summaryrefslogtreecommitdiffstats
path: root/src/VBox/Runtime/common/fs/xfsvfs.cpp
diff options
context:
space:
mode:
Diffstat (limited to 'src/VBox/Runtime/common/fs/xfsvfs.cpp')
-rw-r--r--src/VBox/Runtime/common/fs/xfsvfs.cpp2240
1 files changed, 2240 insertions, 0 deletions
diff --git a/src/VBox/Runtime/common/fs/xfsvfs.cpp b/src/VBox/Runtime/common/fs/xfsvfs.cpp
new file mode 100644
index 00000000..84432e35
--- /dev/null
+++ b/src/VBox/Runtime/common/fs/xfsvfs.cpp
@@ -0,0 +1,2240 @@
+/* $Id: xfsvfs.cpp $ */
+/** @file
+ * IPRT - XFS Virtual Filesystem.
+ */
+
+/*
+ * Copyright (C) 2018-2019 Oracle Corporation
+ *
+ * This file is part of VirtualBox Open Source Edition (OSE), as
+ * available from http://www.virtualbox.org. This file is free software;
+ * you can redistribute it and/or modify it under the terms of the GNU
+ * General Public License (GPL) as published by the Free Software
+ * Foundation, in version 2 as it comes in the "COPYING" file of the
+ * VirtualBox OSE distribution. VirtualBox OSE is distributed in the
+ * hope that it will be useful, but WITHOUT ANY WARRANTY of any kind.
+ *
+ * The contents of this file may alternatively be used under the terms
+ * of the Common Development and Distribution License Version 1.0
+ * (CDDL) only, as it comes in the "COPYING.CDDL" file of the
+ * VirtualBox OSE distribution, in which case the provisions of the
+ * CDDL are applicable instead of those of the GPL.
+ *
+ * You may elect to license modified versions of this file under the
+ * terms and conditions of either the GPL or the CDDL or both.
+ */
+
+
+/*********************************************************************************************************************************
+* Header Files *
+*********************************************************************************************************************************/
+#define LOG_GROUP RTLOGGROUP_FS
+#include <iprt/fsvfs.h>
+
+#include <iprt/asm.h>
+#include <iprt/assert.h>
+#include <iprt/avl.h>
+#include <iprt/file.h>
+#include <iprt/err.h>
+#include <iprt/list.h>
+#include <iprt/log.h>
+#include <iprt/mem.h>
+#include <iprt/string.h>
+#include <iprt/vfs.h>
+#include <iprt/vfslowlevel.h>
+#include <iprt/formats/xfs.h>
+
+
+/*********************************************************************************************************************************
+* Defined Constants And Macros *
+*********************************************************************************************************************************/
+/** The maximum allocation group cache size (in bytes). */
+#if ARCH_BITS >= 64
+# define RTFSXFS_MAX_AG_CACHE_SIZE _512K
+#else
+# define RTFSXFS_MAX_AG_CACHE_SIZE _128K
+#endif
+/** The maximum inode cache size (in bytes). */
+#if ARCH_BITS >= 64
+# define RTFSXFS_MAX_INODE_CACHE_SIZE _512K
+#else
+# define RTFSXFS_MAX_INODE_CACHE_SIZE _128K
+#endif
+/** The maximum extent tree cache size (in bytes). */
+#if ARCH_BITS >= 64
+# define RTFSXFS_MAX_BLOCK_CACHE_SIZE _512K
+#else
+# define RTFSXFS_MAX_BLOCK_CACHE_SIZE _128K
+#endif
+
+
+/*********************************************************************************************************************************
+* Structures and Typedefs *
+*********************************************************************************************************************************/
+/** Pointer to the XFS filesystem data. */
+typedef struct RTFSXFSVOL *PRTFSXFSVOL;
+
+
+/**
+ * Cached allocation group descriptor data.
+ */
+typedef struct RTFSXFSAG
+{
+ /** AVL tree node, indexed by the allocation group number. */
+ AVLU32NODECORE Core;
+ /** List node for the LRU list used for eviction. */
+ RTLISTNODE NdLru;
+ /** Reference counter. */
+ volatile uint32_t cRefs;
+ /** @todo */
+} RTFSXFSAG;
+/** Pointer to allocation group descriptor data. */
+typedef RTFSXFSAG *PRTFSXFSAG;
+
+
+/**
+ * In-memory inode.
+ */
+typedef struct RTFSXFSINODE
+{
+ /** AVL tree node, indexed by the inode number. */
+ AVLU64NODECORE Core;
+ /** List node for the inode LRU list used for eviction. */
+ RTLISTNODE NdLru;
+ /** Reference counter. */
+ volatile uint32_t cRefs;
+ /** Byte offset in the backing file where the inode is stored.. */
+ uint64_t offInode;
+ /** Inode data. */
+ RTFSOBJINFO ObjInfo;
+ /** Inode data fork format. */
+ uint8_t enmFormat;
+ /** Inode flags. */
+ uint16_t fFlags;
+} RTFSXFSINODE;
+/** Pointer to an in-memory inode. */
+typedef RTFSXFSINODE *PRTFSXFSINODE;
+
+
+/**
+ * Block cache entry.
+ */
+typedef struct RTFSXFSBLOCKENTRY
+{
+ /** AVL tree node, indexed by the filesystem block number. */
+ AVLU64NODECORE Core;
+ /** List node for the inode LRU list used for eviction. */
+ RTLISTNODE NdLru;
+ /** Reference counter. */
+ volatile uint32_t cRefs;
+ /** The block data. */
+ uint8_t abData[1];
+} RTFSXFSBLOCKENTRY;
+/** Pointer to a block cache entry. */
+typedef RTFSXFSBLOCKENTRY *PRTFSXFSBLOCKENTRY;
+
+
+/**
+ * Open directory instance.
+ */
+typedef struct RTFSXFSDIR
+{
+ /** Volume this directory belongs to. */
+ PRTFSXFSVOL pVol;
+ /** The underlying inode structure. */
+ PRTFSXFSINODE pInode;
+ /** Set if we've reached the end of the directory enumeration. */
+ bool fNoMoreFiles;
+ /** Current offset into the directory where the next entry should be read. */
+ uint64_t offEntry;
+ /** Next entry index (for logging purposes). */
+ uint32_t idxEntry;
+} RTFSXFSDIR;
+/** Pointer to an open directory instance. */
+typedef RTFSXFSDIR *PRTFSXFSDIR;
+
+
+/**
+ * Open file instance.
+ */
+typedef struct RTFSXFSFILE
+{
+ /** Volume this directory belongs to. */
+ PRTFSXFSVOL pVol;
+ /** The underlying inode structure. */
+ PRTFSXFSINODE pInode;
+ /** Current offset into the file for I/O. */
+ RTFOFF offFile;
+} RTFSXFSFILE;
+/** Pointer to an open file instance. */
+typedef RTFSXFSFILE *PRTFSXFSFILE;
+
+
+/**
+ * XFS filesystem volume.
+ */
+typedef struct RTFSXFSVOL
+{
+ /** Handle to itself. */
+ RTVFS hVfsSelf;
+ /** The file, partition, or whatever backing the ext volume. */
+ RTVFSFILE hVfsBacking;
+ /** The size of the backing thingy. */
+ uint64_t cbBacking;
+
+ /** RTVFSMNT_F_XXX. */
+ uint32_t fMntFlags;
+ /** RTFSXFSVFS_F_XXX (currently none defined). */
+ uint32_t fXfsFlags;
+
+ /** Size of one sector. */
+ size_t cbSector;
+ /** Size of one block. */
+ size_t cbBlock;
+ /** Number of bits to shift for converting a block number to byte offset. */
+ uint32_t cBlockShift;
+ /** Number of blocks per allocation group. */
+ XFSAGNUMBER cBlocksPerAg;
+ /** Number of blocks per allocation group as log2. */
+ uint32_t cAgBlocksLog;
+ /** Number of allocation groups for this volume. */
+ uint32_t cAgs;
+ /** inode of the root directory. */
+ XFSINO uInodeRoot;
+ /** Inode size in bytes. */
+ size_t cbInode;
+ /** Number of inodes per block. */
+ uint32_t cInodesPerBlock;
+ /** Number of inodes per block as log2. */
+ uint32_t cInodesPerBlockLog;
+
+ /** @name Allocation group cache.
+ * @{ */
+ /** LRU list anchor. */
+ RTLISTANCHOR LstAgLru;
+ /** Root of the cached allocation group tree. */
+ AVLU32TREE AgRoot;
+ /** Size of the cached allocation groups. */
+ size_t cbAgs;
+ /** @} */
+
+ /** @name Inode cache.
+ * @{ */
+ /** LRU list anchor for the inode cache. */
+ RTLISTANCHOR LstInodeLru;
+ /** Root of the cached inode tree. */
+ AVLU64TREE InodeRoot;
+ /** Size of the cached inodes. */
+ size_t cbInodes;
+ /** @} */
+
+ /** @name Block cache.
+ * @{ */
+ /** LRU list anchor for the block cache. */
+ RTLISTANCHOR LstBlockLru;
+ /** Root of the cached block tree. */
+ AVLU64TREE BlockRoot;
+ /** Size of cached blocks. */
+ size_t cbBlocks;
+ /** @} */
+} RTFSXFSVOL;
+
+
+
+/*********************************************************************************************************************************
+* Internal Functions *
+*********************************************************************************************************************************/
+static int rtFsXfsVol_OpenDirByInode(PRTFSXFSVOL pThis, uint32_t iInode, PRTVFSDIR phVfsDir);
+
+#ifdef LOG_ENABLED
+/**
+ * Logs the XFS filesystem superblock.
+ *
+ * @returns nothing.
+ * @param iAg The allocation group number for the given super block.
+ * @param pSb Pointer to the superblock.
