| Author | Tokens | Token Proportion | Commits | Commit Proportion |
|---|---|---|---|---|
| Christoph Hellwig | 8036 | 58.51% | 30 | 17.75% |
| David Chinner | 3494 | 25.44% | 51 | 30.18% |
| Darrick J. Wong | 1394 | 10.15% | 34 | 20.12% |
| Nathan Scott | 228 | 1.66% | 11 | 6.51% |
| Brian Foster | 128 | 0.93% | 8 | 4.73% |
| Zhang Tianci | 89 | 0.65% | 1 | 0.59% |
| Alex Elder | 87 | 0.63% | 1 | 0.59% |
| Barry Naujok | 54 | 0.39% | 2 | 1.18% |
| Glen Overby | 49 | 0.36% | 2 | 1.18% |
| Russell Cattelan | 37 | 0.27% | 2 | 1.18% |
| Eric Sandeen | 32 | 0.23% | 6 | 3.55% |
| Carlos Maiolino | 20 | 0.15% | 4 | 2.37% |
| Allison Henderson | 17 | 0.12% | 1 | 0.59% |
| Mark Tinguely | 11 | 0.08% | 2 | 1.18% |
| Qian Cai | 10 | 0.07% | 1 | 0.59% |
| Linus Torvalds (pre-git) | 10 | 0.07% | 3 | 1.78% |
| Shida Zhang | 9 | 0.07% | 1 | 0.59% |
| Yingping Lu | 7 | 0.05% | 1 | 0.59% |
| Lachlan McIlroy | 5 | 0.04% | 1 | 0.59% |
| Fabian Frederick | 4 | 0.03% | 1 | 0.59% |
| Matt Mackall | 4 | 0.03% | 1 | 0.59% |
| Yang Guang | 4 | 0.03% | 1 | 0.59% |
| Michael Hayes | 2 | 0.01% | 1 | 0.59% |
| Gustavo A. R. Silva | 2 | 0.01% | 1 | 0.59% |
| David S. Miller | 1 | 0.01% | 1 | 0.59% |
| Ingo Molnar | 1 | 0.01% | 1 | 0.59% |
| Total | 13735 | 169 |
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696697698699700701702703704705706707708709710711712713714715716717718719720721722723724725726727728729730731732733734735736737738739740741742743744745746747748749750751752753754755756757758759760761762763764765766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804805806807808809810811812813814815816817818819820821822823824825826827828829830831832833834835836837838839840841842843844845846847848849850851852853854855856857858859860861862863864865866867868869870871872873874875876877878879880881882883884885886887888889890891892893894895896897898899900901902903904905906907908909910911912913914915916917918919920921922923924925926927928929930931932933934935936937938939940941942943944945946947948949950951952953954955956957958959960961962963964965966967968969970971972973974975976977978979980981982983984985986987988989990991992993994995996997998999100010011002100310041005100610071008100910101011101210131014101510161017101810191020102110221023102410251026102710281029103010311032103310341035103610371038103910401041104210431044104510461047104810491050105110521053105410551056105710581059106010611062106310641065106610671068106910701071107210731074107510761077107810791080108110821083108410851086108710881089109010911092109310941095109610971098109911001101110211031104110511061107110811091110111111121113111411151116111711181119112011211122112311241125112611271128112911301131113211331134113511361137113811391140114111421143114411451146114711481149115011511152115311541155115611571158115911601161116211631164116511661167116811691170117111721173117411751176117711781179118011811182118311841185118611871188118911901191119211931194119511961197119811991200120112021203120412051206120712081209121012111212121312141215121612171218121912201221122212231224122512261227122812291230123112321233123412351236123712381239124012411242124312441245124612471248124912501251125212531254125512561257125812591260126112621263126412651266126712681269127012711272127312741275127612771278127912801281128212831284128512861287128812891290129112921293129412951296129712981299130013011302130313041305130613071308130913101311131213131314131513161317131813191320132113221323132413251326132713281329133013311332133313341335133613371338133913401341134213431344134513461347134813491350135113521353135413551356135713581359136013611362136313641365136613671368136913701371137213731374137513761377137813791380138113821383138413851386138713881389139013911392139313941395139613971398139914001401140214031404140514061407140814091410141114121413141414151416141714181419142014211422142314241425142614271428142914301431143214331434143514361437143814391440144114421443144414451446144714481449145014511452145314541455145614571458145914601461146214631464146514661467146814691470147114721473147414751476147714781479148014811482148314841485148614871488148914901491149214931494149514961497149814991500150115021503150415051506150715081509151015111512151315141515151615171518151915201521152215231524152515261527152815291530153115321533153415351536153715381539154015411542154315441545154615471548154915501551155215531554155515561557155815591560156115621563156415651566156715681569157015711572157315741575157615771578157915801581158215831584158515861587158815891590159115921593159415951596159715981599160016011602160316041605160616071608160916101611161216131614161516161617161816191620162116221623162416251626162716281629163016311632163316341635163616371638163916401641164216431644164516461647164816491650165116521653165416551656165716581659166016611662166316641665166616671668166916701671167216731674167516761677167816791680168116821683168416851686168716881689169016911692169316941695169616971698169917001701170217031704170517061707170817091710171117121713171417151716171717181719172017211722172317241725172617271728172917301731173217331734173517361737173817391740174117421743174417451746174717481749175017511752175317541755175617571758175917601761176217631764176517661767176817691770177117721773177417751776177717781779178017811782178317841785178617871788178917901791179217931794179517961797179817991800180118021803180418051806180718081809181018111812181318141815181618171818181918201821182218231824182518261827182818291830183118321833183418351836183718381839184018411842184318441845184618471848184918501851185218531854185518561857185818591860186118621863186418651866186718681869187018711872187318741875187618771878187918801881188218831884188518861887188818891890189118921893189418951896189718981899190019011902190319041905190619071908190919101911191219131914191519161917191819191920192119221923192419251926192719281929193019311932193319341935193619371938193919401941194219431944194519461947194819491950195119521953195419551956195719581959196019611962196319641965196619671968196919701971197219731974197519761977197819791980198119821983198419851986198719881989199019911992199319941995199619971998199920002001200220032004200520062007200820092010201120122013201420152016201720182019202020212022202320242025202620272028202920302031203220332034203520362037203820392040204120422043204420452046204720482049205020512052205320542055205620572058205920602061206220632064206520662067206820692070207120722073207420752076207720782079208020812082208320842085208620872088208920902091209220932094209520962097209820992100210121022103210421052106210721082109211021112112211321142115211621172118211921202121212221232124212521262127212821292130213121322133213421352136213721382139214021412142214321442145214621472148214921502151215221532154215521562157215821592160216121622163216421652166216721682169217021712172217321742175217621772178217921802181218221832184218521862187218821892190219121922193219421952196219721982199220022012202220322042205220622072208220922102211221222132214221522162217221822192220222122222223222422252226222722282229223022312232223322342235223622372238223922402241224222432244224522462247224822492250225122522253225422552256225722582259226022612262226322642265226622672268226922702271227222732274227522762277227822792280228122822283228422852286228722882289229022912292229322942295229622972298229923002301230223032304230523062307230823092310231123122313231423152316231723182319232023212322232323242325232623272328232923302331233223332334233523362337233823392340234123422343234423452346234723482349235023512352235323542355235623572358235923602361236223632364236523662367236823692370237123722373237423752376237723782379238023812382238323842385238623872388238923902391239223932394239523962397239823992400240124022403240424052406240724082409241024112412241324142415241624172418241924202421242224232424242524262427242824292430243124322433243424352436243724382439244024412442244324442445244624472448244924502451245224532454245524562457245824592460246124622463246424652466246724682469247024712472247324742475247624772478247924802481248224832484248524862487248824892490249124922493249424952496249724982499250025012502250325042505250625072508250925102511251225132514251525162517251825192520252125222523252425252526252725282529253025312532253325342535253625372538253925402541254225432544254525462547254825492550255125522553255425552556255725582559256025612562256325642565256625672568256925702571257225732574257525762577257825792580258125822583258425852586258725882589259025912592259325942595259625972598259926002601260226032604260526062607260826092610261126122613261426152616261726182619262026212622262326242625262626272628262926302631263226332634263526362637263826392640264126422643264426452646264726482649265026512652265326542655265626572658265926602661266226632664266526662667266826692670267126722673267426752676267726782679268026812682268326842685268626872688268926902691269226932694269526962697269826992700270127022703270427052706270727082709271027112712271327142715271627172718271927202721272227232724272527262727272827292730273127322733273427352736273727382739274027412742274327442745274627472748274927502751275227532754275527562757275827592760276127622763276427652766276727682769277027712772277327742775277627772778277927802781278227832784278527862787278827892790279127922793279427952796279727982799280028012802280328042805280628072808280928102811281228132814281528162817281828192820282128222823282428252826282728282829283028312832283328342835283628372838283928402841284228432844284528462847284828492850285128522853285428552856285728582859286028612862286328642865286628672868286928702871287228732874287528762877287828792880
