid, location->type, location->offset);
	}
out:
	btrfs_free_path(path);
	return ret;
}

/*
 * when we hit a tree root in a directory, the btrfs part of the inode
 * needs to be changed to reflect the root directory of the tree root.  This
 * is kind of like crossing a mount point.
 */
static int fixup_tree_root_location(struct btrfs_fs_info *fs_info,
				    struct inode *dir,
				    struct dentry *dentry,
				    struct btrfs_key *location,
				    struct btrfs_root **sub_root)
{
	struct btrfs_path *path;
	struct btrfs_root *new_root;
	struct btrfs_root_ref *ref;
	struct extent_buffer *leaf;
	struct btrfs_key key;
	int ret;
	int err = 0;

	path = btrfs_alloc_path();
	if (!path) {
		err = -ENOMEM;
		goto out;
	}

	err = -ENOENT;
	key.objectid = BTRFS_I(dir)->root->root_key.objectid;
	key.type = BTRFS_ROOT_REF_KEY;
	key.offset = location->objectid;

	ret = btrfs_search_slot(NULL, fs_info->tree_root, &key, path, 0, 0);
	if (ret) {
		if (ret < 0)
			err = ret;
		goto out;
	}

	leaf = path->nodes[0];
	ref = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_root_ref);
	if (btrfs_root_ref_dirid(leaf, ref) != btrfs_ino(BTRFS_I(dir)) ||
	    btrfs_root_ref_name_len(leaf, ref) != dentry->d_name.len)
		goto out;

	ret = memcmp_extent_buffer(leaf, dentry->d_name.name,
				   (unsigned long)(ref + 1),
				   dentry->d_name.len);
	if (ret)
		goto out;

	btrfs_release_path(path);

	new_root = btrfs_read_fs_root_no_name(fs_info, location);
	if (IS_ERR(new_root)) {
		err = PTR_ERR(new_root);
		goto out;
	}

	*sub_root = new_root;
	location->objectid = btrfs_root_dirid(&new_root->root_item);
	location->type = BTRFS_INODE_ITEM_KEY;
	location->offset = 0;
	err = 0;
out:
	btrfs_free_path(path);
	return err;
}

static void inode_tree_add(struct inode *inode)
{
	struct btrfs_root *root = BTRFS_I(inode)->root;
	struct btrfs_inode *entry;
	struct rb_node **p;
	struct rb_node *parent;
	struct rb_node *new = &BTRFS_I(inode)->rb_node;
	u64 ino = btrfs_ino(BTRFS_I(inode));

	if (inode_unhashed(inode))
		return;
	parent = NULL;
	spin_lock(&root->inode_lock);
	p = &root->inode_tree.rb_node;
	while (*p) {
		parent = *p;
		entry = rb_entry(parent, struct btrfs_inode, rb_node);

		if (ino < btrfs_ino(BTRFS_I(&entry->vfs_inode)))
			p = &parent->rb_left;
		else if (ino > btrfs_ino(BTRFS_I(&entry->vfs_inode)))
			p = &parent->rb_right;
		else {
			WARN_ON(!(entry->vfs_inode.i_state &
				  (I_WILL_FREE | I_FREEING)));
			rb_replace_node(parent, new, &root->inode_tree);
			RB_CLEAR_NODE(parent);
			spin_unlock(&root->inode_lock);
			return;
		}
	}
	rb_link_node(new, parent, p);
	rb_insert_color(new, &root->inode_tree);
	spin_unlock(&root->inode_lock);
}

static void inode_tree_del(struct inode *inode)
{
	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
	struct btrfs_root *root = BTRFS_I(inode)->root;
	int empty = 0;

	spin_lock(&root->inode_lock);
	if (!RB_EMPTY_NODE(&BTRFS_I(inode)->rb_node)) {
		rb_erase(&BTRFS_I(inode)->rb_node, &root->inode_tree);
		RB_CLEAR_NODE(&BTRFS_I(inode)->rb_node);
		empty = RB_EMPTY_ROOT(&root->inode_tree);
	}
	spin_unlock(&root->inode_lock);

	if (empty && btrfs_root_refs(&root->root_item) == 0) {
		synchronize_srcu(&fs_info->subvol_srcu);
		spin_lock(&root->inode_lock);
		empty =