+ */
+static void rtFsXfsSb_Log(uint32_t iAg, PCXFSSUPERBLOCK pSb)
+{
+ if (LogIs2Enabled())
+ {
+ Log2(("XFS: Superblock %#RX32:\n", iAg));
+ Log2(("XFS: u32Magic %#RX32\n", RT_BE2H_U32(pSb->u32Magic)));
+ Log2(("XFS: cbBlock %RU32\n", RT_BE2H_U32(pSb->cbBlock)));
+ Log2(("XFS: cBlocks %RU64\n", RT_BE2H_U64(pSb->cBlocks)));
+ Log2(("XFS: cBlocksRtDev %RU64\n", RT_BE2H_U64(pSb->cBlocksRtDev)));
+ Log2(("XFS: cExtentsRtDev %RU64\n", RT_BE2H_U64(pSb->cExtentsRtDev)));
+ Log2(("XFS: abUuid <todo>\n"));
+ Log2(("XFS: uBlockJournal %#RX64\n", RT_BE2H_U64(pSb->uBlockJournal)));
+ Log2(("XFS: uInodeRoot %#RX64\n", RT_BE2H_U64(pSb->uInodeRoot)));
+ Log2(("XFS: uInodeBitmapRtExt %#RX64\n", RT_BE2H_U64(pSb->uInodeBitmapRtExt)));
+ Log2(("XFS: uInodeBitmapSummary %#RX64\n", RT_BE2H_U64(pSb->uInodeBitmapSummary)));
+ Log2(("XFS: cRtExtent %RU32\n", RT_BE2H_U32(pSb->cRtExtent)));
+ Log2(("XFS: cAgBlocks %RU32\n", RT_BE2H_U32(pSb->cAgBlocks)));
+ Log2(("XFS: cAg %RU32\n", RT_BE2H_U32(pSb->cAg)));
+ Log2(("XFS: cRtBitmapBlocks %RU32\n", RT_BE2H_U32(pSb->cRtBitmapBlocks)));
+ Log2(("XFS: cJournalBlocks %RU32\n", RT_BE2H_U32(pSb->cJournalBlocks)));
+ Log2(("XFS: fVersion %#RX16%s%s%s%s%s%s%s%s%s%s%s\n", RT_BE2H_U16(pSb->fVersion),
+ RT_BE2H_U16(pSb->fVersion) & XFS_SB_VERSION_F_ATTR ? " attr" : "",
+ RT_BE2H_U16(pSb->fVersion) & XFS_SB_VERSION_F_NLINK ? " nlink" : "",
+ RT_BE2H_U16(pSb->fVersion) & XFS_SB_VERSION_F_QUOTA ? " quota" : "",
+ RT_BE2H_U16(pSb->fVersion) & XFS_SB_VERSION_F_ALIGN ? " align" : "",
+ RT_BE2H_U16(pSb->fVersion) & XFS_SB_VERSION_F_DALIGN ? " dalign" : "",
+ RT_BE2H_U16(pSb->fVersion) & XFS_SB_VERSION_F_SHARED ? " shared" : "",
+ RT_BE2H_U16(pSb->fVersion) & XFS_SB_VERSION_F_LOGV2 ? " logv2" : "",
+ RT_BE2H_U16(pSb->fVersion) & XFS_SB_VERSION_F_SECTOR ? " sector" : "",
+ RT_BE2H_U16(pSb->fVersion) & XFS_SB_VERSION_F_EXTFLG ? " extflg" : "",
+ RT_BE2H_U16(pSb->fVersion) & XFS_SB_VERSION_F_DIRV2 ? " dirv2" : "",
+ RT_BE2H_U16(pSb->fVersion) & XFS_SB_VERSION_F_FEAT2 ? " feat2" : ""));
+ Log2(("XFS: cbSector %RU16\n", RT_BE2H_U16(pSb->cbSector)));
+ Log2(("XFS: cbInode %RU16\n", RT_BE2H_U16(pSb->cbInode)));
+ Log2(("XFS: cIndoesPerBlock %RU16\n", RT_BE2H_U16(pSb->cInodesPerBlock)));
+ Log2(("XFS: achFsName %12s\n", &pSb->achFsName[0]));
+ Log2(("XFS: cBlockSzLog %RU8\n", pSb->cBlockSzLog));
+ Log2(("XFS: cSectorSzLog %RU8\n", pSb->cSectorSzLog));
+ Log2(("XFS: cInodeSzLog %RU8\n", pSb->cInodeSzLog));
+ Log2(("XFS: cInodesPerBlockLog %RU8\n", pSb->cInodesPerBlockLog));
+ Log2(("XFS: cAgBlocksLog %RU8\n", pSb->cAgBlocksLog));
+ Log2(("XFS: cExtentsRtDevLog %RU8\n", pSb->cExtentsRtDevLog));
+ Log2(("XFS: fInProgress %RU8\n", pSb->fInProgress));
+ Log2(("XFS: cInodeMaxPct %RU8\n", pSb->cInodeMaxPct));
+ Log2(("XFS: cInodesGlobal %#RX64\n", RT_BE2H_U64(pSb->cInodesGlobal)));
+ Log2(("XFS: cInodesGlobalFree %#RX64\n", RT_BE2H_U64(pSb->cInodesGlobalFree)));
+ Log2(("XFS: cBlocksFree %#RX64\n", RT_BE2H_U64(pSb->cBlocksFree)));
+ Log2(("XFS: cExtentsRtFree %#RX64\n", RT_BE2H_U64(pSb->cExtentsRtFree)));
+ Log2(("XFS: uInodeQuotaUsr %#RX64\n", RT_BE2H_U64(pSb->uInodeQuotaUsr)));
+ Log2(("XFS: uInodeQuotaGrp %#RX64\n", RT_BE2H_U64(pSb->uInodeQuotaGrp)));
+ Log2(("XFS: fQuotaFlags %#RX16\n", RT_BE2H_U16(pSb->fQuotaFlags)));
+ Log2(("XFS: fFlagsMisc %#RX8\n", pSb->fFlagsMisc));
+ Log2(("XFS: uSharedVn %#RX8\n", pSb->uSharedVn));
+ Log2(("XFS: cBlocksInodeAlignment %#RX32\n", RT_BE2H_U32(pSb->cBlocksInodeAlignment)));
+ Log2(("XFS: cBlocksRaidStripe %#RX32\n", RT_BE2H_U32(pSb->cBlocksRaidStripe)));
+ Log2(("XFS: cBlocksRaidWidth %#RX32\n", RT_BE2H_U32(pSb->cBlocksRaidWidth)));
+ Log2(("XFS: cDirBlockAllocLog %RU8\n", pSb->cDirBlockAllocLog));
+ Log2(("XFS: cLogDevSubVolSectorSzLog %RU8\n", pSb->cLogDevSubVolSectorSzLog));
+ Log2(("XFS: cLogDevSectorSzLog %RU16\n", RT_BE2H_U16(pSb->cLogDevSectorSzLog)));
+ Log2(("XFS: cLogDevRaidStripe %RU32\n", RT_BE2H_U32(pSb->cLogDevRaidStripe)));
+ Log2(("XFS: fFeatures2 %#RX32\n", RT_BE2H_U32(pSb->fFeatures2)));
+ Log2(("XFS: fFeaturesRw %#RX32\n", RT_BE2H_U32(pSb->fFeaturesRw)));
+ Log2(("XFS: fFeaturesRo %#RX32\n", RT_BE2H_U32(pSb->fFeaturesRo)));
+ Log2(("XFS: fFeaturesIncompatRw %#RX32\n", RT_BE2H_U32(pSb->fFeaturesIncompatRw)));
+ Log2(("XFS: fFeaturesJrnlIncompatRw %#RX32\n", RT_BE2H_U32(pSb->fFeaturesJrnlIncompatRw)));
+ Log2(("XFS: u32Chksum %#RX32\n", RT_BE2H_U32(pSb->u32Chksum)));
+ Log2(("XFS: u32SparseInodeAlignment %#RX32\n", RT_BE2H_U32(pSb->u32SparseInodeAlignment)));
+ Log2(("XFS: uInodeProjectQuota %#RX64\n", RT_BE2H_U64(pSb->uInodeProjectQuota)));
+ Log2(("XFS: uJrnlSeqSbUpdate %#RX64\n", RT_BE2H_U64(pSb->uJrnlSeqSbUpdate)));
+ Log2(("XFS: abUuidMeta <todo>\n"));
+ Log2(("XFS: uInodeRm %#RX64\n", RT_BE2H_U64(pSb->uInodeRm)));
+ }
+}
+
+
+/**
+ * Logs a AG free space block.
+ *
+ * @returns nothing.
+ * @param iAg The allocation group number for the given free space block.
+ * @param pAgf The AG free space block.
+ */
+static void rtFsXfsAgf_Log(uint32_t iAg, PCXFSAGF pAgf)
+{
+ if (LogIs2Enabled())
+ {
+ Log2(("XFS: AGF %#RX32:\n", iAg));
+ Log2(("XFS: u32Magic %#RX32\n", RT_BE2H_U32(pAgf->u32Magic)));
+ Log2(("XFS: uVersion %#RX32\n", RT_BE2H_U32(pAgf->uVersion)));
+ Log2(("XFS: uSeqNo %#RX32\n", RT_BE2H_U32(pAgf->uSeqNo)));
+ Log2(("XFS: cLengthBlocks %#RX32\n", RT_BE2H_U32(pAgf->cLengthBlocks)));
+ Log2(("XFS: auRoots[0] %#RX32\n", RT_BE2H_U32(pAgf->auRoots[0])));
+ Log2(("XFS: auRoots[1] %#RX32\n", RT_BE2H_U32(pAgf->auRoots[1])));
+ Log2(("XFS: auRoots[2] %#RX32\n", RT_BE2H_U32(pAgf->auRoots[2])));
+ Log2(("XFS: acLvls[0] %RU32\n", RT_BE2H_U32(pAgf->acLvls[0])));
+ Log2(("XFS: acLvls[1] %RU32\n", RT_BE2H_U32(pAgf->acLvls[1])));
+ Log2(("XFS: acLvls[2] %RU32\n", RT_BE2H_U32(pAgf->acLvls[2])));
+ Log2(("XFS: idxFreeListFirst %RU32\n", RT_BE2H_U32(pAgf->idxFreeListFirst)));
+ Log2(("XFS: idxFreeListLast %RU32\n", RT_BE2H_U32(pAgf->idxFreeListLast)));
+ Log2(("XFS: cFreeListBlocks %RU32\n", RT_BE2H_U32(pAgf->cFreeListBlocks)));
+ Log2(("XFS: cFreeBlocks %RU32\n", RT_BE2H_U32(pAgf->cFreeBlocks)));
+ Log2(("XFS: cFreeBlocksLongest %RU32\n", RT_BE2H_U32(pAgf->cFreeBlocksLongest)));
+ Log2(("XFS: cBlocksBTrees %RU32\n", RT_BE2H_U32(pAgf->cBlocksBTrees)));
+ Log2(("XFS: abUuid <todo>\n"));
+ Log2(("XFS: cBlocksRevMap %RU32\n", RT_BE2H_U32(pAgf->cBlocksRevMap)));
+ Log2(("XFS: cBlocksRefcountBTree %RU32\n", RT_BE2H_U32(pAgf->cBlocksRefcountBTree)));
+ Log2(("XFS: uRootRefcount %#RX32\n", RT_BE2H_U32(pAgf->uRootRefcount)));
+ Log2(("XFS: cLvlRefcount %RU32\n", RT_BE2H_U32(pAgf->cLvlRefcount)));
+ Log2(("XFS: uSeqNoLastWrite %#RX64\n", RT_BE2H_U64(pAgf->uSeqNoLastWrite)));
+ Log2(("XFS: uChkSum %#RX32\n", RT_BE2H_U32(pAgf->uChkSum)));
+ }
+}
+
+
+/**
+ * Loads an AG inode information block.
+ *
+ * @returns nothing.
+ * @param iAg The allocation group number for the given inode information block.
+ * @param pAgi The AG inode information block.
+ */
+static void rtFsXfsAgi_Log(uint32_t iAg, PCXFSAGI pAgi)
+{
+ if (LogIs2Enabled())
+ {
+ Log2(("XFS: AGI %#RX32:\n", iAg));
+ Log2(("XFS: u32Magic %#RX32\n", RT_BE2H_U32(pAgi->u32Magic)));
+ Log2(("XFS: uVersion %#RX32\n", RT_BE2H_U32(pAgi->uVersion)));
+ Log2(("XFS: uSeqNo %#RX32\n", RT_BE2H_U32(pAgi->uSeqNo)));
+ Log2(("XFS: cLengthBlocks %#RX32\n", RT_BE2H_U32(pAgi->cLengthBlocks)));
+ Log2(("XFS: cInodesAlloc %#RX32\n", RT_BE2H_U32(pAgi->cInodesAlloc)));
+ Log2(("XFS: uRootInode %#RX32\n", RT_BE2H_U32(pAgi->uRootInode)));
+ Log2(("XFS: cLvlsInode %RU32\n", RT_BE2H_U32(pAgi->cLvlsInode)));
+ Log2(("XFS: uInodeNew %#RX32\n", RT_BE2H_U32(pAgi->uInodeNew)));
+ Log2(("XFS: uInodeDir %#RX32\n", RT_BE2H_U32(pAgi->uInodeDir)));
+ Log2(("XFS: au32HashUnlinked[0..63] <todo>\n"));
+ Log2(("XFS: abUuid <todo>\n"));
+ Log2(("XFS: uChkSum %#RX32\n", RT_BE2H_U32(pAgi->uChkSum)));
+ Log2(("XFS: uSeqNoLastWrite %#RX64\n", RT_BE2H_U64(pAgi->uSeqNoLastWrite)));
+ Log2(("XFS: uRootFreeInode %#RX32\n", RT_BE2H_U32(pAgi->uRootFreeInode)));
+ Log2(("XFS: cLvlsFreeInode %RU32\n", RT_BE2H_U32(pAgi->cLvlsFreeInode)));
+ }
+}
+
+
+/**
+ * Logs a XFS filesystem inode.
+ *
+ * @returns nothing.
+ * @param pThis The XFS volume instance.
+ * @param iInode Inode number.
+ * @param pInode Pointer to the inode.