// SPDX-License-Identifier: GPL-2.0 /* * Copyright (c) 2000-2005 Silicon Graphics, Inc. * Copyright (c) 2013 Red Hat, Inc. * All Rights Reserved. */ #include "xfs.h" #include "xfs_fs.h" #include "xfs_shared.h" #include "xfs_format.h" #include "xfs_log_format.h" #include "xfs_trans_resv.h" #include "xfs_bit.h" #include "xfs_mount.h" #include "xfs_inode.h" #include "xfs_dir2.h" #include "xfs_dir2_priv.h" #include "xfs_trans.h" #include "xfs_bmap.h" #include "xfs_attr_leaf.h" #include "xfs_error.h" #include "xfs_trace.h" #include "xfs_buf_item.h" #include "xfs_log.h" #include "xfs_errortag.h" #include "xfs_health.h" /* * xfs_da_btree.c * * Routines to implement directories as Btrees of hashed names. */ /*======================================================================== * Function prototypes for the kernel. *========================================================================*/ /* * Routines used for growing the Btree. */ STATIC int xfs_da3_root_split(xfs_da_state_t *state, xfs_da_state_blk_t *existing_root, xfs_da_state_blk_t *new_child); STATIC int xfs_da3_node_split(xfs_da_state_t *state, xfs_da_state_blk_t *existing_blk, xfs_da_state_blk_t *split_blk, xfs_da_state_blk_t *blk_to_add, int treelevel, int *result); STATIC void xfs_da3_node_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *node_blk_1, xfs_da_state_blk_t *node_blk_2); STATIC void xfs_da3_node_add(xfs_da_state_t *state, xfs_da_state_blk_t *old_node_blk, xfs_da_state_blk_t *new_node_blk); /* * Routines used for shrinking the Btree. */ STATIC int xfs_da3_root_join(xfs_da_state_t *state, xfs_da_state_blk_t *root_blk); STATIC int xfs_da3_node_toosmall(xfs_da_state_t *state, int *retval); STATIC void xfs_da3_node_remove(xfs_da_state_t *state, xfs_da_state_blk_t *drop_blk); STATIC void xfs_da3_node_unbalance(xfs_da_state_t *state, xfs_da_state_blk_t *src_node_blk, xfs_da_state_blk_t *dst_node_blk); /* * Utility routines. */ STATIC int xfs_da3_blk_unlink(xfs_da_state_t *state, xfs_da_state_blk_t *drop_blk, xfs_da_state_blk_t *save_blk); struct kmem_cache *xfs_da_state_cache; /* anchor for dir/attr state */ /* * Allocate a dir-state structure. * We don't put them on the stack since they're large. */ struct xfs_da_state * xfs_da_state_alloc( struct xfs_da_args *args) { struct xfs_da_state *state; state = kmem_cache_zalloc(xfs_da_state_cache, GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_NOFAIL); state->args = args; state->mp = args->dp->i_mount; return state; } /* * Kill the altpath contents of a da-state structure. */ STATIC void xfs_da_state_kill_altpath(xfs_da_state_t *state) { int i; for (i = 0; i < state->altpath.active; i++) state->altpath.blk[i].bp = NULL; state->altpath.active = 0; } /* * Free a da-state structure. */ void xfs_da_state_free(xfs_da_state_t *state) { xfs_da_state_kill_altpath(state); #ifdef DEBUG memset((char *)state, 0, sizeof(*state)); #endif /* DEBUG */ kmem_cache_free(xfs_da_state_cache, state); } void xfs_da_state_reset( struct xfs_da_state *state, struct xfs_da_args *args) { xfs_da_state_kill_altpath(state); memset(state, 0, sizeof(struct xfs_da_state)); state->args = args; state->mp = state->args->dp->i_mount; } static inline int xfs_dabuf_nfsb(struct xfs_mount *mp, int whichfork) { if (whichfork == XFS_DATA_FORK) return mp->m_dir_geo->fsbcount; return mp->m_attr_geo->fsbcount; } void xfs_da3_node_hdr_from_disk( struct xfs_mount *mp, struct xfs_da3_icnode_hdr *to, struct xfs_da_intnode *from) { if (xfs_has_crc(mp)) { struct xfs_da3_intnode *from3 = (struct xfs_da3_intnode *)from; to->forw = be32_to_cpu(from3->hdr.info.hdr.forw); to->back = be32_to_cpu(from3->hdr.info.hdr.back); to->magic = be16_to_cpu(from3->hdr.info.hdr.magic); to->count = be16_to_cpu(from3->hdr.__count); to->level = be16_to_cpu(from3->hdr.__level); to->btree = from3->__btree; ASSERT(to->magic == XFS_DA3_NODE_MAGIC); } else { to->forw = be32_to_cpu(from->hdr.info.forw); to->back = be32_to_cpu(from->hdr.info.back); to->magic = be16_to_cpu(from->hdr.info.magic); to->count = be16_to_cpu(from->hdr.__count); to->level = be16_to_cpu(from->hdr.__level); to->btree = from->__btree; ASSERT(to->magic == XFS_DA_NODE_MAGIC); } } void xfs_da3_node_hdr_to_disk( struct xfs_mount *mp, struct xfs_da_intnode *to, struct xfs_da3_icnode_hdr *from) { if (xfs_has_crc(mp)) { struct xfs_da3_intnode *to3 = (struct xfs_da3_intnode *)to; ASSERT(from->magic == XFS_DA3_NODE_MAGIC); to3->hdr.info.hdr.forw = cpu_to_be32(from->forw); to3->hdr.info.hdr.back = cpu_to_be32(from->back); to3->hdr.info.hdr.magic = cpu_to_be16(from->magic); to3->hdr.__count = cpu_to_be16(from->count); to3->hdr.__level = cpu_to_be16(from->level); } else { ASSERT(from->magic == XFS_DA_NODE_MAGIC); to->hdr.info.forw = cpu_to_be32(from->forw); to->hdr.info.back = cpu_to_be32(from->back); to->hdr.info.magic = cpu_to_be16(from->magic); to->hdr.__count = cpu_to_be16(from->count); to->hdr.__level = cpu_to_be16(from->level); } } /* * Verify an xfs_da3_blkinfo structure. Note that the da3 fields are only * accessible on v5 filesystems. This header format is common across da node, * attr leaf and dir leaf blocks. */ xfs_failaddr_t xfs_da3_blkinfo_verify( struct xfs_buf *bp, struct xfs_da3_blkinfo *hdr3) { struct xfs_mount *mp = bp->b_mount; struct xfs_da_blkinfo *hdr = &hdr3->hdr; if (!xfs_verify_magic16(bp, hdr->magic)) return __this_address; if (xfs_has_crc(mp)) { if (!uuid_equal(&hdr3->uuid, &mp->m_sb.sb_meta_uuid)) return __this_address; if (be64_to_cpu(hdr3->blkno) != xfs_buf_daddr(bp)) return __this_address; if (!xfs_log_check_lsn(mp, be64_to_cpu(hdr3->lsn))) return __this_address; } return NULL; } static xfs_failaddr_t xfs_da3_node_verify( struct xfs_buf *bp) { struct xfs_mount *mp = bp->b_mount; struct xfs_da_intnode *hdr = bp->b_addr; struct xfs_da3_icnode_hdr ichdr; xfs_failaddr_t fa; xfs_da3_node_hdr_from_disk(mp, &ichdr, hdr); fa = xfs_da3_blkinfo_verify(bp, bp->b_addr); if (fa) return fa; if (ichdr.level == 0) return __this_address; if (ichdr.level > XFS_DA_NODE_MAXDEPTH) return __this_address; if (ichdr.count == 0) return __this_address; /* * we don't know if the node is for and attribute or directory tree, * so only fail if the count is outside both bounds */ if (ichdr.count > mp->m_dir_geo->node_ents && ichdr.count > mp->m_attr_geo->node_ents) return __this_address; /* XXX: hash order check? */ return NULL; } xfs_failaddr_t xfs_da3_node_header_check( struct xfs_buf *bp, xfs_ino_t owner) { struct xfs_mount *mp = bp->b_mount; if (xfs_has_crc(mp)) { struct xfs_da3_blkinfo *hdr3 = bp->b_addr; if (hdr3->hdr.magic != cpu_to_be16(XFS_DA3_NODE_MAGIC)) return __this_address; if (be64_to_cpu(hdr3->owner) != owner) return __this_address; } return NULL; } xfs_failaddr_t xfs_da3_header_check( struct xfs_buf *bp, xfs_ino_t owner) { struct xfs_mount *mp = bp->b_mount; struct xfs_da_blkinfo *hdr = bp->b_addr; if (!xfs_has_crc(mp)) return NULL; switch (hdr->magic) { case cpu_to_be16(XFS_ATTR3_LEAF_MAGIC): return xfs_attr3_leaf_header_check(bp, owner); case cpu_to_be16(XFS_DA3_NODE_MAGIC): return xfs_da3_node_header_check(bp, owner); case cpu_to_be16(XFS_DIR3_LEAF1_MAGIC): case cpu_to_be16(XFS_DIR3_LEAFN_MAGIC): return xfs_dir3_leaf_header_check(bp, owner); } ASSERT(0); return NULL; } static void xfs_da3_node_write_verify( struct xfs_buf *bp) { struct xfs_mount *mp = bp->b_mount; struct xfs_buf_log_item *bip = bp->b_log_item; struct xfs_da3_node_hdr *hdr3 = bp->b_addr; xfs_failaddr_t fa; fa = xfs_da3_node_verify(bp); if (fa) { xfs_verifier_error(bp, -EFSCORRUPTED, fa); return; } if (!xfs_has_crc(mp)) return; if (bip) hdr3->info.lsn = cpu_to_be64(bip->bli_item.li_lsn); xfs_buf_update_cksum(bp, XFS_DA3_NODE_CRC_OFF); } /* * leaf/node format detection on trees is sketchy, so a node read can be done on * leaf level blocks when detection identifies the tree as a node format tree * incorrectly. In this case, we need to swap the verifier to match the correct * format of the block being read. */ static void xfs_da3_node_read_verify( struct xfs_buf *bp) { struct xfs_da_blkinfo *info = bp->b_addr; xfs_failaddr_t fa; switch (be16_to_cpu(info->magic)) { case XFS_DA3_NODE_MAGIC: if (!xfs_buf_verify_cksum(bp, XFS_DA3_NODE_CRC_OFF)) { xfs_verifier_error(bp, -EFSBADCRC, __this_address); break; } fallthrough; case XFS_DA_NODE_MAGIC: fa = xfs_da3_node_verify(bp); if (fa) xfs_verifier_error(bp, -EFSCORRUPTED, fa); return; case XFS_ATTR_LEAF_MAGIC: case XFS_ATTR3_LEAF_MAGIC: bp->b_ops = &xfs_attr3_leaf_buf_ops; bp->b_ops->verify_read(bp); return; case XFS_DIR2_LEAFN_MAGIC: case XFS_DIR3_LEAFN_MAGIC: bp->b_ops = &xfs_dir3_leafn_buf_ops; bp->b_ops->verify_read(bp); return; default: xfs_verifier_error(bp, -EFSCORRUPTED, __this_address); break; } } /* Verify the structure of a da3 block. */ static xfs_failaddr_t xfs_da3_node_verify_struct( struct xfs_buf *bp) { struct xfs_da_blkinfo *info = bp->b_addr; switch (be16_to_cpu(info->magic)) { case XFS_DA3_NODE_MAGIC: case XFS_DA_NODE_MAGIC: return xfs_da3_node_verify(bp); case XFS_ATTR_LEAF_MAGIC: case XFS_ATTR3_LEAF_MAGIC: bp->b_ops = &xfs_attr3_leaf_buf_ops; return bp->b_ops->verify_struct(bp); case XFS_DIR2_LEAFN_MAGIC: case XFS_DIR3_LEAFN_MAGIC: bp->b_ops = &xfs_dir3_leafn_buf_ops; return bp->b_ops->verify_struct(bp); default: return __this_address; } } const struct xfs_buf_ops xfs_da3_node_buf_ops = { .name = "xfs_da3_node", .magic16 = { cpu_to_be16(XFS_DA_NODE_MAGIC), cpu_to_be16(XFS_DA3_NODE_MAGIC) }, .verify_read = xfs_da3_node_read_verify, .verify_write = xfs_da3_node_write_verify, .verify_struct = xfs_da3_node_verify_struct, }; static int xfs_da3_node_set_type( struct xfs_trans *tp, struct xfs_inode *dp, int whichfork, struct xfs_buf *bp) { struct xfs_da_blkinfo *info = bp->b_addr; switch (be16_to_cpu(info->magic)) { case XFS_DA_NODE_MAGIC: case XFS_DA3_NODE_MAGIC: xfs_trans_buf_set_type(tp, bp, XFS_BLFT_DA_NODE_BUF); return 0; case XFS_ATTR_LEAF_MAGIC: case XFS_ATTR3_LEAF_MAGIC: xfs_trans_buf_set_type(tp, bp, XFS_BLFT_ATTR_LEAF_BUF); return 0; case XFS_DIR2_LEAFN_MAGIC: case XFS_DIR3_LEAFN_MAGIC: xfs_trans_buf_set_type(tp, bp, XFS_BLFT_DIR_LEAFN_BUF); return 0; default: XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, tp->t_mountp, info, sizeof(*info)); xfs_trans_brelse(tp, bp); xfs_dirattr_mark_sick(dp, whichfork); return -EFSCORRUPTED; } } int xfs_da3_node_read( struct xfs_trans *tp, struct xfs_inode *dp, xfs_dablk_t bno, struct xfs_buf **bpp, int whichfork) { int error; error = xfs_da_read_buf(tp, dp, bno, 0, bpp, whichfork, &xfs_da3_node_buf_ops); if (error || !