+ */
+static void rtFsXfsInode_Log(PRTFSXFSVOL pThis, XFSINO iInode, PCXFSINODECORE pInode)
+{
+ RT_NOREF(pThis);
+
+ if (LogIs2Enabled())
+ {
+ RTTIMESPEC Spec;
+ char sz[80];
+
+ Log2(("XFS: Inode %#RX64:\n", iInode));
+ Log2(("XFS: u16Magic %#RX16\n", RT_BE2H_U16(pInode->u16Magic)));
+ Log2(("XFS: fMode %#RX16\n", RT_BE2H_U16(pInode->fMode)));
+ Log2(("XFS: iVersion %#RX8\n", pInode->iVersion));
+ Log2(("XFS: enmFormat %#RX8\n", pInode->enmFormat));
+ Log2(("XFS: cOnLinks %RU16\n", RT_BE2H_U16(pInode->cOnLinks)));
+ Log2(("XFS: uUid %#RX32\n", RT_BE2H_U32(pInode->uUid)));
+ Log2(("XFS: uGid %#RX32\n", RT_BE2H_U32(pInode->uGid)));
+ Log2(("XFS: cLinks %#RX32\n", RT_BE2H_U32(pInode->cLinks)));
+ Log2(("XFS: uProjIdLow %#RX16\n", RT_BE2H_U16(pInode->uProjIdLow)));
+ Log2(("XFS: uProjIdHigh %#RX16\n", RT_BE2H_U16(pInode->uProjIdHigh)));
+ Log2(("XFS: cFlush %RU16\n", RT_BE2H_U16(pInode->cFlush)));
+ Log2(("XFS: TsLastAccessed %#RX32:%#RX32 %s\n", RT_BE2H_U32(pInode->TsLastAccessed.cSecEpoch),
+ RT_BE2H_U32(pInode->TsLastAccessed.cNanoSec),
+ RTTimeSpecToString(RTTimeSpecAddNano(RTTimeSpecSetSeconds(&Spec, RT_BE2H_U32(pInode->TsLastAccessed.cSecEpoch)),
+ RT_BE2H_U32(pInode->TsLastAccessed.cNanoSec)),
+ sz, sizeof(sz))));
+ Log2(("XFS: TsLastModified %#RX32:%#RX32 %s\n", RT_BE2H_U32(pInode->TsLastModified.cSecEpoch),
+ RT_BE2H_U32(pInode->TsLastModified.cNanoSec),
+ RTTimeSpecToString(RTTimeSpecAddNano(RTTimeSpecSetSeconds(&Spec, RT_BE2H_U32(pInode->TsLastModified.cSecEpoch)),
+ RT_BE2H_U32(pInode->TsLastModified.cNanoSec)),
+ sz, sizeof(sz))));
+ Log2(("XFS: TsCreatedModified %#RX32:%#RX32 %s\n", RT_BE2H_U32(pInode->TsCreatedModified.cSecEpoch),
+ RT_BE2H_U32(pInode->TsCreatedModified.cNanoSec),
+ RTTimeSpecToString(RTTimeSpecAddNano(RTTimeSpecSetSeconds(&Spec, RT_BE2H_U32(pInode->TsCreatedModified.cSecEpoch)),
+ RT_BE2H_U32(pInode->TsCreatedModified.cNanoSec)),
+ sz, sizeof(sz))));
+ Log2(("XFS: cbInode %#RX64\n", RT_BE2H_U64(pInode->cbInode)));
+ Log2(("XFS: cBlocks %#RX64\n", RT_BE2H_U64(pInode->cBlocks)));
+ Log2(("XFS: cExtentBlocksMin %#RX32\n", RT_BE2H_U32(pInode->cExtentBlocksMin)));
+ Log2(("XFS: cExtentsData %#RX32\n", RT_BE2H_U32(pInode->cExtentsData)));
+ Log2(("XFS: cExtentsAttr %#RX16\n", RT_BE2H_U16(pInode->cExtentsAttr)));
+ Log2(("XFS: offAttrFork %#RX8\n", pInode->offAttrFork));
+ Log2(("XFS: enmFormatAttr %#RX8\n", pInode->enmFormatAttr));
+ Log2(("XFS: fEvtMaskDmig %#RX32\n", RT_BE2H_U32(pInode->fEvtMaskDmig)));
+ Log2(("XFS: uStateDmig %#RX16\n", RT_BE2H_U16(pInode->uStateDmig)));
+ Log2(("XFS: fFlags %#RX16\n", RT_BE2H_U16(pInode->fFlags)));
+ Log2(("XFS: cGeneration %#RX32\n", RT_BE2H_U32(pInode->cGeneration)));
+ Log2(("XFS: offBlockUnlinkedNext %#RX32\n", RT_BE2H_U32(pInode->offBlockUnlinkedNext)));
+ Log2(("XFS: uChkSum %#RX32\n", RT_BE2H_U32(pInode->uChkSum)));
+ Log2(("XFS: cAttrChanges %#RX64\n", RT_BE2H_U64(pInode->cAttrChanges)));
+ Log2(("XFS: uFlushSeqNo %#RX64\n", RT_BE2H_U64(pInode->uFlushSeqNo)));
+ Log2(("XFS: fFlags2 %#RX64\n", RT_BE2H_U64(pInode->fFlags2)));
+ Log2(("XFS: cExtentCowMin %#RX32\n", RT_BE2H_U32(pInode->cExtentCowMin)));
+ Log2(("XFS: TsCreation %#RX32:%#RX32 %s\n", RT_BE2H_U32(pInode->TsCreation.cSecEpoch),
+ RT_BE2H_U32(pInode->TsCreation.cNanoSec),
+ RTTimeSpecToString(RTTimeSpecAddNano(RTTimeSpecSetSeconds(&Spec, RT_BE2H_U32(pInode->TsCreation.cSecEpoch)),
+ RT_BE2H_U32(pInode->TsCreation.cNanoSec)),
+ sz, sizeof(sz))));
+ Log2(("XFS: uInode %#RX64\n", RT_BE2H_U64(pInode->uInode)));
+ Log2(("XFS: abUuid <todo>\n"));
+ }
+}
+
+
+#if 0
+/**
+ * Logs a XFS filesystem directory entry.
+ *
+ * @returns nothing.
+ * @param pThis The XFS volume instance.
+ * @param idxDirEntry Directory entry index number.
+ * @param pDirEntry The directory entry.
+ */
+static void rtFsXfsDirEntry_Log(PRTFSXFSVOL pThis, uint32_t idxDirEntry, PCXFSDIRENTRYEX pDirEntry)
+{
+ if (LogIs2Enabled())
+ {
+ }
+}
+#endif
+#endif
+
+
+/**
+ * Converts a block number to a byte offset.
+ *
+ * @returns Offset in bytes for the given block number.
+ * @param pThis The XFS volume instance.
+ * @param iBlock The block number to convert.
+ */
+DECLINLINE(uint64_t) rtFsXfsBlockIdxToDiskOffset(PRTFSXFSVOL pThis, uint64_t iBlock)
+{
+ return iBlock << pThis->cBlockShift;
+}
+
+
+/**
+ * Converts a byte offset to a block number.
+ *
+ * @returns Block number.
+ * @param pThis The XFS volume instance.
+ * @param iBlock The offset to convert.
+ */
+DECLINLINE(uint64_t) rtFsXfsDiskOffsetToBlockIdx(PRTFSXFSVOL pThis, uint64_t off)
+{
+ return off >> pThis->cBlockShift;
+}
+
+
+/**
+ * Splits the given absolute inode number into the AG number, block inside the AG
+ * and the offset into the block where to find the inode structure.
+ *
+ * @returns nothing.
+ * @param pThis The XFS volume instance.
+ * @param iInode The inode to split.
+ * @param piAg Where to store the AG number.
+ * @param puBlock Where to store the block number inside the AG.
+ * @param poffBlock Where to store the offset into the block.
+ */
+DECLINLINE(void) rtFsXfsInodeSplitAbs(PRTFSXFSVOL pThis, XFSINO iInode,
+ uint32_t *piAg, uint32_t *puBlock,
+ uint32_t *poffBlock)
+{
+ *poffBlock = iInode & (pThis->cInodesPerBlock - 1);
+ iInode >>= pThis->cInodesPerBlockLog;
+ *puBlock = iInode & (RT_BIT_32(pThis->cAgBlocksLog) - 1); /* Using the log2 value here as it is rounded. */
+ iInode >>= RT_BIT_32(pThis->cAgBlocksLog) - 1;
+ *piAg = (uint32_t)iInode;
+}
+
+
+/**
+ * Allocates a new block group.
+ *
+ * @returns Pointer to the new block group descriptor or NULL if out of memory.
+ * @param pThis The XFS volume instance.
+ * @param cbAlloc How much to allocate.
+ * @param iBlockGroup Block group number.
+ */
+static PRTFSXFSBLOCKENTRY rtFsXfsVol_BlockAlloc(PRTFSXFSVOL pThis, size_t cbAlloc, uint64_t iBlock)
+{
+ PRTFSXFSBLOCKENTRY pBlock = (PRTFSXFSBLOCKENTRY)RTMemAllocZ(cbAlloc);
+ if (RT_LIKELY(pBlock))
+ {
+ pBlock->Core.Key = iBlock;
+ pBlock->cRefs = 0;
+ pThis->cbBlocks += cbAlloc;
+ }
+
+ return pBlock;
+}
+
+
+/**
+ * Returns a new block entry utilizing the cache if possible.
+ *
+ * @returns Pointer to the new block entry or NULL if out of memory.
+ * @param pThis The XFS volume instance.
+ * @param iBlock Block number.
+ */
+static PRTFSXFSBLOCKENTRY rtFsXfsVol_BlockGetNew(PRTFSXFSVOL pThis, uint64_t iBlock)
+{
+ PRTFSXFSBLOCKENTRY pBlock = NULL;
+ size_t cbAlloc = RT_UOFFSETOF_DYN(RTFSXFSBLOCKENTRY, abData[pThis->cbBlock]);
+ if (pThis->cbBlocks + cbAlloc <= RTFSXFS_MAX_BLOCK_CACHE_SIZE)
+ pBlock = rtFsXfsVol_BlockAlloc(pThis, cbAlloc, iBlock);
+ else
+ {
+ pBlock = RTListRemoveLast(&pThis->LstBlockLru, RTFSXFSBLOCKENTRY, NdLru);
+ if (!pBlock)
+ pBlock = rtFsXfsVol_BlockAlloc(pThis, cbAlloc, iBlock);
+ else
+ {
+ /* Remove the block group from the tree because it gets a new key. */
+ PAVLU64NODECORE pCore = RTAvlU64Remove(&pThis->BlockRoot, pBlock->Core.Key);
+ Assert(pCore == &pBlock->Core); RT_NOREF(pCore);
+ }
+ }
+
+ Assert(!pBlock->cRefs);
+ pBlock->Core.Key = iBlock;
+ pBlock->cRefs = 1;
+
+ return pBlock;
+}
+
+
+/**
+ * Frees the given block.
+ *
+ * @returns nothing.
+ * @param pThis The XFS volume instance.
+ * @param pBlock The block to free.
+ */
+static void rtFsXfsVol_BlockFree(PRTFSXFSVOL pThis, PRTFSXFSBLOCKENTRY pBlock)
+{
+ Assert(!pBlock->cRefs);
+
+ /*
+ * Put it into the cache if the limit wasn't exceeded, otherwise the block group
+ * is freed right away.
+ */
+ if (pThis->cbBlocks <= RTFSXFS_MAX_BLOCK_CACHE_SIZE)
+ {
+ /* Put onto the LRU list. */
+ RTListPrepend(&pThis->LstBlockLru, &pBlock->NdLru);
+ }
+ else
+ {
+ /* Remove from the tree and free memory. */
+ PAVLU64NODECORE pCore = RTAvlU64Remove(&pThis->BlockRoot, pBlock->Core.Key);
+ Assert(pCore == &pBlock->Core); RT_NOREF(pCore);
+ RTMemFree(pBlock);
+ pThis->cbBlocks -= RT_UOFFSETOF_DYN(RTFSXFSBLOCKENTRY, abData[pThis->cbBlock]);
+ }
+}
+
+
+/**
+ * Gets the specified block data from the volume.
+ *
+ * @returns IPRT status code.
+ * @param pThis The XFS volume instance.
+ * @param iBlock The filesystem block to load.
+ * @param ppBlock Where to return the pointer to the block entry on success.
+ * @param ppvData Where to return the pointer to the block data on success.
+ */
+static int rtFsXfsVol_BlockLoad(PRTFSXFSVOL pThis, uint64_t iBlock, PRTFSXFSBLOCKENTRY *ppBlock, void **ppvData)
+{
+ int rc = VINF_SUCCESS;
+
+ /* Try to fetch the block group from the cache first. */
+ PRTFSXFSBLOCKENTRY pBlock = (PRTFSXFSBLOCKENTRY)RTAvlU64Get(&pThis->BlockRoot, iBlock);
+ if (!pBlock)
+ {
+ /* Slow path, load from disk. */
+ pBlock = rtFsXfsVol_BlockGetNew(pThis, iBlock);
+ if (RT_LIKELY(pBlock))
+ {
+ uint64_t offRead = rtFsXfsBlockIdxToDiskOffset(pThis, iBlock);
+ rc = RTVfsFileReadAt(pThis->hVfsBacking, offRead, &pBlock->abData[0], pThis->cbBlock, NULL);
+ if (RT_SUCCESS(rc))
+ {
+ bool fIns = RTAvlU64Insert(&pThis->BlockRoot, &pBlock->Core);
+ Assert(fIns); RT_NOREF(fIns);
+ }
+ }
+ else
+ rc = VERR_NO_MEMORY;
+ }
+ else
+ {
+ /* Remove from current LRU list position and add to the beginning. */
+ uint32_t cRefs = ASMAtomicIncU32(&pBlock->cRefs);
+ if (cRefs == 1) /* Blocks get removed from the LRU list if they are referenced. */
+ RTListNodeRemove(&pBlock->NdLru);
+ }
+
+ if (RT_SUCCESS(rc))
+ {
+ *ppBlock = pBlock;
+ *ppvData = &pBlock->abData[0];
+ }
+ else if (pBlock)
+ {
+ ASMAtomicDecU32(&pBlock->cRefs);
+ rtFsXfsVol_BlockFree(pThis, pBlock); /* Free the block. */
+ }
+
+ return rc;
+}
+
+
+/**
+ * Releases a reference of the given block.
+ *
+ * @returns nothing.
+ * @param pThis The XFS volume instance.
+ * @param pBlock The block to release.
+ */
+static void rtFsXfsVol_BlockRelease(PRTFSXFSVOL pThis, PRTFSXFSBLOCKENTRY pBlock)
+{
+ uint32_t cRefs = ASMAtomicDecU32(&pBlock->cRefs);
+ if (!cRefs)
+ rtFsXfsVol_BlockFree(pThis, pBlock);
+}
+
+
+/**
+ * Allocates a new alloction group.
+ *
+ * @returns Pointer to the new allocation group descriptor or NULL if out of memory.
+ * @param pThis The XFS volume instance.
+ * @param iAG Allocation group number.
+ */
+static PRTFSXFSAG rtFsXfsAg_Alloc(PRTFSXFSVOL pThis, uint32_t iAg)
+{
+ PRTFSXFSAG pAg = (PRTFSXFSAG)RTMemAllocZ(sizeof(RTFSXFSAG));
+ if (RT_LIKELY(pAg))
+ {
+ pAg->Core.Key = iAg;
+ pAg->cRefs = 0;
+ pThis->cbAgs += sizeof(RTFSXFSAG);
+ }
+
+ return pAg;
+}
+
+
+/**
+ * Frees the given allocation group.
+ *
+ * @returns nothing.
+ * @param pThis The XFS volume instance.
+ * @param pAg The allocation group to free.
+ */
+static void rtFsXfsAg_Free(PRTFSXFSVOL pThis, PRTFSXFSAG pAg)
+{
+ Assert(!pAg->cRefs);
+
+ /*
+ * Put it into the cache if the limit wasn't exceeded, otherwise the allocation group
+ * is freed right away.
+ */
+ if (pThis->cbAgs <= RTFSXFS_MAX_AG_CACHE_SIZE)
+ {
+ /* Put onto the LRU list. */
+ RTListPrepend(&pThis->LstAgLru, &pAg->NdLru);
+ }
+ else
+ {
+ /* Remove from the tree and free memory. */
+ PAVLU32NODECORE pCore = RTAvlU32Remove(&pThis->AgRoot, pAg->Core.Key);
+ Assert(pCore == &pAg->Core); RT_NOREF(pCore);
+ RTMemFree(pAg);
+ pThis->cbAgs -= sizeof(RTFSXFSAG);
+ }
+}
+
+
+/**
+ * Returns a new block group utilizing the cache if possible.
+ *
+ * @returns Pointer to the new block group descriptor or NULL if out of memory.
+ * @param pThis The XFS volume instance.
+ * @param iAg Allocation group number.
+ */
+static PRTFSXFSAG rtFsXfsAg_GetNew(PRTFSXFSVOL pThis, uint32_t iAg)
+{
+ PRTFSXFSAG pAg = NULL;
+ if (pThis->cbAgs + sizeof(RTFSXFSAG) <= RTFSXFS_MAX_AG_CACHE_SIZE)
+ pAg = rtFsXfsAg_Alloc(pThis, iAg);
+ else
+ {
+ pAg = RTListRemoveLast(&pThis->LstAgLru, RTFSXFSAG, NdLru);
+ if (!pAg)
+ pAg = rtFsXfsAg_Alloc(pThis, iAg);
+ else
+ {
+ /* Remove the block group from the tree because it gets a new key. */
+ PAVLU32NODECORE pCore = RTAvlU32Remove(&pThis->AgRoot, pAg->Core.Key);
+ Assert(pCore == &pAg->Core); RT_NOREF(pCore);
+ }
+ }
+
+ Assert(!pAg->cRefs);
+ pAg->Core.Key = iAg;
+ pAg->cRefs = 1;
+
+ return pAg;
+}
+
+
+/**
+ * Loads the given allocation group number and returns it on success.
+ *
+ * @returns IPRT status code.
+ * @param pThis The XFS volume instance.
+ * @param iAg The allocation group to load.
+ * @param ppAg Where to store the allocation group on success.
+ */
+static int rtFsXfsAg_Load(PRTFSXFSVOL pThis, uint32_t iAg, PRTFSXFSAG *ppAg)
+{
+ int rc = VINF_SUCCESS;
+
+ AssertReturn(iAg < pThis->cAgs, VERR_VFS_BOGUS_FORMAT);
+
+ /* Try to fetch the allocation group from the cache first. */
+ PRTFSXFSAG pAg = (PRTFSXFSAG)RTAvlU32Get(&pThis->AgRoot, iAg);
+ if (!pAg)
+ {
+ /* Slow path, load from disk. */
+ pAg = rtFsXfsAg_GetNew(pThis, iAg);
+ if (RT_LIKELY(pAg))
+ {
+ uint64_t offRead = rtFsXfsBlockIdxToDiskOffset(pThis, iAg * pThis->cBlocksPerAg);
+ XFSSUPERBLOCK Sb;
+ rc = RTVfsFileReadAt(pThis->hVfsBacking, offRead, &Sb, sizeof(Sb), NULL);
+ if (RT_SUCCESS(rc))
+ {
+#ifdef LOG_ENABLED
+ rtFsXfsSb_Log(iAg, &Sb);
+#endif
+ }
+ }
+ else
+ rc = VERR_NO_MEMORY;
+ }
+ else
+ {
+ /* Remove from current LRU list position and add to the beginning. */
+ uint32_t cRefs = ASMAtomicIncU32(&pAg->cRefs);
+ if (cRefs == 1) /* Block groups get removed from the LRU list if they are referenced. */
+ RTListNodeRemove(&pAg->NdLru);
+ }
+
+ if (RT_SUCCESS(rc))
+ *ppAg = pAg;
+ else if (pAg)
+ {
+ ASMAtomicDecU32(&pAg->cRefs);
+ rtFsXfsAg_Free(pThis, pAg); /* Free the allocation group. */
+ }
+
+ return rc;
+}
+
+
+/**
+ * Releases a reference of the given allocation group.
+ *
+ * @returns nothing.
+ * @param pThis The XFS volume instance.
+ * @param pAg The allocation group to release.
+ */
+static void rtFsXfsAg_Release(PRTFSXFSVOL pThis, PRTFSXFSAG pAg)
+{
+ uint32_t cRefs = ASMAtomicDecU32(&pAg->cRefs);
+ if (!cRefs)
+ rtFsXfsAg_Free(pThis, pAg);
+}
+
+
+/**
+ * Allocates a new inode.
+ *
+ * @returns Pointer to the new inode or NULL if out of memory.
+ * @param pThis The XFS volume instance.
+ * @param iInode Inode number.
+ */
+static PRTFSXFSINODE rtFsXfsInode_Alloc(PRTFSXFSVOL pThis, uint32_t iInode)
+{
+ PRTFSXFSINODE pInode = (PRTFSXFSINODE)RTMemAllocZ(sizeof(RTFSXFSINODE));
+ if (RT_LIKELY(pInode))
+ {
+ pInode->Core.Key = iInode;
+ pInode->cRefs = 0;
+ pThis->cbInodes += sizeof(RTFSXFSINODE);
+ }
+
+ return pInode;
+}
+
+
+/**
+ * Frees the given inode.
+ *
+ * @returns nothing.
+ * @param pThis The XFS volume instance.
+ * @param pInode The inode to free.
+ */
+static void rtFsXfsInode_Free(PRTFSXFSVOL pThis, PRTFSXFSINODE pInode)
+{
+ Assert(!pInode->cRefs);
+
+ /*
+ * Put it into the cache if the limit wasn't exceeded, otherwise the inode
+ * is freed right away.
+ */
+ if (pThis->cbInodes <= RTFSXFS_MAX_INODE_CACHE_SIZE)
+ {
+ /* Put onto the LRU list. */
+ RTListPrepend(&pThis->LstInodeLru, &pInode->NdLru);
+ }
+ else
+ {
+ /* Remove from the tree and free memory. */
+ PAVLU64NODECORE pCore = RTAvlU64Remove(&pThis->InodeRoot, pInode->Core.Key);
+ Assert(pCore == &pInode->Core); RT_NOREF(pCore);
+ RTMemFree(pInode);
+ pThis->cbInodes -= sizeof(RTFSXFSINODE);
+ }
+}
+
+
+/**
+ * Returns a new inodep utilizing the cache if possible.
+ *
+ * @returns Pointer to the new inode or NULL if out of memory.
+ * @param pThis The XFS volume instance.
+ * @param iInode Inode number.
+ */
+static PRTFSXFSINODE rtFsXfsInode_GetNew(PRTFSXFSVOL pThis, XFSINO iInode)
+{
+ PRTFSXFSINODE pInode = NULL;
+ if (pThis->cbInodes + sizeof(RTFSXFSINODE) <= RTFSXFS_MAX_INODE_CACHE_SIZE)
+ pInode = rtFsXfsInode_Alloc(pThis, iInode);
+ else
+ {
+ pInode = RTListRemoveLast(&pThis->LstInodeLru, RTFSXFSINODE, NdLru);
+ if (!pInode)
+ pInode = rtFsXfsInode_Alloc(pThis, iInode);
+ else
+ {
+ /* Remove the block group from the tree because it gets a new key. */
+ PAVLU64NODECORE pCore = RTAvlU64Remove(&pThis->InodeRoot, pInode->Core.Key);
+ Assert(pCore == &pInode->Core); RT_NOREF(pCore);
+ }
+ }
+
+ Assert(!pInode->cRefs);
+ pInode->Core.Key = iInode;
+ pInode->cRefs = 1;
+
+ return pInode;
+}
+
+
+/**
+ * Loads the given inode number and returns it on success.
+ *
+ * @returns IPRT status code.
+ * @param pThis The XFS volume instance.
+ * @param iInode The inode to load.
+ * @param ppInode Where to store the inode on success.
+ */
+static int rtFsXfsInode_Load(PRTFSXFSVOL pThis, XFSINO iInode, PRTFSXFSINODE *ppInode)
+{
+ int rc = VINF_SUCCESS;
+
+ /* Try to fetch the inode from the cache first. */
+ PRTFSXFSINODE pInode = (PRTFSXFSINODE)RTAvlU64Get(&pThis->InodeRoot, iInode);
+ if (!pInode)
+ {
+ /* Slow path, load from disk. */
+ pInode = rtFsXfsInode_GetNew(pThis, iInode);
+ if (RT_LIKELY(pInode))
+ {
+ uint32_t iAg;
+ uint32_t uBlock;
+ uint32_t offBlock;
+
+ rtFsXfsInodeSplitAbs(pThis, iInode, &iAg, &uBlock, &offBlock);
+
+ uint64_t offRead = (iAg * pThis->cBlocksPerAg + uBlock) * pThis->cbBlock + offBlock;
+ XFSINODECORE Inode;
+ rc = RTVfsFileReadAt(pThis->hVfsBacking, offRead, &Inode, RT_MIN(sizeof(Inode), pThis->cbInode), NULL);
+ if (RT_SUCCESS(rc))
+ {
+#ifdef LOG_ENABLED
+ rtFsXfsInode_Log(pThis, iInode, &Inode);
+#endif
+
+ pInode->offInode = offRead;
+ pInode->fFlags = RT_BE2H_U16(Inode.fFlags);
+ pInode->enmFormat = Inode.enmFormat;
+ pInode->ObjInfo.cbObject = RT_BE2H_U64(Inode.cbInode);
+ pInode->ObjInfo.cbAllocated = RT_BE2H_U64(Inode.cBlocks) * pThis->cbBlock;
+ RTTimeSpecSetSeconds(&pInode->ObjInfo.AccessTime, RT_BE2H_U32(Inode.TsLastAccessed.cSecEpoch));
+ RTTimeSpecAddNano(&pInode->ObjInfo.AccessTime, RT_BE2H_U32(Inode.TsLastAccessed.cNanoSec));
+ RTTimeSpecSetSeconds(&pInode->ObjInfo.ModificationTime, RT_BE2H_U32(Inode.TsLastModified.cSecEpoch));
+ RTTimeSpecAddNano(&pInode->ObjInfo.ModificationTime, RT_BE2H_U32(Inode.TsLastModified.cNanoSec));
+ RTTimeSpecSetSeconds(&pInode->ObjInfo.ChangeTime, RT_BE2H_U32(Inode.TsCreatedModified.cSecEpoch));
+ RTTimeSpecAddNano(&pInode->ObjInfo.ChangeTime, RT_BE2H_U32(Inode.TsCreatedModified.cNanoSec));
+ pInode->ObjInfo.Attr.enmAdditional = RTFSOBJATTRADD_UNIX;
+ pInode->ObjInfo.Attr.u.Unix.uid = RT_BE2H_U32(Inode.uUid);
+ pInode->ObjInfo.Attr.u.Unix.gid = RT_BE2H_U32(Inode.uGid);
+ pInode->ObjInfo.Attr.u.Unix.cHardlinks = RT_BE2H_U16(Inode.cOnLinks); /** @todo v2 inodes. */
+ pInode->ObjInfo.Attr.u.Unix.INodeIdDevice = 0;
+ pInode->ObjInfo.Attr.u.Unix.INodeId = iInode;
+ pInode->ObjInfo.Attr.u.Unix.fFlags = 0;
+ pInode->ObjInfo.Attr.u.Unix.GenerationId = RT_BE2H_U32(Inode.cGeneration);
+ pInode->ObjInfo.Attr.u.Unix.Device = 0;
+ if (Inode.iVersion >= 3)
+ {
+ RTTimeSpecSetSeconds(&pInode->ObjInfo.BirthTime, RT_BE2H_U32(Inode.TsCreation.cSecEpoch));
+ RTTimeSpecAddNano(&pInode->ObjInfo.BirthTime, RT_BE2H_U32(Inode.TsCreation.cNanoSec));
+ }
+ else
+ pInode->ObjInfo.BirthTime = pInode->ObjInfo.ChangeTime;
+
+ /* Fill in the mode. */
+ pInode->ObjInfo.Attr.fMode = 0;
+ uint16_t fInodeMode = RT_BE2H_U16(Inode.fMode);
+ switch (XFS_INODE_MODE_TYPE_GET_TYPE(fInodeMode))
+ {
+ case XFS_INODE_MODE_TYPE_FIFO:
+ pInode->ObjInfo.Attr.fMode |= RTFS_TYPE_FIFO;
+ break;
+ case XFS_INODE_MODE_TYPE_CHAR:
+ pInode->ObjInfo.Attr.fMode |= RTFS_TYPE_DEV_CHAR;
+ break;
+ case XFS_INODE_MODE_TYPE_DIR:
+ pInode->ObjInfo.Attr.fMode |= RTFS_TYPE_DIRECTORY;
+ break;
+ case XFS_INODE_MODE_TYPE_BLOCK:
+ pInode->ObjInfo.Attr.fMode |= RTFS_TYPE_DEV_BLOCK;
+ break;
+ case XFS_INODE_MODE_TYPE_REGULAR:
+ pInode->ObjInfo.Attr.fMode |= RTFS_TYPE_FILE;
+ break;
+ case XFS_INODE_MODE_TYPE_SYMLINK:
+ pInode->ObjInfo.Attr.fMode |= RTFS_TYPE_SYMLINK;
+ break;
+ case XFS_INODE_MODE_TYPE_SOCKET:
+ pInode->ObjInfo.Attr.fMode |= RTFS_TYPE_SOCKET;
+ break;
+ default:
+ rc = VERR_VFS_BOGUS_FORMAT;
+ }
+ if (fInodeMode & XFS_INODE_MODE_EXEC_OTHER)
+ pInode->ObjInfo.Attr.fMode |= RTFS_UNIX_IXOTH;
+ if (fInodeMode & XFS_INODE_MODE_WRITE_OTHER)
+ pInode->ObjInfo.Attr.fMode |= RTFS_UNIX_IWOTH;
+ if (fInodeMode & XFS_INODE_MODE_READ_OTHER)
+ pInode->ObjInfo.Attr.fMode |= RTFS_UNIX_IROTH;
+ if (fInodeMode & XFS_INODE_MODE_EXEC_GROUP)
+ pInode->ObjInfo.Attr.fMode |= RTFS_UNIX_IXGRP;
+ if (fInodeMode & XFS_INODE_MODE_WRITE_GROUP)
+ pInode->ObjInfo.Attr.fMode |= RTFS_UNIX_IWGRP;
+ if (fInodeMode & XFS_INODE_MODE_READ_GROUP)
+ pInode->ObjInfo.Attr.fMode |= RTFS_UNIX_IRGRP;
+ if (fInodeMode & XFS_INODE_MODE_EXEC_OWNER)
+ pInode->ObjInfo.Attr.fMode |= RTFS_UNIX_IXUSR;
+ if (fInodeMode & XFS_INODE_MODE_WRITE_OWNER)
+ pInode->ObjInfo.Attr.fMode |= RTFS_UNIX_IWUSR;
+ if (fInodeMode & XFS_INODE_MODE_READ_OWNER)
+ pInode->ObjInfo.Attr.fMode |= RTFS_UNIX_IRUSR;
+ if (fInodeMode & XFS_INODE_MODE_STICKY)
+ pInode->ObjInfo.Attr.fMode |= RTFS_UNIX_ISTXT;
+ if (fInodeMode & XFS_INODE_MODE_SET_GROUP_ID)
+ pInode->ObjInfo.Attr.fMode |= RTFS_UNIX_ISGID;
+ if (fInodeMode & XFS_INODE_MODE_SET_USER_ID)
+ pInode->ObjInfo.Attr.fMode |= RTFS_UNIX_ISUID;
+ }
+ }
+ else
+ rc = VERR_NO_MEMORY;
+ }
+ else
+ {
+ /* Remove from current LRU list position and add to the beginning. */
+ uint32_t cRefs = ASMAtomicIncU32(&pInode->cRefs);
+ if (cRefs == 1) /* Inodes get removed from the LRU list if they are referenced. */
+ RTListNodeRemove(&pInode->NdLru);
+ }
+
+ if (RT_SUCCESS(rc))
+ *ppInode = pInode;
+ else if (pInode)
+ {
+ ASMAtomicDecU32(&pInode->cRefs);
+ rtFsXfsInode_Free(pThis, pInode); /* Free the inode. */
+ }
+
+ return rc;
+}
+
+
+/**
+ * Releases a reference of the given inode.