*bpp || !tp) return error; return xfs_da3_node_set_type(tp, dp, whichfork, *bpp); } int xfs_da3_node_read_mapped( struct xfs_trans *tp, struct xfs_inode *dp, xfs_daddr_t mappedbno, struct xfs_buf **bpp, int whichfork) { struct xfs_mount *mp = dp->i_mount; int error; error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp, mappedbno, XFS_FSB_TO_BB(mp, xfs_dabuf_nfsb(mp, whichfork)), 0, bpp, &xfs_da3_node_buf_ops); if (xfs_metadata_is_sick(error)) xfs_dirattr_mark_sick(dp, whichfork); if (error || !*bpp) return error; if (whichfork == XFS_ATTR_FORK) xfs_buf_set_ref(*bpp, XFS_ATTR_BTREE_REF); else xfs_buf_set_ref(*bpp, XFS_DIR_BTREE_REF); if (!tp) return 0; return xfs_da3_node_set_type(tp, dp, whichfork, *bpp); } /* * Copy src directory/attr leaf/node buffer to the dst. * For v5 file systems make sure the right blkno is stamped in. */ void xfs_da_buf_copy( struct xfs_buf *dst, struct xfs_buf *src, size_t size) { struct xfs_da3_blkinfo *da3 = dst->b_addr; memcpy(dst->b_addr, src->b_addr, size); dst->b_ops = src->b_ops; xfs_trans_buf_copy_type(dst, src); if (xfs_has_crc(dst->b_mount)) da3->blkno = cpu_to_be64(xfs_buf_daddr(dst)); } /*======================================================================== * Routines used for growing the Btree. *========================================================================*/ /* * Create the initial contents of an intermediate node. */ int xfs_da3_node_create( struct xfs_da_args *args, xfs_dablk_t blkno, int level, struct xfs_buf **bpp, int whichfork) { struct xfs_da_intnode *node; struct xfs_trans *tp = args->trans; struct xfs_mount *mp = tp->t_mountp; struct xfs_da3_icnode_hdr ichdr = {0}; struct xfs_buf *bp; int error; struct xfs_inode *dp = args->dp; trace_xfs_da_node_create(args); ASSERT(level <= XFS_DA_NODE_MAXDEPTH); error = xfs_da_get_buf(tp, dp, blkno, &bp, whichfork); if (error) return error; bp->b_ops = &xfs_da3_node_buf_ops; xfs_trans_buf_set_type(tp, bp, XFS_BLFT_DA_NODE_BUF); node = bp->b_addr; if (xfs_has_crc(mp)) { struct xfs_da3_node_hdr *hdr3 = bp->b_addr; memset(hdr3, 0, sizeof(struct xfs_da3_node_hdr)); ichdr.magic = XFS_DA3_NODE_MAGIC; hdr3->info.blkno = cpu_to_be64(xfs_buf_daddr(bp)); hdr3->info.owner = cpu_to_be64(args->owner); uuid_copy(&hdr3->info.uuid, &mp->m_sb.sb_meta_uuid); } else { ichdr.magic = XFS_DA_NODE_MAGIC; } ichdr.level = level; xfs_da3_node_hdr_to_disk(dp->i_mount, node, &ichdr); xfs_trans_log_buf(tp, bp, XFS_DA_LOGRANGE(node, &node->hdr, args->geo->node_hdr_size)); *bpp = bp; return 0; } /* * Split a leaf node, rebalance, then possibly split * intermediate nodes, rebalance, etc. */ int /* error */ xfs_da3_split( struct xfs_da_state *state) { struct xfs_da_state_blk *oldblk; struct xfs_da_state_blk *newblk; struct xfs_da_state_blk *addblk; struct xfs_da_intnode *node; int max; int action = 0; int error; int i; trace_xfs_da_split(state->args); if (XFS_TEST_ERROR(false, state->mp, XFS_ERRTAG_DA_LEAF_SPLIT)) return -EIO; /* * Walk back up the tree splitting/inserting/adjusting as necessary. * If we need to insert and there isn't room, split the node, then * decide which fragment to insert the new block from below into. * Note that we may split the root this way, but we need more fixup. */ max = state->path.active - 1; ASSERT((max >= 0) && (max < XFS_DA_NODE_MAXDEPTH)); ASSERT(state->path.blk[max].magic == XFS_ATTR_LEAF_MAGIC || state->path.blk[max].magic == XFS_DIR2_LEAFN_MAGIC); addblk = &state->path.blk[max]; /* initial dummy value */ for (i = max; (i >= 0) && addblk; state->path.active--, i--) { oldblk = &state->path.blk[i]; newblk = &state->altpath.blk[i]; /* * If a leaf node then * Allocate a new leaf node, then rebalance across them. * else if an intermediate node then * We split on the last layer, must we split the node? */ switch (oldblk->magic) { case XFS_ATTR_LEAF_MAGIC: error = xfs_attr3_leaf_split(state, oldblk, newblk); if (error < 0) return error; /* GROT: attr is inconsistent */ if (!error) { addblk = newblk; break; } /* * Entry wouldn't fit, split the leaf again. The new * extrablk will be consumed by xfs_da3_node_split if * the node is split. */ state->extravalid = 1; if (state->inleaf) { state->extraafter = 0; /* before newblk */ trace_xfs_attr_leaf_split_before(state->args); error = xfs_attr3_leaf_split(state, oldblk, &state->extrablk); } else { state->extraafter = 1; /* after newblk */ trace_xfs_attr_leaf_split_after(state->args); error = xfs_attr3_leaf_split(state, newblk, &state->extrablk); } if (error == 1) return -ENOSPC; if (error) return error; /* GROT: attr inconsistent */ addblk = newblk; break; case XFS_DIR2_LEAFN_MAGIC: error = xfs_dir2_leafn_split(state, oldblk, newblk); if (error) return error; addblk = newblk; break; case XFS_DA_NODE_MAGIC: error = xfs_da3_node_split(state, oldblk, newblk, addblk, max - i, &action); addblk->bp = NULL; if (error) return error; /* GROT: dir is inconsistent */ /* * Record the newly split block for the next time thru? */ if (action) addblk = newblk; else addblk = NULL; break; } /* * Update the btree to show the new hashval for this child. */ xfs_da3_fixhashpath(state, &state->path); } if (!addblk) return 0; /* * xfs_da3_node_split() should have consumed any extra blocks we added * during a double leaf split in the attr fork. This is guaranteed as * we can't be here if the attr fork only has a single leaf block. */ ASSERT(state->extravalid == 0 || state->path.blk[max].magic == XFS_DIR2_LEAFN_MAGIC); /* * Split the root node. */ ASSERT(state->path.active == 0); oldblk = &state->path.blk[0]; error = xfs_da3_root_split(state, oldblk, addblk); if (error) goto out; /* * Update pointers to the node which used to be block 0 and just got * bumped because of the addition of a new root node. Note that the * original block 0 could be at any position in the list of blocks in * the tree. * * Note: the magic numbers and sibling pointers are in the same physical * place for both v2 and v3 headers (by design). Hence it doesn't matter * which version of the xfs_da_intnode structure we use here as the * result will be the same using either structure. */ node = oldblk->bp->b_addr; if (node->hdr.info.forw) { if (be32_to_cpu(node->hdr.info.forw) != addblk->blkno) { xfs_buf_mark_corrupt(oldblk->bp); xfs_da_mark_sick(state->args); error = -EFSCORRUPTED; goto out; } node = addblk->bp->b_addr; node->hdr.info.back = cpu_to_be32(oldblk->blkno); xfs_trans_log_buf(state->args->trans, addblk->bp, XFS_DA_LOGRANGE(node, &node->hdr.info, sizeof(node->hdr.info))); } node = oldblk->bp->b_addr; if (node->hdr.info.back) { if (be32_to_cpu(node->hdr.info.back) != addblk->blkno) { xfs_buf_mark_corrupt(oldblk->bp); xfs_da_mark_sick(state->args); error = -EFSCORRUPTED; goto out; } node = addblk->bp->b_addr; node->hdr.info.forw = cpu_to_be32(oldblk->blkno); xfs_trans_log_buf(state->args->trans, addblk->bp, XFS_DA_LOGRANGE(node, &node->hdr.info, sizeof(node->hdr.info))); } out: addblk->bp = NULL; return error; } /* * Split the root. We have to create a new root and point to the two * parts (the split old root) that we just created. Copy block zero to * the EOF, extending the inode in process. */ STATIC int /* error */ xfs_da3_root_split( struct xfs_da_state *state, struct xfs_da_state_blk *blk1, struct xfs_da_state_blk *blk2) { struct xfs_da_intnode *node; struct xfs_da_intnode *oldroot; struct xfs_da_node_entry *btree; struct xfs_da3_icnode_hdr nodehdr; struct xfs_da_args *args; struct xfs_buf *bp; struct xfs_inode *dp; struct xfs_trans *tp; struct xfs_dir2_leaf *leaf; xfs_dablk_t blkno; int level; int error; int size; trace_xfs_da_root_split(state->args); /* * Copy the existing (incorrect) block from the root node position * to a free space somewhere. */ args = state->args; error = xfs_da_grow_inode(args, &blkno); if (error) return error; dp = args->dp; tp = args->trans; error = xfs_da_get_buf(tp, dp, blkno, &bp, args->whichfork); if (error) return error; node = bp->b_addr; oldroot = blk1->bp->b_addr; if (oldroot->hdr.info.magic == cpu_to_be16(XFS_DA_NODE_MAGIC) || oldroot->hdr.info.magic == cpu_to_be16(XFS_DA3_NODE_MAGIC)) { struct xfs_da3_icnode_hdr icnodehdr; xfs_da3_node_hdr_from_disk(dp->i_mount, &icnodehdr, oldroot); btree = icnodehdr.btree; size = (int)((char *)&btree[icnodehdr.count] - (char *)oldroot); level = icnodehdr.level; } else { struct xfs_dir3_icleaf_hdr leafhdr; leaf = (xfs_dir2_leaf_t *)oldroot; xfs_dir2_leaf_hdr_from_disk(dp->i_mount, &leafhdr, leaf); ASSERT(leafhdr.magic == XFS_DIR2_LEAFN_MAGIC || leafhdr.magic == XFS_DIR3_LEAFN_MAGIC); size = (int)((char *)&leafhdr.ents[leafhdr.count] - (char *)leaf); level = 0; } /* * Copy old root to new buffer and log it. */ xfs_da_buf_copy(bp, blk1->bp, size); xfs_trans_log_buf(tp, bp, 0, size - 1); /* * Update blk1 to point to new buffer. */ blk1->bp = bp; blk1->blkno = blkno; /* * Set up the new root node. */ error = xfs_da3_node_create(args, (args->whichfork == XFS_DATA_FORK) ? args->geo->leafblk : 0, level + 1, &bp, args->whichfork); if (error) return error; node = bp->b_addr; xfs_da3_node_hdr_from_disk(dp->i_mount, &nodehdr, node); btree = nodehdr.btree; btree[0].hashval = cpu_to_be32(blk1->hashval); btree[0].before = cpu_to_be32(blk1->blkno); btree[1].hashval = cpu_to_be32(blk2->hashval); btree[1].before = cpu_to_be32(blk2->blkno); nodehdr.count = 2; xfs_da3_node_hdr_to_disk(dp->i_mount, node, &nodehdr); #ifdef DEBUG if (oldroot->hdr.info.magic == cpu_to_be16(XFS_DIR2_LEAFN_MAGIC) || oldroot->hdr.info.magic == cpu_to_be16(XFS_DIR3_LEAFN_MAGIC)) { ASSERT(blk1->blkno >= args->geo->leafblk && blk1->blkno < args->geo->freeblk); ASSERT(blk2->blkno >= args->geo->leafblk && blk2->blkno < args->geo->freeblk); } #endif /* Header is already logged by xfs_da_node_create */ xfs_trans_log_buf(tp, bp, XFS_DA_LOGRANGE(node, btree, sizeof(xfs_da_node_entry_t) * 2)); return 0; } /* * Split the node, rebalance, then add the new entry. */ STATIC int /* error */ xfs_da3_node_split( struct xfs_da_state *state, struct xfs_da_state_blk *oldblk, struct xfs_da_state_blk *newblk, struct xfs_da_state_blk *addblk, int treelevel, int *result) { struct xfs_da_intnode *node; struct xfs_da3_icnode_hdr nodehdr; xfs_dablk_t blkno; int newcount; int error; int useextra; struct xfs_inode *dp = state->args->dp; trace_xfs_da_node_split(state->args); node = oldblk->bp->b_addr; xfs_da3_node_hdr_from_disk(dp->i_mount, &nodehdr, node); /* * With V2 dirs the extra block is data or freespace. */ useextra = state->extravalid && state->args->whichfork == XFS_ATTR_FORK; newcount = 1 + useextra; /* * Do we have to split the node? */ if (nodehdr.count + newcount > state->args->geo->node_ents) { /* * Allocate a new node, add to the doubly linked chain of * nodes, then move some of our excess entries into it. */ error = xfs_da_grow_inode(state->args, &blkno); if (error) return error; /* GROT: dir is inconsistent */ error = xfs_da3_node_create(state->args, blkno, treelevel, &newblk->bp, state->args->whichfork); if (error) return error; /* GROT: dir is inconsistent */ newblk->blkno = blkno; newblk->magic = XFS_DA_NODE_MAGIC; xfs_da3_node_rebalance(state, oldblk, newblk); error = xfs_da3_blk_link(state, oldblk, newblk); if (error) return error; *result = 1; } else { *result = 0; } /* * Insert the new entry(s) into the correct block * (updating last hashval in the process). * * xfs_da3_node_add() inserts BEFORE the given index, * and as a result of using node_lookup_int() we always * point to a valid entry (not after one), but a split * operation always results in a new block whose hashvals * FOLLOW the current block. * * If we had double-split op below us, then add the extra block too. */ node = oldblk->bp->b_addr; xfs_da3_node_hdr_from_disk(dp->i_mount, &nodehdr, node); if (oldblk->index <= nodehdr.count) { oldblk->index++; xfs_da3_node_add(state, oldblk, addblk); if (useextra) { if (state->extraafter) oldblk->index++; xfs_da3_node_add(state, oldblk, &state->extrablk); state->extravalid = 0; } } else { newblk->index++; xfs_da3_node_add(state, newblk, addblk); if (useextra) { if (state->extraafter) newblk->index++; xfs_da3_node_add(state, newblk, &state->extrablk); state->extravalid = 0; } } return 0; } /* * Balance the btree elements between two intermediate nodes, * usually one full and one empty. * * NOTE: if blk2 is empty, then it will get the upper half of blk1. */ STATIC void xfs_da3_node_rebalance( struct xfs_da_state *state, struct xfs_da_state_blk *blk1, struct xfs_da_state_blk *blk2) { struct xfs_da_intnode *node1; struct xfs_da_intnode *node2; struct xfs_da_node_entry *btree1; struct xfs_da_node_entry *btree2; struct xfs_da_node_entry *btree_s; struct xfs_da_node_entry *btree_d; struct xfs_da3_icnode_hdr nodehdr1; struct xfs_da3_icnode_hdr nodehdr2; struct xfs_trans *tp; int count; int tmp; int swap = 0; struct xfs_inode *dp = state->args->dp; trace_xfs_da_node_rebalance(state->args); node1 = blk1->bp->b_addr; node2 = blk2->bp->b_addr; xfs_da3_node_hdr_from_disk(dp->i_mount, &nodehdr1, node1); xfs_da3_node_hdr_from_disk(dp->i_mount, &nodehdr2, node2); btree1 = nodehdr1.btree; btree2 = nodehdr2.btree; /* * Figure out how many entries need to move, and in which direction. * Swap the nodes around if that makes it simpler. */ if (nodehdr1.count > 0 && nodehdr2.count > 0 && ((be32_to_cpu(btree2[0].hashval) < be32_to_cpu(btree1[0].hashval)) || (be32_to_cpu(btree2[nodehdr2.count - 1].hashval) < be32_to_cpu(btree1[nodehdr1.count - 1].hashval)))) { swap(node1, node2); xfs_da3_node_hdr_from_disk(dp->i_mount, &nodehdr1, node1); xfs_da3_node_hdr_from_disk(dp->i_mount, &nodehdr2, node2); btree1 = nodehdr1.btree; btree2 = nodehdr2.btree; swap = 1; } count = (nodehdr1.count - nodehdr2.count) / 2; if (count == 0) return; tp = state->args->trans; /* * Two cases: high-to-low and low-to-high. */ if (count > 0) { /* * Move elements in node2 up to make a hole. */ tmp = nodehdr2.count; if (tmp > 0) { tmp *= (uint)sizeof(xfs_da_node_entry_t); btree_s = &btree2[0]; btree_d = &btree2[count]; memmove(btree_d, btree_s, tmp); } /* * Move the req'd B-tree elements from high in node1 to * low in node2. */ nodehdr2.count += count; tmp = count * (uint)sizeof(xfs_da_node_entry_t); btree_s = &btree1[nodehdr1.count - count]; btree_d = &btree2[0]; memcpy(btree_d, btree_s, tmp); nodehdr1.count -= count; } else { /* * Move the req'd B-tree elements from low in node2 to * high in node1. */ count = -count; tmp = count * (uint)sizeof(xfs_da_node_entry_t); btree_s = &btree2[0]; btree_d = &btree1[nodehdr1.count]; memcpy(btree_d, btree_s, tmp); nodehdr1.count += count; xfs_trans_log_buf(tp, blk1->bp, XFS_DA_LOGRANGE(node1, btree_d, tmp)); /* * Move elements in node2 down to fill the hole. */ tmp = nodehdr2.count - count; tmp *= (uint)sizeof(xfs_da_node_entry_t); btree_s = &btree2[count]; btree_d = &btree2[0]; memmove(btree_d, btree_s, tmp); nodehdr2.count -= count; } /* * Log header of node 1 and all current bits of node 2. */ xfs_da3_node_hdr_to_disk(dp->i_mount, node1, &nodehdr1); xfs_trans_log_buf(tp, blk1->bp, XFS_DA_LOGRANGE(node1, &node1->hdr, state->args->geo->node_hdr_size)); xfs_da3_node_hdr_to_disk(dp->i_mount, node2, &nodehdr2); xfs_trans_log_buf(tp, blk2->bp, XFS_DA_LOGRANGE(node2, &node2->hdr, state->args->geo->node_hdr_size + (sizeof(btree2[0]) * nodehdr2.count))); /* * Record the last hashval from each block for upward propagation. * (note: don't use the swapped node pointers) */ if (swap) { node1 = blk1->bp->b_addr; node2 = blk2->bp->b_addr; xfs_da3_node_hdr_from_disk(dp->i_mount, &nodehdr1, node1); xfs_da3_node_hdr_from_disk(dp->i_mount, &nodehdr2, node2); btree1 = nodehdr1.btree; btree2 = nodehdr2.btree; } blk1->hashval = be32_to_cpu(btree1[nodehdr1.count - 1].hashval); blk2->hashval = be32_to_cpu(btree2[nodehdr2.count - 1].hashval); /* * Adjust the expected index for insertion. */ if (blk1->index >= nodehdr1.count) { blk2->index = blk1->index - nodehdr1.count; blk1->index = nodehdr1.count + 1; /* make it invalid */ } } /* * Add a new entry to an intermediate node. */ STATIC void xfs_da3_node_add( struct xfs_da_state *state, struct xfs_da_state_blk *oldblk, struct xfs_da_state_blk *newblk) { struct xfs_da_intnode *node; struct xfs_da3_icnode_hdr nodehdr; struct xfs_da_node_entry *btree; int tmp; struct xfs_inode *dp = state->args->dp; trace_xfs_da_node_add(state->args); node = oldblk->bp->b_addr; xfs_da3_node_hdr_from_disk(dp->i_mount, &nodehdr, node); btree = nodehdr.btree; ASSERT(oldblk->index >= 0 && oldblk->index <= nodehdr.count); ASSERT(newblk->blkno != 0); if (state->args->whichfork == XFS_DATA_FORK) ASSERT(newblk->blkno >= state->args->geo->leafblk && newblk->blkno < state->args->geo->freeblk); /* * We may need to make some room before we insert the new node. */ tmp = 0; if (oldblk->index < nodehdr.count) { tmp = (nodehdr.count - oldblk->index) * (uint)sizeof(*btree); memmove(&btree[oldblk->index + 1], &btree[oldblk->index], tmp); } btree[oldblk->index].hashval = cpu_to_be32(newblk->hashval); btree[oldblk->index].before = cpu_to_be32(newblk->blkno); xfs_trans_log_buf(state->args->trans, oldblk->bp, XFS_DA_LOGRANGE(node, &btree[oldblk->index], tmp + sizeof(*btree))); nodehdr.count += 1; xfs_da3_node_hdr_to_disk(dp->i_mount, node, &nodehdr); xfs_trans_log_buf(state->args->trans, oldblk->bp, XFS_DA_LOGRANGE(node, &node->hdr, state->args->geo->node_hdr_size)); /* * Copy the last hash value from the oldblk to propagate upwards. */ oldblk->hashval = be32_to_cpu(btree[nodehdr.count - 1].hashval); } /*======================================================================== * Routines used for shrinking the Btree. *========================================================================*/ /* * Deallocate an empty leaf node, remove it from its parent, * possibly deallocating that block, etc... */ int xfs_da3_join( struct xfs_da_state *state) { struct xfs_da_state_blk *drop_blk; struct xfs_da_state_blk *save_blk; int action = 0; int error; trace_xfs_da_join(state->args); drop_blk = &state->path.blk[ state->path.active-1 ]; save_blk = &state->altpath.blk[ state->path.active-1 ]; ASSERT(state->path.blk[0].magic == XFS_DA_NODE_MAGIC); ASSERT(drop_blk->magic == XFS_ATTR_LEAF_MAGIC || drop_blk->magic == XFS_DIR2_LEAFN_MAGIC); /* * Walk back up the tree joining/deallocating as necessary. * When we stop dropping blocks, break out. */ for ( ; state->path.active >= 2; drop_blk--, save_blk--, state->path.active--) { /* * See if we can combine the block with a neighbor. * (action == 0) => no options, just leave * (action == 1) => coalesce, then unlink * (action == 2) => block empty, unlink it */ switch (drop_blk->magic) { case XFS_ATTR_LEAF_MAGIC: error = xfs_attr3_leaf_toosmall(state, &action); if (error) return error; if (action == 0) return 0; xfs_attr3_leaf_unbalance(state, drop_blk, save_blk); break; case XFS_DIR2_LEAFN_MAGIC: error = xfs_dir2_leafn_toosmall(state, &action); if (error) return error; if (action == 0) return 0; xfs_dir2_leafn_unbalance(state, drop_blk, save_blk); break; case XFS_DA_NODE_MAGIC: /* * Remove the offending node, fixup hashvals, * check for a toosmall neighbor. */ xfs_da3_node_remove(state, drop_blk); xfs_da3_fixhashpath(state, &state->path); error = xfs_da3_node_toosmall(state, &action); if (error) return error; if (action == 0) return 0; xfs_da3_node_unbalance(state, drop_blk, save_blk); break; } xfs_da3_fixhashpath(state, &state->altpath); error = xfs_da3_blk_unlink(state, drop_blk, save_blk); xfs_da_state_kill_altpath(state); if (error) return error; error = xfs_da_shrink_inode(state->args, drop_blk->blkno, drop_blk->bp); drop_blk->bp = NULL; if (error) return error; } /* * We joined all the way to the top. If it turns out that * we only have one entry in the root, make the child block * the new root. */ xfs_da3_node_remove(state, drop_blk); xfs_da3_fixhashpath(state, &state->path); error = xfs_da3_root_join(state, &state->path.blk[0]); return error; } #ifdef DEBUG static void xfs_da_blkinfo_onlychild_validate(struct xfs_da_blkinfo *blkinfo, __u16 level) { __be16 magic = blkinfo->magic; if (level == 1) { ASSERT(magic == cpu_to_be16(XFS_DIR2_LEAFN_MAGIC) || magic == cpu_to_be16(XFS_DIR3_LEAFN_MAGIC) || magic == cpu_to_be16(XFS_ATTR_LEAF_MAGIC) || magic == cpu_to_be16(XFS_ATTR3_LEAF_MAGIC)); } else { ASSERT(magic == cpu_to_be16(XFS_DA_NODE_MAGIC) || magic == cpu_to_be16(XFS_DA3_NODE_MAGIC)); } ASSERT(!blkinfo->forw); ASSERT(!blkinfo->back); } #else /* !DEBUG */ #define xfs_da_blkinfo_onlychild_validate(blkinfo, level) #endif /* !DEBUG */ /* * We have only one entry in the root. Copy the only remaining child of * the old root to block 0 as the new root node. */ STATIC int xfs_da3_root_join( struct xfs_da_state *state, struct xfs_da_state_blk *root_blk) { struct xfs_da_intnode *oldroot; struct xfs_da_args *args; xfs_dablk_t child; struct xfs_buf *bp; struct xfs_da3_icnode_hdr oldroothdr; int error; struct xfs_inode *dp = state->args->dp; xfs_failaddr_t fa; trace_xfs_da_root_join(state->args); ASSERT(root_blk->magic == XFS_DA_NODE_MAGIC); args = state->args; oldroot = root_blk->bp->b_addr; xfs_da3_node_hdr_from_disk(dp->i_mount, &oldroothdr, oldroot); ASSERT(oldroothdr.forw == 0); ASSERT(oldroothdr.back == 0); /* * If the root has more than one child, then don't do anything. */ if (oldroothdr.count > 1) return 0; /* * Read in the (only) child block, then copy those bytes into * the root block's buffer and free the original child block. */ child = be32_to_cpu(oldroothdr.btree[0].before); ASSERT(child != 0); error = xfs_da3_node_read(args->trans, dp, child, &bp, args->whichfork); if (error) return error; fa = xfs_da3_header_check(bp, args->owner); if (fa) { __xfs_buf_mark_corrupt(bp, fa); xfs_trans_brelse(args->trans, bp); xfs_da_mark_sick(args); return -EFSCORRUPTED; } xfs_da_blkinfo_onlychild_validate(bp->b_addr, oldroothdr.level); /* * Copy child to root buffer and log it. */ xfs_da_buf_copy(root_blk->bp, bp, args->geo->blksize); xfs_trans_log_buf(args->trans, root_blk->bp, 0, args->geo->blksize - 1); /* * Now we can drop the child buffer. */ error = xfs_da_shrink_inode(args, child, bp); return error; } /* * Check a node block and its neighbors to see if the block should be * collapsed into one or the other neighbor. Always keep the block * with the smaller block number. * If the current block is over 50% full, don't try to join it, return 0. * If the block is empty, fill in the state structure and return 2. * If it can be collapsed, fill in the state structure and return 1. * If nothing can be done, return 0. */ STATIC int xfs_da3_node_toosmall( struct xfs_da_state *state, int *action) { struct xfs_da_intnode *node; struct xfs_da_state_blk *blk; struct xfs_da_blkinfo *info; xfs_dablk_t blkno; struct xfs_buf *bp; xfs_failaddr_t fa; struct xfs_da3_icnode_hdr nodehdr; int count; int forward; int error; int retval; int i; struct xfs_inode *dp = state->args->dp; trace_xfs_da_node_toosmall(state->args); /* * Check for the degenerate case of the block being over 50% full. * If so, it's not worth even looking to see if we might be able * to coalesce with a sibling. */ blk = &state->path.blk[ state->path.active-1 ]; info = blk->bp->b_addr; node = (xfs_da_intnode_t *)info; xfs_da3_node_hdr_from_disk(dp->i_mount, &nodehdr, node); if (nodehdr.count > (state->args->geo->node_ents >> 1)) { *action = 0; /* blk over 50%, don't try to join */ return 0; /* blk over 50%, don't try to join */ } /* * Check for the degenerate case of the block being empty. * If the block is empty, we'll simply delete it, no need to * coalesce it with a sibling block. We choose (arbitrarily) * to merge with the forward block unless it is NULL. */ if (nodehdr.count == 0) { /* * Make altpath point to the block we want to keep and * path point to the block we want to drop (this one). */ forward = (info->forw != 0); memcpy(&state->altpath, &state->path, sizeof(state->path)); error = xfs_da3_path_shift(state, &state->altpath, forward, 0, &retval); if (error) return error; if (retval) { *action = 0; } else { *action = 2; } return 0; } /* * Examine each sibling block to see if we can coalesce with * at least 25% free space to spare. We need to figure out * whether to merge with the forward or the backward block. * We prefer coalescing with the lower numbered sibling so as * to shrink a directory over time. */ count = state->args->geo->node_ents; count -= state->args->geo->node_ents >> 2; count -= nodehdr.count; /* start with smaller blk num */ forward = nodehdr.forw < nodehdr.back; for (i = 0; i < 2; forward = !forward, i++) { struct xfs_da3_icnode_hdr thdr; if (forward) blkno = nodehdr.forw; else blkno = nodehdr.back; if (blkno == 0) continue; error = xfs_da3_node_read(state->args->trans, dp, blkno, &bp, state->args->whichfork); if (error) return error; fa = xfs_da3_node_header_check(bp, state->args->owner); if (fa) { __xfs_buf_mark_corrupt(bp, fa); xfs_trans_brelse(state->args->trans, bp); xfs_da_mark_sick(state->args); return -EFSCORRUPTED; } node = bp->b_addr; xfs_da3_node_hdr_from_disk(dp->i_mount, &thdr, node); xfs_trans_brelse(state->args->trans, bp); if (count - thdr.count >= 0) break; /* fits with at least 25% to spare */ } if (i >= 2) { *action = 0; return 0; } /* * Make altpath point to the block we want to keep (the lower * numbered block) and path point to the block we want to drop. */ memcpy(&state->altpath, &state->path, sizeof(state->path)); if (blkno < blk->blkno) { error = xfs_da3_path_shift(state, &state->altpath, forward, 0, &retval); } else { error = xfs_da3_path_shift(state, &state->path, forward, 0, &retval); } if (error) return error; if (retval) { *action = 0; return 0; } *action = 1; return 0; } /* * Pick up the last hashvalue from an intermediate node. */ STATIC uint xfs_da3_node_lasthash( struct xfs_inode *dp, struct xfs_buf *bp, int *count) { struct xfs_da3_icnode_hdr nodehdr; xfs_da3_node_hdr_from_disk(dp->i_mount, &nodehdr, bp->b_addr); if (count) *count = nodehdr.count; if (!nodehdr.count) return 0; return be32_to_cpu(nodehdr.btree[nodehdr.count - 1].hashval); } /* * Walk back up the tree adjusting hash values as necessary, * when we stop making changes, return. */ void xfs_da3_fixhashpath( struct xfs_da_state *state, struct xfs_da_state_path *path) { struct xfs_da_state_blk *blk; struct xfs_da_intnode *node; struct xfs_da_node_entry *btree; xfs_dahash_t lasthash=0; int level; int count; struct xfs_inode *dp = state->args->dp; trace_xfs_da_fixhashpath(state->args); level = path->active-1; blk = &path->blk[ level ]; switch (blk->magic) { case XFS_ATTR_LEAF_MAGIC: lasthash = xfs_attr_leaf_lasthash(blk->bp, &count); if (count == 0) return; break; case XFS_DIR2_LEAFN_MAGIC: lasthash = xfs_dir2_leaf_lasthash(dp, blk->bp, &count); if (count == 0) return; break; case XFS_DA_NODE_MAGIC: lasthash = xfs_da3_node_lasthash(dp, blk->bp, &count); if (count == 0) return; break; } for (blk--, level--; level >= 0; blk--, level--) { struct xfs_da3_icnode_hdr nodehdr; node = blk->bp->b_addr; xfs_da3_node_hdr_from_disk(dp->i_mount, &nodehdr, node); btree = nodehdr.btree; if (be32_to_cpu(btree[blk->index].hashval) == lasthash) break; blk->hashval = lasthash; btree[blk->index].hashval = cpu_to_be32(lasthash); xfs_trans_log_buf(state->args->trans, blk->bp, XFS_DA_LOGRANGE(node, &btree[blk->index], sizeof(*btree))); lasthash = be32_to_cpu(btree[nodehdr.count - 1].hashval); } } /* * Remove an entry from an intermediate node. */ STATIC void xfs_da3_node_remove( struct xfs_da_state *state, struct xfs_da_state_blk *drop_blk) { struct xfs_da_intnode *node; struct xfs_da3_icnode_hdr nodehdr; struct xfs_da_node_entry *btree; int index; int tmp; struct xfs_inode *dp = state->args->dp; trace_xfs_da_node_remove(state->args); node = drop_blk->bp->b_addr; xfs_da3_node_hdr_from_disk(dp->i_mount, &nodehdr, node); ASSERT(drop_blk->index < nodehdr.count); ASSERT(drop_blk->index >= 0); /* * Copy over the offending entry, or just zero it out. */ index = drop_blk->index; btree = nodehdr.btree; if (index < nodehdr.count - 1) { tmp = nodehdr.count - index - 1; tmp *= (uint)sizeof(xfs_da_node_entry_t); memmove(&btree[index], &btree[index + 1], tmp); xfs_trans_log_buf(state->args->trans, drop_blk->bp, XFS_DA_LOGRANGE(node, &btree[index], tmp)); index = nodehdr.count - 1; } memset(&btree[index], 0, sizeof(xfs_da_node_entry_t)); xfs_trans_log_buf(state->args->trans, drop_blk->bp, XFS_DA_LOGRANGE(node, &btree[index], sizeof(btree[index]))); nodehdr.count -= 1; xfs_da3_node_hdr_to_disk(dp->i_mount, node, &nodehdr); xfs_trans_log_buf(state->args->trans, drop_blk->bp, XFS_DA_LOGRANGE(node, &node->hdr, state->args->geo->node_hdr_size)); /* * Copy the last hash value from the block to propagate upwards. */ drop_blk->hashval = be32_to_cpu(btree[index - 1].hashval); } /* * Unbalance the elements between two intermediate nodes, * move all Btree elements from one node into another. */ STATIC void xfs_da3_node_unbalance( struct xfs_da_state *state, struct xfs_da_state_blk *drop_blk, struct xfs_da_state_blk *save_blk) { struct xfs_da_intnode *drop_node; struct xfs_da_intnode *save_node; struct xfs_da_node_entry *drop_btree; struct xfs_da_node_entry *save_btree; struct xfs_da3_icnode_hdr drop_hdr; struct xfs_da3_icnode_hdr save_hdr; struct xfs_trans *tp; int sindex; int tmp; struct xfs_inode *dp = state->args->dp; trace_xfs_da_node_unbalance(state->args); drop_node = drop_blk->bp->b_addr; save_node = save_blk->bp->b_addr; xfs_da3_node_hdr_from_disk(dp->i_mount, &drop_hdr, drop_node); xfs_da3_node_hdr_from_disk(dp->i_mount, &save_hdr, save_node); drop_btree = drop_hdr.btree; save_btree = save_hdr.btree; tp = state->args->trans; /* * If the dying block has lower hashvals, then move all the * elements in the remaining block up to make a hole. */ if ((be32_to_cpu(drop_btree[0].hashval) < be32_to_cpu(save_btree[0].hashval)) || (be32_to_cpu(drop_btree[drop_hdr.count - 1].hashval) < be32_to_cpu(save_btree[save_hdr.count - 1].hashval))) { /* XXX: check this - is memmove dst correct? */ tmp = save_hdr.count * sizeof(xfs_da_node_entry_t); memmove(&save_btree[drop_hdr.count], &save_btree[0], tmp); sindex = 0; xfs_trans_log_buf(tp, save_blk->bp, XFS_DA_LOGRANGE(save_node, &save_btree[0], (save_hdr.count + drop_hdr.count) * sizeof(xfs_da_node_entry_t))); } else { sindex = save_hdr.count; xfs_trans_log_buf(tp, save_blk->bp, XFS_DA_LOGRANGE(save_node, &save_btree[sindex], drop_hdr.count * sizeof(xfs_da_node_entry_t))); } /* * Move all the B-tree elements from drop_blk to save_blk. */ tmp = drop_hdr.count * (uint)sizeof(xfs_da_node_entry_t); memcpy(&save_btree[sindex], &drop_btree[0], tmp); save_hdr.count += drop_hdr.count; xfs_da3_node_hdr_to_disk(dp->i_mount, save_node, &save_hdr); xfs_trans_log_buf(tp, save_blk->bp, XFS_DA_LOGRANGE(save_node, &save_node->hdr, state->args->geo->node_hdr_size)); /* * Save the last hashval in the remaining block for upward propagation. */ save_blk->hashval = be32_to_cpu(save_btree[save_hdr.count - 1].hashval); } /*======================================================================== * Routines used for finding things in the Btree. *========================================================================*/ /* * Walk down the Btree looking for a particular filename, filling * in the state structure as we go. * * We will set the state structure to point to each of the elements * in each of the nodes where either the hashval is or should be. * * We support duplicate hashval's so for each entry in the current * node that could contain the desired hashval, descend. This is a * pruned depth-first tree search. */ int /* error */ xfs_da3_node_lookup_int( struct xfs_da_state *state, int *result) { struct xfs_da_state_blk *blk; struct xfs_da_blkinfo *curr; struct xfs_da_intnode *node; struct xfs_da_node_entry *btree; struct xfs_da3_icnode_hdr nodehdr; struct xfs_da_args *args; xfs_failaddr_t fa; xfs_dablk_t blkno; xfs_dahash_t hashval; xfs_dahash_t btreehashval; int probe; int span; int max; int error; int retval; unsigned int expected_level = 0; uint16_t magic; struct xfs_inode *dp = state->args->dp; args = state->args; /* * Descend thru the B-tree searching each level for the right * node to use, until the right hashval is found. */ blkno = args->geo->leafblk; for (blk = &state->path.blk[0], state->path.active = 1; state->path.active <= XFS_DA_NODE_MAXDEPTH; blk++, state->path.active++) { /* * Read the next node down in the tree. */ blk->blkno = blkno; error = xfs_da3_node_read(args->trans, args->dp, blkno, &blk->bp, args->whichfork); if (error) { blk->blkno = 0; state->path.active--; return error; } curr = blk->bp->b_addr; magic = be16_to_cpu(curr->magic); if (magic == XFS_ATTR_LEAF_MAGIC || magic == XFS_ATTR3_LEAF_MAGIC) { fa = xfs_attr3_leaf_header_check(blk->bp, args->owner); if (fa) { __xfs_buf_mark_corrupt(blk->bp, fa); xfs_da_mark_sick(args); return -EFSCORRUPTED; } blk->magic = XFS_ATTR_LEAF_MAGIC; blk->hashval = xfs_attr_leaf_lasthash(blk->bp, NULL); break; } if (magic == XFS_DIR2_LEAFN_MAGIC || magic == XFS_DIR3_LEAFN_MAGIC) { fa = xfs_dir3_leaf_header_check(blk->bp, args->owner); if (fa) { __xfs_buf_mark_corrupt(blk->bp, fa); xfs_da_mark_sick(args); return -EFSCORRUPTED; } blk->magic = XFS_DIR2_LEAFN_MAGIC; blk->hashval = xfs_dir2_leaf_lasthash(args->dp, blk->bp, NULL); break; } if (magic != XFS_DA_NODE_MAGIC && magic != XFS_DA3_NODE_MAGIC) { xfs_buf_mark_corrupt(blk->bp); xfs_da_mark_sick(args); return -EFSCORRUPTED; } fa = xfs_da3_node_header_check(blk->bp, args->owner); if (fa) { __xfs_buf_mark_corrupt(blk->bp, fa); xfs_da_mark_sick(args); return -EFSCORRUPTED; } blk->magic = XFS_DA_NODE_MAGIC; /* * Search an intermediate node for a match. */ node = blk->bp->b_addr; xfs_da3_node_hdr_from_disk(dp->i_mount, &nodehdr, node); btree = nodehdr.btree; /* Tree taller than we can handle; bail out! */ if (nodehdr.level >= XFS_DA_NODE_MAXDEPTH) { xfs_buf_mark_corrupt(blk->bp); xfs_da_mark_sick(args); return -EFSCORRUPTED; } /* Check the level from the root. */ if (blkno == args->geo->leafblk) expected_level = nodehdr.level - 1; else if (expected_level != nodehdr.level) { xfs_buf_mark_corrupt(blk->bp); xfs_da_mark_sick(args); return -EFSCORRUPTED; } else expected_level--; max = nodehdr.count; blk->hashval = be32_to_cpu(btree[max - 1].hashval); /* * Binary search. (note: small blocks will skip loop) */ probe = span = max / 2; hashval = args->hashval; while (span > 4) { span /= 2; btreehashval = be32_to_cpu(btree[probe].hashval); if (btreehashval < hashval) probe += span; else if (btreehashval > hashval) probe -= span; else break; } ASSERT((probe >= 0) && (probe < max)); ASSERT((span <= 4) || (be32_to_cpu(btree[probe].hashval) == hashval)); /* * Since we may have duplicate hashval's, find the first * matching hashval in the node. */ while (probe > 0 && be32_to_cpu(btree[probe].hashval) >= hashval) { probe--; } while (probe < max && be32_to_cpu(btree[probe].hashval) < hashval) { probe++; } /* * Pick the right block to descend on. */ if (probe == max) { blk->index = max - 1; blkno = be32_to_cpu(btree[max - 1].before); } else { blk->index = probe; blkno = be32_to_cpu(btree[probe].before); } /* We can't point back to the root. */ if (XFS_IS_CORRUPT(dp->i_mount, blkno == args->geo->leafblk)) { xfs_da_mark_sick(args); return -EFSCORRUPTED; } } if (XFS_IS_CORRUPT(dp->i_mount, expected_level != 0)) { xfs_da_mark_sick(args); return -EFSCORRUPTED; } /* * A leaf block that ends in the hashval that we are interested in * (final hashval == search hashval) means that the next block may * contain more entries with the same hashval, shift upward to the * next leaf and keep searching. */ for (;;) { if (blk->magic == XFS_DIR2_LEAFN_MAGIC) { retval = xfs_dir2_leafn_lookup_int(blk->bp, args, &blk->index, state); } else if (blk->magic == XFS_ATTR_LEAF_MAGIC) { retval = xfs_attr3_leaf_lookup_int(blk->bp, args); blk->index = args->index; args->blkno = blk->blkno; } else { ASSERT(0); xfs_da_mark_sick(args); return -EFSCORRUPTED; } if (((retval == -ENOENT) || (retval == -ENOATTR)) && (blk->hashval == args->hashval)) { error = xfs_da3_path_shift(state, &state->path, 1, 1, &retval); if (error) return error; if (retval == 0) { continue; } else if (blk->magic == XFS_ATTR_LEAF_MAGIC) { /* path_shift() gives ENOENT */ retval = -ENOATTR; } } break; } *result = retval; return 0; } /*======================================================================== * Utility routines. *========================================================================*/ /* * Compare two intermediate nodes for "order". */ STATIC int xfs_da3_node_order( struct xfs_inode *dp, struct xfs_buf *node1_bp, struct xfs_buf *node2_bp) { struct xfs_da_intnode *node1; struct xfs_da_intnode *node2; struct xfs_da_node_entry *btree1; struct xfs_da_node_entry *btree2; struct xfs_da3_icnode_hdr node1hdr; struct xfs_da3_icnode_hdr node2hdr; node1 = node1_bp->b_addr; node2 = node2_bp->b_addr; xfs_da3_node_hdr_from_disk(dp->i_mount, &node1hdr, node1); xfs_da3_node_hdr_from_disk(dp->i_mount, &node2hdr, node2); btree1 = node1hdr.btree; btree2 = node2hdr.btree; if (node1hdr.count > 0 && node2hdr.count > 0 && ((be32_to_cpu(btree2[0].hashval) < be32_to_cpu(btree1[0].hashval)) || (be32_to_cpu(btree2[node2hdr.count - 1].hashval) < be32_to_cpu(btree1[node1hdr.count - 1].hashval)))) { return 1; } return 0; } /* * Link a new block into a doubly linked list of blocks (of whatever type). */ int /* error */ xfs_da3_blk_link( struct xfs_da_state *state, struct xfs_da_state_blk *old_blk, struct xfs_da_state_blk *new_blk) { struct xfs_da_blkinfo *old_info; struct xfs_da_blkinfo *new_info; struct xfs_da_blkinfo *tmp_info; struct xfs_da_args *args; struct xfs_buf *bp; xfs_failaddr_t fa; int before = 0; int error; struct xfs_inode *dp = state->args->dp; /* * Set up environment. */ args = state->args; ASSERT(args != NULL); old_info = old_blk->bp->b_addr; new_info = new_blk->bp->b_addr; ASSERT(old_blk->magic == XFS_DA_NODE_MAGIC || old_blk->magic == XFS_DIR2_LEAFN_MAGIC || old_blk->magic == XFS_ATTR_LEAF_MAGIC); switch (old_blk->magic) { case XFS_ATTR_LEAF_MAGIC: before = xfs_attr_leaf_order(old_blk->bp, new_blk->bp); break; case XFS_DIR2_LEAFN_MAGIC: before = xfs_dir2_leafn_order(dp, old_blk->bp, new_blk->bp); break; case XFS_DA_NODE_MAGIC: before = xfs_da3_node_order(dp, old_blk->bp, new_blk->bp); break; } /* * Link blocks in appropriate order. */ if (before) { /* * Link new block in before existing block. */ trace_xfs_da_link_before(args); new_info->forw = cpu_to_be32(old_blk->blkno); new_info->back = old_info->back; if (old_info->back) { error = xfs_da3_node_read(args->trans, dp, be32_to_cpu(old_info->back), &bp, args->whichfork); if (error) return error; fa = xfs_da3_header_check(bp, args->owner); if (fa) { __xfs_buf_mark_corrupt(bp, fa); xfs_trans_brelse(args->trans, bp); xfs_da_mark_sick(args); return -EFSCORRUPTED; } ASSERT(bp != NULL); tmp_info = bp->b_addr; ASSERT(tmp_info->magic == old_info->magic); ASSERT(be32_to_cpu(tmp_info->forw) == old_blk->blkno); tmp_info->forw = cpu_to_be32(new_blk->blkno); xfs_trans_log_buf(args->trans, bp, 0, sizeof(*tmp_info)-1); } old_info->back = cpu_to_be32(new_blk->blkno); } else { /* * Link new block in after existing block. */ trace_xfs_da_link_after(args); new_info->forw = old_info->forw; new_info->back = cpu_to_be32(old_blk->blkno); if (old_info->forw) { error = xfs_da3_node_read(args->trans, dp, be32_to_cpu(old_info->forw), &bp, args->whichfork); if (error) return error; fa = xfs_da3_header_check(bp, args->owner); if (fa) { __xfs_buf_mark_corrupt(bp, fa); xfs_trans_brelse(args->trans, bp); xfs_da_mark_sick(args); return -EFSCORRUPTED; } ASSERT(bp != NULL); tmp_info = bp->b_addr; ASSERT(tmp_info->magic == old_info->magic); ASSERT(be32_to_cpu(tmp_info->back) == old_blk->blkno); tmp_info->back = cpu_to_be32(new_blk->blkno); xfs_trans_log_buf(args->trans, bp, 0, sizeof(*tmp_info)-1); } old_info->forw = cpu_to_be32(new_blk->blkno); } xfs_trans_log_buf(args->trans, old_blk->bp, 0, sizeof(*tmp_info) - 1); xfs_trans_log_buf(args->trans, new_blk->bp, 0, sizeof(*tmp_info) - 1); return 0; } /* * Unlink a block from a doubly linked list of blocks. */ STATIC int /* error */ xfs_da3_blk_unlink( struct xfs_da_state *state, struct xfs_da_state_blk *drop_blk, struct xfs_da_state_blk *save_blk) { struct xfs_da_blkinfo *drop_info; struct xfs_da_blkinfo *save_info; struct xfs_da_blkinfo *tmp_info; struct xfs_da_args *args; struct xfs_buf *bp; xfs_failaddr_t fa; int error; /* * Set up environment. */ args = state->args; ASSERT(args != NULL); save_info = save_blk->bp->b_addr; drop_info = drop_blk->bp->b_addr; ASSERT(save_blk->magic == XFS_DA_NODE_MAGIC || save_blk->magic == XFS_DIR2_LEAFN_MAGIC || save_blk->magic == XFS_ATTR_LEAF_MAGIC); ASSERT(save_blk->magic == drop_blk->magic); ASSERT((be32_to_cpu(save_info->forw) == drop_blk->blkno) || (be32_to_cpu(save_info->back) == drop_blk->blkno)); ASSERT((be32_to_cpu(drop_info->forw) == save_blk->blkno) || (be32_to_cpu(drop_info->back) == save_blk->blkno)); /* * Unlink the leaf block from the doubly linked chain of leaves. */ if (be32_to_cpu(save_info->back) == drop_blk->blkno) { trace_xfs_da_unlink_back(args); save_info->back = drop_info->back; if (drop_info->back) { error = xfs_da3_node_read(args->trans, args->dp, be32_to_cpu(drop_info->back), &bp, args->whichfork); if (error) return error; fa = xfs_da3_header_check(bp, args->owner); if (fa) { __xfs_buf_mark_corrupt(bp, fa); xfs_trans_brelse(args->trans, bp); xfs_da_mark_sick(args); return -EFSCORRUPTED; } ASSERT(bp != NULL); tmp_info = bp->b_addr; ASSERT(tmp_info->magic == save_info->magic); ASSERT(be32_to_cpu(tmp_info->forw) == drop_blk->blkno); tmp_info->forw = cpu_to_be32(save_blk->blkno); xfs_trans_log_buf(args->trans, bp, 0, sizeof(*tmp_info) - 1); } } else { trace_xfs_da_unlink_forward(args); save_info->forw = drop_info->forw; if (drop_info->forw) { error = xfs_da3_node_read(args->trans, args->dp, be32_to_cpu(drop_info->forw), &bp, args->whichfork); if (error) return error; fa = xfs_da3_header_check(bp, args->owner); if (fa) { __xfs_buf_mark_corrupt(bp, fa); xfs_trans_brelse(args->trans, bp); xfs_da_mark_sick(args); return -EFSCORRUPTED; } ASSERT(bp != NULL); tmp_info = bp->b_addr; ASSERT(tmp_info->magic == save_info->magic); ASSERT(be32_to_cpu(tmp_info->back) == drop_blk->blkno); tmp_info->back = cpu_to_be32(save_blk->blkno); xfs_trans_log_buf(args->trans, bp, 0, sizeof(*tmp_info) - 1); } } xfs_trans_log_buf(args->trans, save_blk->bp, 0, sizeof(*save_info) - 1); return 0; } /* * Move a path "forward" or "!forward" one block at the current level. * * This routine will adjust a "path" to point to the next block * "forward" (higher hashvalues) or "!forward" (lower hashvals) in the * Btree, including updating pointers to the intermediate nodes between * the new bottom and the root. */ int /* error */ xfs_da3_path_shift( struct xfs_da_state *state, struct xfs_da_state_path *path, int forward, int release, int *result) { struct xfs_da_state_blk *blk; struct xfs_da_blkinfo *info; struct xfs_da_args *args; struct xfs_da_node_entry *btree; struct xfs_da3_icnode_hdr nodehdr; struct xfs_buf *bp; xfs_failaddr_t fa; xfs_dablk_t blkno = 0; int level; int error; struct xfs_inode *dp = state->args->dp; trace_xfs_da_path_shift(state->args); /* * Roll up the Btree looking for the first block where our * current index is not at the edge of the block. Note that * we skip the bottom layer because we want the sibling block. */ args = state->args; ASSERT(args != NULL); ASSERT(path != NULL); ASSERT((path->active > 0) && (path->active < XFS_DA_NODE_MAXDEPTH)); level = (path->active-1) - 1; /* skip bottom layer in path */ for (; level >= 0; level--) { blk = &path->blk[level]; xfs_da3_node_hdr_from_disk(dp->i_mount, &nodehdr, blk->bp->b_addr); if (forward && (blk->index < nodehdr.count - 1)) { blk->index++; blkno = be32_to_cpu(nodehdr.btree[blk->index].before); break; } else if (!forward && (blk->index > 0)) { blk->index--; blkno = be32_to_cpu(nodehdr.btree[blk->index].before); break; } } if (level < 0) { *result = -ENOENT; /* we're out of our tree */ ASSERT(args->op_flags & XFS_DA_OP_OKNOENT); return 0; } /* * Roll down the edge of the subtree until we reach the * same depth we were at originally. */ for (blk++, level++; level < path->active; blk++, level++) { /* * Read the next child block into a local buffer. */ error = xfs_da3_node_read(args->trans, dp, blkno, &bp, args->whichfork); if (error) return error; /* * Release the old block (if it's dirty, the trans doesn't * actually let go) and swap the local buffer into the path * structure. This ensures failure of the above read doesn't set * a NULL buffer in an active slot in the path. */ if (release) xfs_trans_brelse(args->trans, blk->bp); blk->blkno = blkno; blk->bp = bp; info = blk->bp->b_addr; ASSERT(info->magic == cpu_to_be16(XFS_DA_NODE_MAGIC) || info->magic == cpu_to_be16(XFS_DA3_NODE_MAGIC) || info->magic == cpu_to_be16(XFS_DIR2_LEAFN_MAGIC) || info->magic == cpu_to_be16(XFS_DIR3_LEAFN_MAGIC) || info->magic == cpu_to_be16(XFS_ATTR_LEAF_MAGIC) || info->magic == cpu_to_be16(XFS_ATTR3_LEAF_MAGIC)); /* * Note: we flatten the magic number to a single type so we * don't have to compare against crc/non-crc types elsewhere. */ switch (be16_to_cpu(info->magic)) { case XFS_DA_NODE_MAGIC: case XFS_DA3_NODE_MAGIC: fa = xfs_da3_node_header_check(blk->bp, args->owner); if (fa) { __xfs_buf_mark_corrupt(blk->bp, fa); xfs_da_mark_sick(args); return -EFSCORRUPTED; } blk->magic = XFS_DA_NODE_MAGIC; xfs_da3_node_hdr_from_disk(dp->i_mount, &nodehdr, bp->b_addr); btree = nodehdr.btree; blk->hashval = be32_to_cpu(btree[nodehdr.count - 1].hashval); if (forward) blk->index = 0; else blk->index = nodehdr.count - 1; blkno = be32_to_cpu(btree[blk->index].before); break; case XFS_ATTR_LEAF_MAGIC: case XFS_ATTR3_LEAF_MAGIC: fa = xfs_attr3_leaf_header_check(blk->bp, args->owner); if (fa) { __xfs_buf_mark_corrupt(blk->bp, fa); xfs_da_mark_sick(args); return -EFSCORRUPTED; } blk->magic = XFS_ATTR_LEAF_MAGIC; ASSERT(level == path->active-1); blk->index = 0; blk->hashval = xfs_attr_leaf_lasthash(blk->bp, NULL); break; case XFS_DIR2_LEAFN_MAGIC: case XFS_DIR3_LEAFN_MAGIC: fa = xfs_dir3_leaf_header_check(blk->bp, args->owner); if (fa) { __xfs_buf_mark_corrupt(blk->bp, fa); xfs_da_mark_sick(args); return -EFSCORRUPTED; } blk->magic = XFS_DIR2_LEAFN_MAGIC; ASSERT(level == path->active-1); blk->index = 0; blk->hashval = xfs_dir2_leaf_lasthash(args->dp, blk->bp, NULL); break; default: ASSERT(0); break; } } *result = 0; return 0; } /*======================================================================== * Utility routines. *========================================================================*/ /* * Implement a simple hash on a character string. * Rotate the hash value by 7 bits, then XOR each character in. * This is implemented with some source-level loop unrolling. */ xfs_dahash_t xfs_da_hashname(const uint8_t *name, int namelen) { xfs_dahash_t hash; /* * Do four characters at a time as long as we can. */ for (hash = 0; namelen >= 4; namelen -= 4, name += 4) hash = (name[0] << 21) ^ (name[1] << 14) ^ (name[2] << 7) ^ (name[3] << 0) ^ rol32(hash, 7 * 4); /* * Now do the rest of the characters. */ switch (namelen) { case 3: return (name[0] << 14) ^ (name[1] << 7) ^ (name[2] << 0) ^ rol32(hash, 7 * 3); case 2: return (name[0] << 7) ^ (name[1] << 0) ^ rol32(hash, 7 * 2); case 1: return (name[0] << 0) ^ rol32(hash, 7 * 1); default: /* case 0: */ return hash; } } enum xfs_dacmp xfs_da_compname( struct xfs_da_args *args, const unsigned char *name, int len) { return (args->namelen == len && memcmp(args->name, name, len) == 0) ? XFS_CMP_EXACT : XFS_CMP_DIFFERENT; } int xfs_da_grow_inode_int( struct xfs_da_args *args, xfs_fileoff_t *bno, int count) { struct xfs_trans *tp = args->trans; struct xfs_inode *dp = args->dp; int w = args->whichfork; xfs_rfsblock_t nblks = dp->i_nblocks; struct xfs_bmbt_irec map, *mapp = ↦ int nmap, error, got, i, mapi = 1; /* * Find a spot in the file space to put the new block. */ error = xfs_bmap_first_unused(tp, dp, count, bno, w); if (error) return error; /* * Try mapping it in one filesystem block. */ nmap = 1; error = xfs_bmapi_write(tp, dp, *bno, count, xfs_bmapi_aflag(w)|XFS_BMAPI_METADATA|XFS_BMAPI_CONTIG, args->total, &map, &nmap); if (error == -ENOSPC && count > 1) { xfs_fileoff_t b; int c; /* * If we didn't get it and the block might work if fragmented, * try without the CONTIG flag. Loop until we get it all. */ mapp = kmalloc(sizeof(*mapp) * count, GFP_KERNEL | __GFP_NOFAIL); for (b = *bno, mapi = 0; b < *bno + count; ) { c = (int)(*bno + count - b); nmap = min(XFS_BMAP_MAX_NMAP, c); error = xfs_bmapi_write(tp, dp, b, c, xfs_bmapi_aflag(w)|XFS_BMAPI_METADATA, args->total, &mapp[mapi], &nmap); if (error) goto out_free_map; mapi += nmap; b = mapp[mapi - 1].br_startoff + mapp[mapi - 1].br_blockcount; } } if (error) goto out_free_map; /* * Count the blocks we got, make sure it matches the total. */ for (i = 0, got = 0; i < mapi; i++) got += mapp[i].br_blockcount; if (got != count || mapp[0].br_startoff != *bno || mapp[mapi - 1].br_startoff + mapp[mapi - 1].br_blockcount != *bno + count) { error = -ENOSPC; goto out_free_map; } /* account for newly allocated blocks in reserved blocks total */ args->total -= dp->i_nblocks - nblks; out_free_map: if (mapp != &map) kfree(mapp); return error; } /* * Add a block to the btree ahead of the file. * Return the new block number to the caller. */ int xfs_da_grow_inode( struct xfs_da_args *args, xfs_dablk_t *new_blkno) { xfs_fileoff_t bno; int error; trace_xfs_da_grow_inode(args); bno = args->geo->leafblk; error = xfs_da_grow_inode_int(args, &bno, args->geo->fsbcount); if (!error) *new_blkno = (xfs_dablk_t)bno; return error; } /* * Ick. We need to always be able to remove a btree block, even * if there's no space reservation because the filesystem is full. * This is called if xfs_bunmapi on a btree block fails due to ENOSPC. * It swaps the target block with the last block in the file. The * last block in the file can always be removed since it can't cause * a bmap btree split to do that. */ STATIC int xfs_da3_swap_lastblock( struct xfs_da_args *args, xfs_dablk_t *dead_blknop, struct xfs_buf **dead_bufp) { struct xfs_da_blkinfo *dead_info; struct xfs_da_blkinfo *sib_info; struct xfs_da_intnode *par_node; struct xfs_da_intnode *dead_node; struct xfs_dir2_leaf *dead_leaf2; struct xfs_da_node_entry *btree; struct xfs_da3_icnode_hdr par_hdr; struct xfs_inode *dp; struct xfs_trans *tp; struct xfs_mount *mp; struct xfs_buf *dead_buf; struct xfs_buf *last_buf; struct xfs_buf *sib_buf; struct xfs_buf *par_buf; xfs_failaddr_t fa; xfs_dahash_t dead_hash; xfs_fileoff_t lastoff; xfs_dablk_t dead_blkno; xfs_dablk_t last_blkno; xfs_dablk_t sib_blkno; xfs_dablk_t par_blkno; int error; int w; int entno; int level; int