+ *
+ * @returns nothing.
+ * @param pThis The XFS volume instance.
+ * @param pInode The inode to release.
+ */
+static void rtFsXfsInode_Release(PRTFSXFSVOL pThis, PRTFSXFSINODE pInode)
+{
+ uint32_t cRefs = ASMAtomicDecU32(&pInode->cRefs);
+ if (!cRefs)
+ rtFsXfsInode_Free(pThis, pInode);
+}
+
+
+/**
+ * Worker for various QueryInfo methods.
+ *
+ * @returns IPRT status code.
+ * @param pInode The inode structure to return info for.
+ * @param pObjInfo Where to return object info.
+ * @param enmAddAttr What additional info to return.
+ */
+static int rtFsXfsInode_QueryInfo(PRTFSXFSINODE pInode, PRTFSOBJINFO pObjInfo, RTFSOBJATTRADD enmAddAttr)
+{
+ RT_ZERO(*pObjInfo);
+
+ pObjInfo->cbObject = pInode->ObjInfo.cbObject;
+ pObjInfo->cbAllocated = pInode->ObjInfo.cbAllocated;
+ pObjInfo->AccessTime = pInode->ObjInfo.AccessTime;
+ pObjInfo->ModificationTime = pInode->ObjInfo.ModificationTime;
+ pObjInfo->ChangeTime = pInode->ObjInfo.ChangeTime;
+ pObjInfo->BirthTime = pInode->ObjInfo.BirthTime;
+ pObjInfo->Attr.fMode = pInode->ObjInfo.Attr.fMode;
+ pObjInfo->Attr.enmAdditional = enmAddAttr;
+ switch (enmAddAttr)
+ {
+ case RTFSOBJATTRADD_UNIX:
+ memcpy(&pObjInfo->Attr.u.Unix, &pInode->ObjInfo.Attr.u.Unix, sizeof(pInode->ObjInfo.Attr.u.Unix));
+ break;
+
+ case RTFSOBJATTRADD_UNIX_OWNER:
+ pObjInfo->Attr.u.UnixOwner.uid = pInode->ObjInfo.Attr.u.Unix.uid;
+ break;
+
+ case RTFSOBJATTRADD_UNIX_GROUP:
+ pObjInfo->Attr.u.UnixGroup.gid = pInode->ObjInfo.Attr.u.Unix.gid;
+ break;
+
+ default:
+ break;
+ }
+
+ return VINF_SUCCESS;
+}
+
+
+/**
+ * Maps the given inode block to the destination filesystem block.
+ *
+ * @returns IPRT status code.
+ * @param pThis The XFS volume instance.
+ * @param pInode The inode structure to read from.
+ * @param iBlock The inode block to map.
+ * @param cBlocks Number of blocks requested.
+ * @param piBlockFs Where to store the filesystem block on success.
+ * @param pcBlocks Where to store the number of contiguous blocks on success.
+ * @param pfSparse Where to store the sparse flag on success.
+ *
+ * @todo Optimize
+ */
+static int rtFsXfsInode_MapBlockToFs(PRTFSXFSVOL pThis, PRTFSXFSINODE pInode, uint64_t iBlock, size_t cBlocks,
+ uint64_t *piBlockFs, size_t *pcBlocks, bool *pfSparse)
+{
+ RT_NOREF(pThis, pInode, iBlock, cBlocks, piBlockFs, pcBlocks, pfSparse);
+ return VERR_NOT_IMPLEMENTED;
+}
+
+
+/**
+ * Reads data from the given inode at the given byte offset.
+ *
+ * @returns IPRT status code.
+ * @param pThis The XFS volume instance.
+ * @param pInode The inode structure to read from.
+ * @param off The byte offset to start reading from.
+ * @param pvBuf Where to store the read data to.
+ * @param pcbRead Where to return the amount of data read.
+ */
+static int rtFsXfsInode_Read(PRTFSXFSVOL pThis, PRTFSXFSINODE pInode, uint64_t off, void *pvBuf, size_t cbRead, size_t *pcbRead)
+{
+ int rc = VINF_SUCCESS;
+ uint8_t *pbBuf = (uint8_t *)pvBuf;
+
+ if (((uint64_t)pInode->ObjInfo.cbObject < off + cbRead))
+ {
+ if (!pcbRead)
+ return VERR_EOF;
+ else
+ cbRead = (uint64_t)pInode->ObjInfo.cbObject - off;
+ }
+
+ while ( cbRead
+ && RT_SUCCESS(rc))
+ {
+ uint64_t iBlockStart = rtFsXfsDiskOffsetToBlockIdx(pThis, off);
+ uint32_t offBlockStart = off % pThis->cbBlock;
+
+ /* Resolve the inode block to the proper filesystem block. */
+ uint64_t iBlockFs = 0;
+ size_t cBlocks = 0;
+ bool fSparse = false;
+ rc = rtFsXfsInode_MapBlockToFs(pThis, pInode, iBlockStart, 1, &iBlockFs, &cBlocks, &fSparse);
+ if (RT_SUCCESS(rc))
+ {
+ Assert(cBlocks == 1);
+
+ size_t cbThisRead = RT_MIN(cbRead, pThis->cbBlock - offBlockStart);
+
+ if (!fSparse)
+ {
+ uint64_t offRead = rtFsXfsBlockIdxToDiskOffset(pThis, iBlockFs);
+ rc = RTVfsFileReadAt(pThis->hVfsBacking, offRead + offBlockStart, pbBuf, cbThisRead, NULL);
+ }
+ else
+ memset(pbBuf, 0, cbThisRead);
+
+ if (RT_SUCCESS(rc))
+ {
+ pbBuf += cbThisRead;
+ cbRead -= cbThisRead;
+ off += cbThisRead;
+ if (pcbRead)
+ *pcbRead += cbThisRead;
+ }
+ }
+ }
+
+ return rc;
+}
+
+
+
+/*
+ *
+ * File operations.
+ * File operations.
+ * File operations.
+ *
+ */
+
+/**
+ * @interface_method_impl{RTVFSOBJOPS,pfnClose}
+ */
+static DECLCALLBACK(int) rtFsXfsFile_Close(void *pvThis)
+{
+ PRTFSXFSFILE pThis = (PRTFSXFSFILE)pvThis;
+ LogFlow(("rtFsXfsFile_Close(%p/%p)\n", pThis, pThis->pInode));
+
+ rtFsXfsInode_Release(pThis->pVol, pThis->pInode);
+ pThis->pInode = NULL;
+ pThis->pVol = NULL;
+ return VINF_SUCCESS;
+}
+
+
+/**
+ * @interface_method_impl{RTVFSOBJOPS,pfnQueryInfo}
+ */
+static DECLCALLBACK(int) rtFsXfsFile_QueryInfo(void *pvThis, PRTFSOBJINFO pObjInfo, RTFSOBJATTRADD enmAddAttr)
+{
+ PRTFSXFSFILE pThis = (PRTFSXFSFILE)pvThis;
+ return rtFsXfsInode_QueryInfo(pThis->pInode, pObjInfo, enmAddAttr);
+}
+
+
+/**
+ * @interface_method_impl{RTVFSIOSTREAMOPS,pfnRead}
+ */
+static DECLCALLBACK(int) rtFsXfsFile_Read(void *pvThis, RTFOFF off, PCRTSGBUF pSgBuf, bool fBlocking, size_t *pcbRead)
+{
+ PRTFSXFSFILE pThis = (PRTFSXFSFILE)pvThis;
+ AssertReturn(pSgBuf->cSegs == 1, VERR_INTERNAL_ERROR_3);
+ RT_NOREF(fBlocking);
+
+ if (off == -1)
+ off = pThis->offFile;
+ else
+ AssertReturn(off >= 0, VERR_INTERNAL_ERROR_3);
+
+ int rc;
+ size_t cbRead = pSgBuf->paSegs[0].cbSeg;
+ if (!pcbRead)
+ {
+ rc = rtFsXfsInode_Read(pThis->pVol, pThis->pInode, (uint64_t)off, pSgBuf->paSegs[0].pvSeg, cbRead, NULL);
+ if (RT_SUCCESS(rc))
+ pThis->offFile = off + cbRead;
+ Log6(("rtFsXfsFile_Read: off=%#RX64 cbSeg=%#x -> %Rrc\n", off, pSgBuf->paSegs[0].cbSeg, rc));
+ }
+ else
+ {
+ PRTFSXFSINODE pInode = pThis->pInode;
+ if (off >= pInode->ObjInfo.cbObject)
+ {
+ *pcbRead = 0;
+ rc = VINF_EOF;
+ }
+ else
+ {
+ if ((uint64_t)off + cbRead <= (uint64_t)pInode->ObjInfo.cbObject)
+ rc = rtFsXfsInode_Read(pThis->pVol, pThis->pInode, (uint64_t)off, pSgBuf->paSegs[0].pvSeg, cbRead, NULL);
+ else
+ {
+ /* Return VINF_EOF if beyond end-of-file. */
+ cbRead = (size_t)(pInode->ObjInfo.cbObject - off);
+ rc = rtFsXfsInode_Read(pThis->pVol, pThis->pInode, off, pSgBuf->paSegs[0].pvSeg, cbRead, NULL);
+ if (RT_SUCCESS(rc))
+ rc = VINF_EOF;
+ }
+ if (RT_SUCCESS(rc))
+ {
+ pThis->offFile = off + cbRead;
+ *pcbRead = cbRead;
+ }
+ else
+ *pcbRead = 0;
+ }
+ Log6(("rtFsXfsFile_Read: off=%#RX64 cbSeg=%#x -> %Rrc *pcbRead=%#x\n", off, pSgBuf->paSegs[0].cbSeg, rc, *pcbRead));
+ }
+
+ return rc;
+}
+
+
+/**
+ * @interface_method_impl{RTVFSIOSTREAMOPS,pfnWrite}
+ */
+static DECLCALLBACK(int) rtFsXfsFile_Write(void *pvThis, RTFOFF off, PCRTSGBUF pSgBuf, bool fBlocking, size_t *pcbWritten)
+{
+ RT_NOREF(pvThis, off, pSgBuf, fBlocking, pcbWritten);
+ return VERR_WRITE_PROTECT;
+}
+
+
+/**
+ * @interface_method_impl{RTVFSIOSTREAMOPS,pfnFlush}
+ */
+static DECLCALLBACK(int) rtFsXfsFile_Flush(void *pvThis)
+{
+ RT_NOREF(pvThis);
+ return VINF_SUCCESS;
+}
+
+
+/**
+ * @interface_method_impl{RTVFSIOSTREAMOPS,pfnTell}
+ */
+static DECLCALLBACK(int) rtFsXfsFile_Tell(void *pvThis, PRTFOFF poffActual)
+{
+ PRTFSXFSFILE pThis = (PRTFSXFSFILE)pvThis;
+ *poffActual = pThis->offFile;
+ return VINF_SUCCESS;
+}
+
+
+/**
+ * @interface_method_impl{RTVFSOBJSETOPS,pfnMode}
+ */
+static DECLCALLBACK(int) rtFsXfsFile_SetMode(void *pvThis, RTFMODE fMode, RTFMODE fMask)
+{
+ RT_NOREF(pvThis, fMode, fMask);
+ return VERR_WRITE_PROTECT;
+}
+
+
+/**
+ * @interface_method_impl{RTVFSOBJSETOPS,pfnSetTimes}
+ */
+static DECLCALLBACK(int) rtFsXfsFile_SetTimes(void *pvThis, PCRTTIMESPEC pAccessTime, PCRTTIMESPEC pModificationTime,
+ PCRTTIMESPEC pChangeTime, PCRTTIMESPEC pBirthTime)
+{
+ RT_NOREF(pvThis, pAccessTime, pModificationTime, pChangeTime, pBirthTime);
+ return VERR_WRITE_PROTECT;
+}
+
+
+/**
+ * @interface_method_impl{RTVFSOBJSETOPS,pfnSetOwner}
+ */
+static DECLCALLBACK(int) rtFsXfsFile_SetOwner(void *pvThis, RTUID uid, RTGID gid)
+{
+ RT_NOREF(pvThis, uid, gid);
+ return VERR_WRITE_PROTECT;
+}
+
+
+/**
+ * @interface_method_impl{RTVFSFILEOPS,pfnSeek}
+ */
+static DECLCALLBACK(int) rtFsXfsFile_Seek(void *pvThis, RTFOFF offSeek, unsigned uMethod, PRTFOFF poffActual)
+{
+ PRTFSXFSFILE pThis = (PRTFSXFSFILE)pvThis;
+ RTFOFF offNew;
+ switch (uMethod)
+ {
+ case RTFILE_SEEK_BEGIN:
+ offNew = offSeek;
+ break;
+ case RTFILE_SEEK_END:
+ offNew = pThis->pInode->ObjInfo.cbObject + offSeek;
+ break;
+ case RTFILE_SEEK_CURRENT:
+ offNew = (RTFOFF)pThis->offFile + offSeek;
+ break;
+ default:
+ return VERR_INVALID_PARAMETER;
+ }
+ if (offNew >= 0)
+ {
+ pThis->offFile = offNew;
+ *poffActual = offNew;
+ return VINF_SUCCESS;
+ }
+ return VERR_NEGATIVE_SEEK;
+}
+
+
+/**
+ * @interface_method_impl{RTVFSFILEOPS,pfnQuerySize}
+ */
+static DECLCALLBACK(int) rtFsXfsFile_QuerySize(void *pvThis, uint64_t *pcbFile)
+{
+ PRTFSXFSFILE pThis = (PRTFSXFSFILE)pvThis;
+ *pcbFile = (uint64_t)pThis->pInode->ObjInfo.cbObject;
+ return VINF_SUCCESS;
+}
+
+
+/**
+ * @interface_method_impl{RTVFSFILEOPS,pfnSetSize}
+ */
+static DECLCALLBACK(int) rtFsXfsFile_SetSize(void *pvThis, uint64_t cbFile, uint32_t fFlags)
+{
+ RT_NOREF(pvThis, cbFile, fFlags);
+ return VERR_WRITE_PROTECT;
+}
+
+
+/**
+ * @interface_method_impl{RTVFSFILEOPS,pfnQueryMaxSize}
+ */
+static DECLCALLBACK(int) rtFsXfsFile_QueryMaxSize(void *pvThis, uint64_t *pcbMax)
+{
+ RT_NOREF(pvThis);
+ *pcbMax = INT64_MAX; /** @todo */
+ return VINF_SUCCESS;
+}
+
+
+/**
+ * XFS file operations.
+ */
+static const RTVFSFILEOPS g_rtFsXfsFileOps =
+{
+ { /* Stream */
+ { /* Obj */
+ RTVFSOBJOPS_VERSION,
+ RTVFSOBJTYPE_FILE,
+ "XFS File",
+ rtFsXfsFile_Close,
+ rtFsXfsFile_QueryInfo,
+ RTVFSOBJOPS_VERSION
+ },
+ RTVFSIOSTREAMOPS_VERSION,
+ RTVFSIOSTREAMOPS_FEAT_NO_SG,
+ rtFsXfsFile_Read,
+ rtFsXfsFile_Write,
+ rtFsXfsFile_Flush,
+ NULL /*PollOne*/,
+ rtFsXfsFile_Tell,
+ NULL /*pfnSkip*/,
+ NULL /*pfnZeroFill*/,
+ RTVFSIOSTREAMOPS_VERSION,
+ },
+ RTVFSFILEOPS_VERSION,
+ 0,
+ { /* ObjSet */
+ RTVFSOBJSETOPS_VERSION,
+ RT_UOFFSETOF(RTVFSFILEOPS, ObjSet) - RT_UOFFSETOF(RTVFSFILEOPS, Stream.Obj),
+ rtFsXfsFile_SetMode,
+ rtFsXfsFile_SetTimes,
+ rtFsXfsFile_SetOwner,
+ RTVFSOBJSETOPS_VERSION
+ },
+ rtFsXfsFile_Seek,
+ rtFsXfsFile_QuerySize,
+ rtFsXfsFile_SetSize,
+ rtFsXfsFile_QueryMaxSize,
+ RTVFSFILEOPS_VERSION
+};
+
+
+/**
+ * Creates a new VFS file from the given regular file inode.
+ *
+ * @returns IPRT status code.
+ * @param pThis The XFS volume instance.
+ * @param fOpen Open flags passed.
+ * @param iInode The inode for the file.
+ * @param phVfsFile Where to store the VFS file handle on success.
+ * @param pErrInfo Where to record additional error information on error, optional.
+ * @param pszWhat Logging prefix.
+ */
+static int rtFsXfsVol_NewFile(PRTFSXFSVOL pThis, uint64_t fOpen, uint32_t iInode,
+ PRTVFSFILE phVfsFile, PRTERRINFO pErrInfo, const char *pszWhat)
+{
+ /*
+ * Load the inode and check that it really is a file.
+ */
+ PRTFSXFSINODE pInode = NULL;
+ int rc = rtFsXfsInode_Load(pThis, iInode, &pInode);
+ if (RT_SUCCESS(rc))
+ {
+ if (RTFS_IS_FILE(pInode->ObjInfo.Attr.fMode))
+ {
+ PRTFSXFSFILE pNewFile;
+ rc = RTVfsNewFile(&g_rtFsXfsFileOps, sizeof(*pNewFile), fOpen, pThis->hVfsSelf, NIL_RTVFSLOCK,
+ phVfsFile, (void **)&pNewFile);
+ if (RT_SUCCESS(rc))
+ {
+ pNewFile->pVol = pThis;
+ pNewFile->pInode = pInode;
+ pNewFile->offFile = 0;
+ }
+ }
+ else
+ rc = RTERRINFO_LOG_SET_F(pErrInfo, VERR_NOT_A_FILE, "%s: fMode=%#RX32", pszWhat, pInode->ObjInfo.Attr.fMode);
+
+ if (RT_FAILURE(rc))
+ rtFsXfsInode_Release(pThis, pInode);
+ }
+
+ return rc;
+}
+
+
+
+/*
+ *
+ * XFS directory code.
+ * XFS directory code.
+ * XFS directory code.
+ *
+ */
+
+/**
+ * Looks up an entry in the given directory inode.
+ *
+ * @returns IPRT status code.
+ * @param pThis The XFS volume instance.
+ * @param pInode The directory inode structure to.
+ * @param pszEntry The entry to lookup.
+ * @param piInode Where to store the inode number if the entry was found.
+ */
+static int rtFsXfsDir_Lookup(PRTFSXFSVOL pThis, PRTFSXFSINODE pInode, const char *pszEntry, uint32_t *piInode)
+{
+ uint64_t offEntry = 0;
+ int rc = VERR_FILE_NOT_FOUND;
+ uint32_t idxDirEntry = 0;
+ size_t cchEntry = strlen(pszEntry);
+
+ if (cchEntry > 255)
+ return VERR_FILENAME_TOO_LONG;
+
+ RT_NOREF(pThis, idxDirEntry, offEntry, pInode, piInode);
+
+#if 0 /** @todo */
+ while (offEntry < (uint64_t)pInode->ObjInfo.cbObject)
+ {
+ EXTDIRENTRYEX DirEntry;
+ size_t cbThis = RT_MIN(sizeof(DirEntry), (uint64_t)pInode->ObjInfo.cbObject - offEntry);
+ int rc2 = rtFsXfsInode_Read(pThis, pInode, offEntry, &DirEntry, cbThis, NULL);
+ if (RT_SUCCESS(rc2))
+ {
+#ifdef LOG_ENABLED
+ rtFsExtDirEntry_Log(pThis, idxDirEntry, &DirEntry);
+#endif
+
+ uint16_t cbName = pThis->fFeaturesIncompat & EXT_SB_FEAT_INCOMPAT_DIR_FILETYPE
+ ? DirEntry.Core.u.v2.cbName
+ : RT_LE2H_U16(DirEntry.Core.u.v1.cbName);
+ if ( cchEntry == cbName
+ && !memcmp(pszEntry, &DirEntry.Core.achName[0], cchEntry))
+ {
+ *piInode = RT_LE2H_U32(DirEntry.Core.iInodeRef);
+ rc = VINF_SUCCESS;
+ break;
+ }
+
+ offEntry += RT_LE2H_U16(DirEntry.Core.cbRecord);
+ idxDirEntry++;
+ }
+ else
+ {
+ rc = rc2;
+ break;
+ }
+ }
+#endif
+ return rc;
+}
+
+
+
+/*
+ *
+ * Directory instance methods
+ * Directory instance methods
+ * Directory instance methods
+ *
+ */
+
+/**
+ * @interface_method_impl{RTVFSOBJOPS,pfnClose}
+ */
+static DECLCALLBACK(int) rtFsXfsDir_Close(void *pvThis)
+{
+ PRTFSXFSDIR pThis = (PRTFSXFSDIR)pvThis;
+ LogFlowFunc(("pThis=%p\n", pThis));
+ rtFsXfsInode_Release(pThis->pVol, pThis->pInode);
+ pThis->pInode = NULL;
+ return VINF_SUCCESS;
+}
+
+
+/**
+ * @interface_method_impl{RTVFSOBJOPS,pfnQueryInfo}
+ */
+static DECLCALLBACK(int) rtFsXfsDir_QueryInfo(void *pvThis, PRTFSOBJINFO pObjInfo, RTFSOBJATTRADD enmAddAttr)
+{
+ PRTFSXFSDIR pThis = (PRTFSXFSDIR)pvThis;
+ LogFlowFunc(("\n"));
+ return rtFsXfsInode_QueryInfo(pThis->pInode, pObjInfo, enmAddAttr);
+}
+
+
+/**
+ * @interface_method_impl{RTVFSOBJSETOPS,pfnMode}
+ */
+static DECLCALLBACK(int) rtFsXfsDir_SetMode(void *pvThis, RTFMODE fMode, RTFMODE fMask)
+{
+ LogFlowFunc(("\n"));
+ RT_NOREF(pvThis, fMode, fMask);
+ return VERR_WRITE_PROTECT;
+}
+
+
+/**
+ * @interface_method_impl{RTVFSOBJSETOPS,pfnSetTimes}
+ */
+static DECLCALLBACK(int) rtFsXfsDir_SetTimes(void *pvThis, PCRTTIMESPEC pAccessTime, PCRTTIMESPEC pModificationTime,
+ PCRTTIMESPEC pChangeTime, PCRTTIMESPEC pBirthTime)
+{
+ LogFlowFunc(("\n"));
+ RT_NOREF(pvThis, pAccessTime, pModificationTime, pChangeTime, pBirthTime);
+ return VERR_WRITE_PROTECT;
+}
+
+
+/**
+ * @interface_method_impl{RTVFSOBJSETOPS,pfnSetOwner}
+ */
+static DECLCALLBACK(int) rtFsXfsDir_SetOwner(void *pvThis, RTUID uid, RTGID gid)
+{
+ LogFlowFunc(("\n"));
+ RT_NOREF(pvThis, uid, gid);
+ return VERR_WRITE_PROTECT;
+}
+
+
+/**
+ * @interface_method_impl{RTVFSDIROPS,pfnOpen}
+ */
+static DECLCALLBACK(int) rtFsXfsDir_Open(void *pvThis, const char *pszEntry, uint64_t fOpen,
+ uint32_t fFlags, PRTVFSOBJ phVfsObj)
+{
+ LogFlowFunc(("pszEntry='%s' fOpen=%#RX64 fFlags=%#x\n", pszEntry, fOpen, fFlags));
+ PRTFSXFSDIR pThis = (PRTFSXFSDIR)pvThis;
+ PRTFSXFSVOL pVol = pThis->pVol;
+ int rc = VINF_SUCCESS;
+
+ RT_NOREF(pThis, pVol, phVfsObj, pszEntry, fFlags);
+
+ /*
+ * We cannot create or replace anything, just open stuff.