dead_level; trace_xfs_da_swap_lastblock(args); dead_buf = *dead_bufp; dead_blkno = *dead_blknop; tp = args->trans; dp = args->dp; w = args->whichfork; ASSERT(w == XFS_DATA_FORK); mp = dp->i_mount; lastoff = args->geo->freeblk; error = xfs_bmap_last_before(tp, dp, &lastoff, w); if (error) return error; if (XFS_IS_CORRUPT(mp, lastoff == 0)) { xfs_da_mark_sick(args); return -EFSCORRUPTED; } /* * Read the last block in the btree space. */ last_blkno = (xfs_dablk_t)lastoff - args->geo->fsbcount; error = xfs_da3_node_read(tp, dp, last_blkno, &last_buf, w); if (error) return error; fa = xfs_da3_header_check(last_buf, args->owner); if (fa) { __xfs_buf_mark_corrupt(last_buf, fa); xfs_trans_brelse(tp, last_buf); xfs_da_mark_sick(args); return -EFSCORRUPTED; } /* * Copy the last block into the dead buffer and log it. */ xfs_da_buf_copy(dead_buf, last_buf, args->geo->blksize); xfs_trans_log_buf(tp, dead_buf, 0, args->geo->blksize - 1); dead_info = dead_buf->b_addr; /* * Get values from the moved block. */ if (dead_info->magic == cpu_to_be16(XFS_DIR2_LEAFN_MAGIC) || dead_info->magic == cpu_to_be16(XFS_DIR3_LEAFN_MAGIC)) { struct xfs_dir3_icleaf_hdr leafhdr; struct xfs_dir2_leaf_entry *ents; dead_leaf2 = (xfs_dir2_leaf_t *)dead_info; xfs_dir2_leaf_hdr_from_disk(dp->i_mount, &leafhdr, dead_leaf2); ents = leafhdr.ents; dead_level = 0; dead_hash = be32_to_cpu(ents[leafhdr.count - 1].hashval); } else { struct xfs_da3_icnode_hdr deadhdr; dead_node = (xfs_da_intnode_t *)dead_info; xfs_da3_node_hdr_from_disk(dp->i_mount, &deadhdr, dead_node); btree = deadhdr.btree; dead_level = deadhdr.level; dead_hash = be32_to_cpu(btree[deadhdr.count - 1].hashval); } sib_buf = par_buf = NULL; /* * If the moved block has a left sibling, fix up the pointers. */ if ((sib_blkno = be32_to_cpu(dead_info->back))) { error = xfs_da3_node_read(tp, dp, sib_blkno, &sib_buf, w); if (error) goto done; fa = xfs_da3_header_check(sib_buf, args->owner); if (fa) { __xfs_buf_mark_corrupt(sib_buf, fa); xfs_da_mark_sick(args); error = -EFSCORRUPTED; goto done; } sib_info = sib_buf->b_addr; if (XFS_IS_CORRUPT(mp, be32_to_cpu(sib_info->forw) != last_blkno || sib_info->magic != dead_info->magic)) { xfs_da_mark_sick(args); error = -EFSCORRUPTED; goto done; } sib_info->forw = cpu_to_be32(dead_blkno); xfs_trans_log_buf(tp, sib_buf, XFS_DA_LOGRANGE(sib_info, &sib_info->forw, sizeof(sib_info->forw))); sib_buf = NULL; } /* * If the moved block has a right sibling, fix up the pointers. */ if ((sib_blkno = be32_to_cpu(dead_info->forw))) { error = xfs_da3_node_read(tp, dp, sib_blkno, &sib_buf, w); if (error) goto done; fa = xfs_da3_header_check(sib_buf, args->owner); if (fa) { __xfs_buf_mark_corrupt(sib_buf, fa); xfs_da_mark_sick(args); error = -EFSCORRUPTED; goto done; } sib_info = sib_buf->b_addr; if (XFS_IS_CORRUPT(mp, be32_to_cpu(sib_info->back) != last_blkno || sib_info->magic != dead_info->magic)) { xfs_da_mark_sick(args); error = -EFSCORRUPTED; goto done; } sib_info->back = cpu_to_be32(dead_blkno); xfs_trans_log_buf(tp, sib_buf, XFS_DA_LOGRANGE(sib_info, &sib_info->back, sizeof(sib_info->back))); sib_buf = NULL; } par_blkno = args->geo->leafblk; level = -1; /* * Walk down the tree looking for the parent of the moved block. */ for (;;) { error = xfs_da3_node_read(tp, dp, par_blkno, &par_buf, w); if (error) goto done; fa = xfs_da3_node_header_check(par_buf, args->owner); if (fa) { __xfs_buf_mark_corrupt(par_buf, fa); xfs_da_mark_sick(args); error = -EFSCORRUPTED; goto done; } par_node = par_buf->b_addr; xfs_da3_node_hdr_from_disk(dp->i_mount, &par_hdr, par_node); if (XFS_IS_CORRUPT(mp, level >= 0 && level != par_hdr.level + 1)) { xfs_da_mark_sick(args); error = -EFSCORRUPTED; goto done; } level = par_hdr.level; btree = par_hdr.btree; for (entno = 0; entno < par_hdr.count && be32_to_cpu(btree[entno].hashval) < dead_hash; entno++) continue; if (XFS_IS_CORRUPT(mp, entno == par_hdr.count)) { xfs_da_mark_sick(args); error = -EFSCORRUPTED; goto done; } par_blkno = be32_to_cpu(btree[entno].before); if (level == dead_level + 1) break; xfs_trans_brelse(tp, par_buf); par_buf = NULL; } /* * We're in the right parent block. * Look for the right entry. */ for (;;) { for (; entno < par_hdr.count && be32_to_cpu(btree[entno].before) != last_blkno; entno++) continue; if (entno < par_hdr.count) break; par_blkno = par_hdr.forw; xfs_trans_brelse(tp, par_buf); par_buf = NULL; if (XFS_IS_CORRUPT(mp, par_blkno == 0)) { xfs_da_mark_sick(args); error = -EFSCORRUPTED; goto done; } error = xfs_da3_node_read(tp, dp, par_blkno, &par_buf, w); if (error) goto done; fa = xfs_da3_node_header_check(par_buf, args->owner); if (fa) { __xfs_buf_mark_corrupt(par_buf, fa); xfs_da_mark_sick(args); error = -EFSCORRUPTED; goto done; } par_node = par_buf->b_addr; xfs_da3_node_hdr_from_disk(dp->i_mount, &par_hdr, par_node); if (XFS_IS_CORRUPT(mp, par_hdr.level != level)) { xfs_da_mark_sick(args); error = -EFSCORRUPTED; goto done; } btree = par_hdr.btree; entno = 0; } /* * Update the parent entry pointing to the moved block. */ btree[entno].before = cpu_to_be32(dead_blkno); xfs_trans_log_buf(tp, par_buf, XFS_DA_LOGRANGE(par_node, &btree[entno].before, sizeof(btree[entno].before))); *dead_blknop = last_blkno; *dead_bufp = last_buf; return 0; done: if (par_buf) xfs_trans_brelse(tp, par_buf); if (sib_buf) xfs_trans_brelse(tp, sib_buf); xfs_trans_brelse(tp, last_buf); return error; } /* * Remove a btree block from a directory or attribute. */ int xfs_da_shrink_inode( struct xfs_da_args *args, xfs_dablk_t dead_blkno, struct xfs_buf *dead_buf) { struct xfs_inode *dp; int done, error, w, count; struct xfs_trans *tp; trace_xfs_da_shrink_inode(args); dp = args->dp; w = args->whichfork; tp = args->trans; count = args->geo->fsbcount; for (;;) { /* * Remove extents. If we get ENOSPC for a dir we have to move * the last block to the place we want to kill. */ error = xfs_bunmapi(tp, dp, dead_blkno, count, xfs_bmapi_aflag(w), 0, &done); if (error == -ENOSPC) { if (w != XFS_DATA_FORK) break; error = xfs_da3_swap_lastblock(args, &dead_blkno, &dead_buf); if (error) break; } else { break; } } xfs_trans_binval(tp, dead_buf); return error; } static int xfs_dabuf_map( struct xfs_inode *dp, xfs_dablk_t bno, unsigned int flags, int whichfork, struct xfs_buf_map **mapp, int *nmaps) { struct xfs_mount *mp = dp->i_mount; int nfsb = xfs_dabuf_nfsb(mp, whichfork); struct xfs_bmbt_irec irec, *irecs = &irec; struct xfs_buf_map *map = *mapp; xfs_fileoff_t off = bno; int error = 0, nirecs, i; if (nfsb > 1) irecs = kzalloc(sizeof(irec) * nfsb, GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_NOFAIL); nirecs = nfsb; error = xfs_bmapi_read(dp, bno, nfsb, irecs, &nirecs, xfs_bmapi_aflag(whichfork)); if (error) goto out_free_irecs; /* * Use the caller provided map for the single map case, else allocate a * larger one that needs to be free by the caller. */ if (nirecs > 1) { map = kzalloc(nirecs * sizeof(struct xfs_buf_map), GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_NOFAIL); if (!map) { error = -ENOMEM; goto out_free_irecs; } *mapp = map; } for (i = 0; i < nirecs; i++) { if (irecs[i].br_startblock == HOLESTARTBLOCK || irecs[i].br_startblock == DELAYSTARTBLOCK) goto invalid_mapping; if (off != irecs[i].br_startoff) goto invalid_mapping; map[i].bm_bn = XFS_FSB_TO_DADDR(mp, irecs[i].br_startblock); map[i].bm_len = XFS_FSB_TO_BB(mp, irecs[i].br_blockcount); off += irecs[i].br_blockcount; } if (off != bno + nfsb) goto invalid_mapping; *nmaps = nirecs; out_free_irecs: if (irecs != &irec) kfree(irecs); return error; invalid_mapping: /* Caller ok with no mapping. */ if (XFS_IS_CORRUPT(mp, !(flags & XFS_DABUF_MAP_HOLE_OK))) { xfs_dirattr_mark_sick(dp, whichfork); error = -EFSCORRUPTED; if (xfs_error_level >= XFS_ERRLEVEL_LOW) { xfs_alert(mp, "%s: bno %u inode %llu", __func__, bno, dp->i_ino); for (i = 0; i < nirecs; i++) { xfs_alert(mp, "[%02d] br_startoff %lld br_startblock %lld br_blockcount %lld br_state %d", i, irecs[i].br_startoff, irecs[i].br_startblock, irecs[i].br_blockcount, irecs[i].br_state); } } } else { *nmaps = 0; } goto out_free_irecs; } /* * Get a buffer for the dir/attr block. */ int xfs_da_get_buf( struct xfs_trans *tp, struct xfs_inode *dp, xfs_dablk_t bno, struct xfs_buf **bpp, int whichfork) { struct xfs_mount *mp = dp->i_mount; struct xfs_buf *bp; struct xfs_buf_map map, *mapp = ↦ int nmap = 1; int error; *bpp = NULL; error = xfs_dabuf_map(dp, bno, 0, whichfork, &mapp, &nmap); if (error || nmap == 0) goto out_free; error = xfs_trans_get_buf_map(tp, mp->m_ddev_targp, mapp, nmap, 0, &bp); if (error) goto out_free; *bpp = bp; out_free: if (mapp != &map) kfree(mapp); return error; } /* * Get a buffer for the dir/attr block, fill in the contents. */ int xfs_da_read_buf( struct xfs_trans *tp, struct xfs_inode *dp, xfs_dablk_t bno, unsigned int flags, struct xfs_buf **bpp, int whichfork, const struct xfs_buf_ops *ops) { struct xfs_mount *mp = dp->i_mount; struct xfs_buf *bp; struct xfs_buf_map map, *mapp = ↦ int nmap = 1; int error; *bpp = NULL; error = xfs_dabuf_map(dp, bno, flags, whichfork, &mapp, &nmap); if (error || !nmap) goto out_free; error = xfs_trans_read_buf_map(mp, tp, mp->m_ddev_targp, mapp, nmap, 0, &bp, ops); if (xfs_metadata_is_sick(error)) xfs_dirattr_mark_sick(dp, whichfork); if (error) goto out_free; if (whichfork == XFS_ATTR_FORK) xfs_buf_set_ref(bp, XFS_ATTR_BTREE_REF); else xfs_buf_set_ref(bp, XFS_DIR_BTREE_REF); *bpp = bp; out_free: if (mapp != &map) kfree(mapp); return error; } /* * Readahead the dir/attr block. */ int xfs_da_reada_buf( struct xfs_inode *dp, xfs_dablk_t bno, unsigned int flags, int whichfork, const struct xfs_buf_ops *ops) { struct xfs_buf_map map; struct xfs_buf_map *mapp; int nmap; int error; mapp = ↦ nmap = 1; error = xfs_dabuf_map(dp, bno, flags, whichfork, &mapp, &nmap); if (error || !nmap) goto out_free; xfs_buf_readahead_map(dp->i_mount->m_ddev_targp, mapp, nmap, ops); out_free: if (mapp != &map) kfree(mapp); return error; }
Information contained on this website is for historical information purposes only and does not indicate or represent copyright ownership.
Created with Cregit http://github.com/cregit/cregit
Version 2.0-RC1