+ */
+ if ( (fOpen & RTFILE_O_ACTION_MASK) == RTFILE_O_OPEN
+ || (fOpen & RTFILE_O_ACTION_MASK) == RTFILE_O_OPEN_CREATE)
+ { /* likely */ }
+ else
+ return VERR_WRITE_PROTECT;
+
+ /*
+ * Lookup the entry.
+ */
+ uint32_t iInode = 0;
+ rc = rtFsXfsDir_Lookup(pVol, pThis->pInode, pszEntry, &iInode);
+ if (RT_SUCCESS(rc))
+ {
+ PRTFSXFSINODE pInode = NULL;
+ rc = rtFsXfsInode_Load(pVol, iInode, &pInode);
+ if (RT_SUCCESS(rc))
+ {
+ if (RTFS_IS_DIRECTORY(pInode->ObjInfo.Attr.fMode))
+ {
+ RTVFSDIR hVfsDir;
+ rc = rtFsXfsVol_OpenDirByInode(pVol, iInode, &hVfsDir);
+ if (RT_SUCCESS(rc))
+ {
+ *phVfsObj = RTVfsObjFromDir(hVfsDir);
+ RTVfsDirRelease(hVfsDir);
+ AssertStmt(*phVfsObj != NIL_RTVFSOBJ, rc = VERR_INTERNAL_ERROR_3);
+ }
+ }
+ else if (RTFS_IS_FILE(pInode->ObjInfo.Attr.fMode))
+ {
+ RTVFSFILE hVfsFile;
+ rc = rtFsXfsVol_NewFile(pVol, fOpen, iInode, &hVfsFile, NULL, pszEntry);
+ if (RT_SUCCESS(rc))
+ {
+ *phVfsObj = RTVfsObjFromFile(hVfsFile);
+ RTVfsFileRelease(hVfsFile);
+ AssertStmt(*phVfsObj != NIL_RTVFSOBJ, rc = VERR_INTERNAL_ERROR_3);
+ }
+ }
+ else
+ rc = VERR_NOT_SUPPORTED;
+ }
+ }
+
+ LogFlow(("rtFsXfsDir_Open(%s): returns %Rrc\n", pszEntry, rc));
+ return rc;
+}
+
+
+/**
+ * @interface_method_impl{RTVFSDIROPS,pfnCreateDir}
+ */
+static DECLCALLBACK(int) rtFsXfsDir_CreateDir(void *pvThis, const char *pszSubDir, RTFMODE fMode, PRTVFSDIR phVfsDir)
+{
+ RT_NOREF(pvThis, pszSubDir, fMode, phVfsDir);
+ LogFlowFunc(("\n"));
+ return VERR_WRITE_PROTECT;
+}
+
+
+/**
+ * @interface_method_impl{RTVFSDIROPS,pfnOpenSymlink}
+ */
+static DECLCALLBACK(int) rtFsXfsDir_OpenSymlink(void *pvThis, const char *pszSymlink, PRTVFSSYMLINK phVfsSymlink)
+{
+ RT_NOREF(pvThis, pszSymlink, phVfsSymlink);
+ LogFlowFunc(("\n"));
+ return VERR_NOT_SUPPORTED;
+}
+
+
+/**
+ * @interface_method_impl{RTVFSDIROPS,pfnCreateSymlink}
+ */
+static DECLCALLBACK(int) rtFsXfsDir_CreateSymlink(void *pvThis, const char *pszSymlink, const char *pszTarget,
+ RTSYMLINKTYPE enmType, PRTVFSSYMLINK phVfsSymlink)
+{
+ RT_NOREF(pvThis, pszSymlink, pszTarget, enmType, phVfsSymlink);
+ LogFlowFunc(("\n"));
+ return VERR_WRITE_PROTECT;
+}
+
+
+/**
+ * @interface_method_impl{RTVFSDIROPS,pfnUnlinkEntry}
+ */
+static DECLCALLBACK(int) rtFsXfsDir_UnlinkEntry(void *pvThis, const char *pszEntry, RTFMODE fType)
+{
+ RT_NOREF(pvThis, pszEntry, fType);
+ LogFlowFunc(("\n"));
+ return VERR_WRITE_PROTECT;
+}
+
+
+/**
+ * @interface_method_impl{RTVFSDIROPS,pfnRenameEntry}
+ */
+static DECLCALLBACK(int) rtFsXfsDir_RenameEntry(void *pvThis, const char *pszEntry, RTFMODE fType, const char *pszNewName)
+{
+ RT_NOREF(pvThis, pszEntry, fType, pszNewName);
+ LogFlowFunc(("\n"));
+ return VERR_WRITE_PROTECT;
+}
+
+
+/**
+ * @interface_method_impl{RTVFSDIROPS,pfnRewindDir}
+ */
+static DECLCALLBACK(int) rtFsXfsDir_RewindDir(void *pvThis)
+{
+ PRTFSXFSDIR pThis = (PRTFSXFSDIR)pvThis;
+ LogFlowFunc(("\n"));
+
+ pThis->fNoMoreFiles = false;
+ pThis->offEntry = 0;
+ pThis->idxEntry = 0;
+ return VINF_SUCCESS;
+}
+
+
+/**
+ * @interface_method_impl{RTVFSDIROPS,pfnReadDir}
+ */
+static DECLCALLBACK(int) rtFsXfsDir_ReadDir(void *pvThis, PRTDIRENTRYEX pDirEntry, size_t *pcbDirEntry,
+ RTFSOBJATTRADD enmAddAttr)
+{
+ PRTFSXFSDIR pThis = (PRTFSXFSDIR)pvThis;
+ PRTFSXFSINODE pInode = pThis->pInode;
+ LogFlowFunc(("\n"));
+
+ if (pThis->fNoMoreFiles)
+ return VERR_NO_MORE_FILES;
+
+ RT_NOREF(pInode, pDirEntry, pcbDirEntry, enmAddAttr);
+ return VERR_NOT_IMPLEMENTED;
+}
+
+
+/**
+ * XFS directory operations.
+ */
+static const RTVFSDIROPS g_rtFsXfsDirOps =
+{
+ { /* Obj */
+ RTVFSOBJOPS_VERSION,
+ RTVFSOBJTYPE_DIR,
+ "XFS Dir",
+ rtFsXfsDir_Close,
+ rtFsXfsDir_QueryInfo,
+ RTVFSOBJOPS_VERSION
+ },
+ RTVFSDIROPS_VERSION,
+ 0,
+ { /* ObjSet */
+ RTVFSOBJSETOPS_VERSION,
+ RT_UOFFSETOF(RTVFSDIROPS, ObjSet) - RT_UOFFSETOF(RTVFSDIROPS, Obj),
+ rtFsXfsDir_SetMode,
+ rtFsXfsDir_SetTimes,
+ rtFsXfsDir_SetOwner,
+ RTVFSOBJSETOPS_VERSION
+ },
+ rtFsXfsDir_Open,
+ NULL /* pfnFollowAbsoluteSymlink */,
+ NULL /* pfnOpenFile */,
+ NULL /* pfnOpenDir */,
+ rtFsXfsDir_CreateDir,
+ rtFsXfsDir_OpenSymlink,
+ rtFsXfsDir_CreateSymlink,
+ NULL /* pfnQueryEntryInfo */,
+ rtFsXfsDir_UnlinkEntry,
+ rtFsXfsDir_RenameEntry,
+ rtFsXfsDir_RewindDir,
+ rtFsXfsDir_ReadDir,
+ RTVFSDIROPS_VERSION,
+};
+
+
+/**
+ * Opens a directory by the given inode.
+ *
+ * @returns IPRT status code.
+ * @param pThis The XFS volume instance.
+ * @param iInode The inode to open.
+ * @param phVfsDir Where to store the handle to the VFS directory on success.
+ */
+static int rtFsXfsVol_OpenDirByInode(PRTFSXFSVOL pThis, uint32_t iInode, PRTVFSDIR phVfsDir)
+{
+ PRTFSXFSINODE pInode = NULL;
+ int rc = rtFsXfsInode_Load(pThis, iInode, &pInode);
+ if (RT_SUCCESS(rc))
+ {
+ if (RTFS_IS_DIRECTORY(pInode->ObjInfo.Attr.fMode))
+ {
+ PRTFSXFSDIR pNewDir;
+ rc = RTVfsNewDir(&g_rtFsXfsDirOps, sizeof(*pNewDir), 0 /*fFlags*/, pThis->hVfsSelf, NIL_RTVFSLOCK,
+ phVfsDir, (void **)&pNewDir);
+ if (RT_SUCCESS(rc))
+ {
+ pNewDir->fNoMoreFiles = false;
+ pNewDir->pVol = pThis;
+ pNewDir->pInode = pInode;
+ }
+ }
+ else
+ rc = VERR_VFS_BOGUS_FORMAT;
+
+ if (RT_FAILURE(rc))
+ rtFsXfsInode_Release(pThis, pInode);
+ }
+
+ return rc;
+}
+
+
+
+/*
+ *
+ * Volume level code.
+ * Volume level code.
+ * Volume level code.
+ *
+ */
+
+static DECLCALLBACK(int) rtFsXfsVolAgTreeDestroy(PAVLU32NODECORE pCore, void *pvUser)
+{
+ RT_NOREF(pvUser);
+
+ PRTFSXFSAG pAg = (PRTFSXFSAG)pCore;
+ Assert(!pAg->cRefs);
+ RTMemFree(pAg);
+ return VINF_SUCCESS;
+}
+
+
+static DECLCALLBACK(int) rtFsXfsVolInodeTreeDestroy(PAVLU64NODECORE pCore, void *pvUser)
+{
+ RT_NOREF(pvUser);
+
+ PRTFSXFSINODE pInode = (PRTFSXFSINODE)pCore;
+ Assert(!pInode->cRefs);
+ RTMemFree(pInode);
+ return VINF_SUCCESS;
+}
+
+
+static DECLCALLBACK(int) rtFsXfsVolBlockTreeDestroy(PAVLU64NODECORE pCore, void *pvUser)
+{
+ RT_NOREF(pvUser);
+
+ PRTFSXFSBLOCKENTRY pBlock = (PRTFSXFSBLOCKENTRY)pCore;
+ Assert(!pBlock->cRefs);
+ RTMemFree(pBlock);
+ return VINF_SUCCESS;
+}
+
+
+/**
+ * @interface_method_impl{RTVFSOBJOPS::Obj,pfnClose}
+ */
+static DECLCALLBACK(int) rtFsXfsVol_Close(void *pvThis)
+{
+ PRTFSXFSVOL pThis = (PRTFSXFSVOL)pvThis;
+
+ /* Destroy the block group tree. */
+ RTAvlU32Destroy(&pThis->AgRoot, rtFsXfsVolAgTreeDestroy, pThis);
+ pThis->AgRoot = NULL;
+ RTListInit(&pThis->LstAgLru);
+
+ /* Destroy the inode tree. */
+ RTAvlU64Destroy(&pThis->InodeRoot, rtFsXfsVolInodeTreeDestroy, pThis);
+ pThis->InodeRoot = NULL;
+ RTListInit(&pThis->LstInodeLru);
+
+ /* Destroy the block cache tree. */
+ RTAvlU64Destroy(&pThis->BlockRoot, rtFsXfsVolBlockTreeDestroy, pThis);
+ pThis->BlockRoot = NULL;
+ RTListInit(&pThis->LstBlockLru);
+
+ /*
+ * Backing file and handles.
+ */
+ RTVfsFileRelease(pThis->hVfsBacking);
+ pThis->hVfsBacking = NIL_RTVFSFILE;
+ pThis->hVfsSelf = NIL_RTVFS;
+
+ return VINF_SUCCESS;
+}
+
+
+/**
+ * @interface_method_impl{RTVFSOBJOPS::Obj,pfnQueryInfo}
+ */
+static DECLCALLBACK(int) rtFsXfsVol_QueryInfo(void *pvThis, PRTFSOBJINFO pObjInfo, RTFSOBJATTRADD enmAddAttr)
+{
+ RT_NOREF(pvThis, pObjInfo, enmAddAttr);
+ return VERR_WRONG_TYPE;
+}
+
+
+/**
+ * @interface_method_impl{RTVFSOBJOPS::Obj,pfnOpenRoot}
+ */
+static DECLCALLBACK(int) rtFsXfsVol_OpenRoot(void *pvThis, PRTVFSDIR phVfsDir)
+{
+ PRTFSXFSVOL pThis = (PRTFSXFSVOL)pvThis;
+ int rc = rtFsXfsVol_OpenDirByInode(pThis, pThis->uInodeRoot, phVfsDir);
+ LogFlowFunc(("returns %Rrc\n", rc));
+ return rc;
+}
+
+
+/**
+ * @interface_method_impl{RTVFSOBJOPS::Obj,pfnQueryRangeState}
+ */
+static DECLCALLBACK(int) rtFsXfsVol_QueryRangeState(void *pvThis, uint64_t off, size_t cb, bool *pfUsed)
+{
+ RT_NOREF(pvThis, off, cb, pfUsed);
+ return VERR_NOT_IMPLEMENTED;
+}
+
+
+DECL_HIDDEN_CONST(const RTVFSOPS) g_rtFsXfsVolOps =
+{
+ /* .Obj = */
+ {
+ /* .uVersion = */ RTVFSOBJOPS_VERSION,
+ /* .enmType = */ RTVFSOBJTYPE_VFS,
+ /* .pszName = */ "XfsVol",
+ /* .pfnClose = */ rtFsXfsVol_Close,
+ /* .pfnQueryInfo = */ rtFsXfsVol_QueryInfo,
+ /* .uEndMarker = */ RTVFSOBJOPS_VERSION
+ },
+ /* .uVersion = */ RTVFSOPS_VERSION,
+ /* .fFeatures = */ 0,
+ /* .pfnOpenRoot = */ rtFsXfsVol_OpenRoot,
+ /* .pfnQueryRangeState = */ rtFsXfsVol_QueryRangeState,
+ /* .uEndMarker = */ RTVFSOPS_VERSION
+};
+
+
+/**
+ * Loads and parses the AGI block.
+ *
+ * @returns IPRT status code.
+ * @param pThis The XFS volume instance.
+ * @param pErrInfo Where to return additional error info.
+ */
+static int rtFsXfsVolLoadAgi(PRTFSXFSVOL pThis, PRTERRINFO pErrInfo)
+{
+ XFSAGI Agi;
+ int rc = RTVfsFileReadAt(pThis->hVfsBacking, 2 * pThis->cbSector, &Agi, sizeof(&Agi), NULL);
+ if (RT_SUCCESS(rc))
+ {
+#ifdef LOG_ENABLED
+ rtFsXfsAgi_Log(0, &Agi);
+#endif
+
+ /** @todo Verification */
+ RT_NOREF(pErrInfo);
+ }
+
+ return rc;
+}
+
+
+/**
+ * Loads and parses the superblock of the filesystem.
+ *
+ * @returns IPRT status code.
+ * @param pThis The XFS volume instance.
+ * @param pErrInfo Where to return additional error info.
+ */
+static int rtFsXfsVolLoadAndParseSuperblock(PRTFSXFSVOL pThis, PRTERRINFO pErrInfo)
+{
+ int rc = VINF_SUCCESS;
+ XFSSUPERBLOCK Sb;
+ rc = RTVfsFileReadAt(pThis->hVfsBacking, XFS_SB_OFFSET, &Sb, sizeof(XFSSUPERBLOCK), NULL);
+ if (RT_FAILURE(rc))
+ return RTERRINFO_LOG_SET(pErrInfo, rc, "Error reading super block");
+
+ /* Validate the superblock. */
+ if (RT_BE2H_U32(Sb.u32Magic) != XFS_SB_MAGIC)
+ return RTERRINFO_LOG_SET_F(pErrInfo, VERR_VFS_UNKNOWN_FORMAT, "Not XFS - Signature mismatch: %RX32", RT_BE2H_U32(Sb.u32Magic));
+
+#ifdef LOG_ENABLED
+ rtFsXfsSb_Log(0, &Sb);
+#endif
+
+ /** @todo More verification */
+ pThis->cbSector = RT_BE2H_U32(Sb.cbSector);
+ pThis->cbBlock = RT_BE2H_U32(Sb.cbBlock);
+ pThis->cBlockShift = Sb.cBlockSzLog;
+ pThis->cBlocksPerAg = RT_BE2H_U32(Sb.cAgBlocks);
+ pThis->cAgs = RT_BE2H_U32(Sb.cAg);
+ pThis->uInodeRoot = RT_BE2H_U64(Sb.uInodeRoot);
+ pThis->cbInode = RT_BE2H_U16(Sb.cbInode);
+ pThis->cInodesPerBlock = RT_BE2H_U16(Sb.cInodesPerBlock);
+ pThis->cAgBlocksLog = Sb.cAgBlocksLog;
+ pThis->cInodesPerBlockLog = Sb.cInodesPerBlockLog;
+ return rc;
+}
+
+
+RTDECL(int) RTFsXfsVolOpen(RTVFSFILE hVfsFileIn, uint32_t fMntFlags, uint32_t fXfsFlags, PRTVFS phVfs, PRTERRINFO pErrInfo)
+{
+ AssertPtrReturn(phVfs, VERR_INVALID_POINTER);
+ AssertReturn(!(fMntFlags & ~RTVFSMNT_F_VALID_MASK), VERR_INVALID_FLAGS);
+ AssertReturn(!fXfsFlags, VERR_INVALID_FLAGS);
+
+ uint32_t cRefs = RTVfsFileRetain(hVfsFileIn);
+ AssertReturn(cRefs != UINT32_MAX, VERR_INVALID_HANDLE);
+
+ /*
+ * Create a VFS instance and initialize the data so rtFsXfsVol_Close works.
+ */
+ RTVFS hVfs;
+ PRTFSXFSVOL pThis;
+ int rc = RTVfsNew(&g_rtFsXfsVolOps, sizeof(*pThis), NIL_RTVFS, RTVFSLOCK_CREATE_RW, &hVfs, (void **)&pThis);
+ if (RT_SUCCESS(rc))
+ {
+ pThis->hVfsBacking = hVfsFileIn;
+ pThis->hVfsSelf = hVfs;
+ pThis->fMntFlags = fMntFlags;
+ pThis->fXfsFlags = fXfsFlags;
+ pThis->AgRoot = NULL;
+ pThis->InodeRoot = NULL;
+ pThis->BlockRoot = NULL;
+ pThis->cbAgs = 0;
+ pThis->cbInodes = 0;
+ pThis->cbBlocks = 0;
+ RTListInit(&pThis->LstAgLru);
+ RTListInit(&pThis->LstInodeLru);
+ RTListInit(&pThis->LstBlockLru);
+
+ rc = RTVfsFileGetSize(pThis->hVfsBacking, &pThis->cbBacking);
+ if (RT_SUCCESS(rc))
+ {
+ rc = rtFsXfsVolLoadAndParseSuperblock(pThis, pErrInfo);
+ if (RT_SUCCESS(rc))
+ rc = rtFsXfsVolLoadAgi(pThis, pErrInfo);
+ if (RT_SUCCESS(rc))
+ {
+ *phVfs = hVfs;
+ return VINF_SUCCESS;
+ }
+ }
+
+ RTVfsRelease(hVfs);
+ *phVfs = NIL_RTVFS;
+ }
+ else
+ RTVfsFileRelease(hVfsFileIn);
+
+ return rc;
+}
+
+
+/**
+ * @interface_method_impl{RTVFSCHAINELEMENTREG,pfnValidate}
+ */
+static DECLCALLBACK(int) rtVfsChainXfsVol_Validate(PCRTVFSCHAINELEMENTREG pProviderReg, PRTVFSCHAINSPEC pSpec,
+ PRTVFSCHAINELEMSPEC pElement, uint32_t *poffError, PRTERRINFO pErrInfo)
+{
+ RT_NOREF(pProviderReg);
+
+ /*
+ * Basic checks.
+ */
+ if (pElement->enmTypeIn != RTVFSOBJTYPE_FILE)
+ return pElement->enmTypeIn == RTVFSOBJTYPE_INVALID ? VERR_VFS_CHAIN_CANNOT_BE_FIRST_ELEMENT : VERR_VFS_CHAIN_TAKES_FILE;
+ if ( pElement->enmType != RTVFSOBJTYPE_VFS
+ && pElement->enmType != RTVFSOBJTYPE_DIR)
+ return VERR_VFS_CHAIN_ONLY_DIR_OR_VFS;
+ if (pElement->cArgs > 1)
+ return VERR_VFS_CHAIN_AT_MOST_ONE_ARG;
+
+ /*
+ * Parse the flag if present, save in pElement->uProvider.
+ */
+ bool fReadOnly = (pSpec->fOpenFile & RTFILE_O_ACCESS_MASK) == RTFILE_O_READ;
+ if (pElement->cArgs > 0)
+ {
+ const char *psz = pElement->paArgs[0].psz;
+ if (*psz)
+ {
+ if (!strcmp(psz, "ro"))
+ fReadOnly = true;
+ else if (!strcmp(psz, "rw"))
+ fReadOnly = false;
+ else
+ {
+ *poffError = pElement->paArgs[0].offSpec;
+ return RTErrInfoSet(pErrInfo, VERR_VFS_CHAIN_INVALID_ARGUMENT, "Expected 'ro' or 'rw' as argument");
+ }
+ }
+ }
+
+ pElement->uProvider = fReadOnly ? RTVFSMNT_F_READ_ONLY : 0;
+ return VINF_SUCCESS;
+}
+
+
+/**
+ * @interface_method_impl{RTVFSCHAINELEMENTREG,pfnInstantiate}
+ */
+static DECLCALLBACK(int) rtVfsChainXfsVol_Instantiate(PCRTVFSCHAINELEMENTREG pProviderReg, PCRTVFSCHAINSPEC pSpec,
+ PCRTVFSCHAINELEMSPEC pElement, RTVFSOBJ hPrevVfsObj,
+ PRTVFSOBJ phVfsObj, uint32_t *poffError, PRTERRINFO pErrInfo)
+{
+ RT_NOREF(pProviderReg, pSpec, poffError);
+
+ int rc;
+ RTVFSFILE hVfsFileIn = RTVfsObjToFile(hPrevVfsObj);
+ if (hVfsFileIn != NIL_RTVFSFILE)
+ {
+ RTVFS hVfs;
+ rc = RTFsXfsVolOpen(hVfsFileIn, (uint32_t)pElement->uProvider, (uint32_t)(pElement->uProvider >> 32), &hVfs, pErrInfo);
+ RTVfsFileRelease(hVfsFileIn);
+ if (RT_SUCCESS(rc))
+ {
+ *phVfsObj = RTVfsObjFromVfs(hVfs);
+ RTVfsRelease(hVfs);
+ if (*phVfsObj != NIL_RTVFSOBJ)
+ return VINF_SUCCESS;
+ rc = VERR_VFS_CHAIN_CAST_FAILED;
+ }
+ }
+ else
+ rc = VERR_VFS_CHAIN_CAST_FAILED;
+ return rc;
+}
+
+
+/**
+ * @interface_method_impl{RTVFSCHAINELEMENTREG,pfnCanReuseElement}
+ */
+static DECLCALLBACK(bool) rtVfsChainXfsVol_CanReuseElement(PCRTVFSCHAINELEMENTREG pProviderReg,
+ PCRTVFSCHAINSPEC pSpec, PCRTVFSCHAINELEMSPEC pElement,
+ PCRTVFSCHAINSPEC pReuseSpec, PCRTVFSCHAINELEMSPEC pReuseElement)
+{
+ RT_NOREF(pProviderReg, pSpec, pReuseSpec);
+ if ( pElement->paArgs[0].uProvider == pReuseElement->paArgs[0].uProvider
+ || !pReuseElement->paArgs[0].uProvider)
+ return true;
+ return false;
+}
+
+
+/** VFS chain element 'xfs'. */
+static RTVFSCHAINELEMENTREG g_rtVfsChainXfsVolReg =
+{
+ /* uVersion = */ RTVFSCHAINELEMENTREG_VERSION,
+ /* fReserved = */ 0,
+ /* pszName = */ "xfs",
+ /* ListEntry = */ { NULL, NULL },
+ /* pszHelp = */ "Open a XFS file system, requires a file object on the left side.\n"
+ "First argument is an optional 'ro' (read-only) or 'rw' (read-write) flag.\n",
+ /* pfnValidate = */ rtVfsChainXfsVol_Validate,
+ /* pfnInstantiate = */ rtVfsChainXfsVol_Instantiate,
+ /* pfnCanReuseElement = */ rtVfsChainXfsVol_CanReuseElement,
+ /* uEndMarker = */ RTVFSCHAINELEMENTREG_VERSION
+};
+
+RTVFSCHAIN_AUTO_REGISTER_ELEMENT_PROVIDER(&g_rtVfsChainXfsVolReg, rtVfsChainXfsVolReg);
+