[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <feeb2dc096e3cf080b91eca836957136f580da3a.1667822611.git.ritesh.list@gmail.com>
Date: Mon, 7 Nov 2022 17:51:44 +0530
From: "Ritesh Harjani (IBM)" <ritesh.list@...il.com>
To: Theodore Ts'o <tytso@....edu>
Cc: linux-ext4@...r.kernel.org,
Harshad Shirwadkar <harshadshirwadkar@...il.com>,
Wang Shilong <wshilong@....com>,
Andreas Dilger <adilger.kernel@...ger.ca>,
Li Xi <lixi@....com>, Ritesh Harjani <ritesh.list@...il.com>
Subject: [RFCv1 56/72] e2fsck: wait fix thread finish before checking
From: Wang Shilong <wshilong@....com>
Before proceeding next inodes, waitting existed
fixing finished.
Signed-off-by: Wang Shilong <wshilong@....com>
[unlock from Jan's orphan inode path as well]
Signed-off-by: Ritesh Harjani (IBM) <ritesh.list@...il.com>
---
e2fsck/e2fsck.c | 3 +++
e2fsck/e2fsck.h | 5 +++-
e2fsck/pass1.c | 70 +++++++++++++++++++++++++++++++++++++++++++------
e2fsck/util.c | 56 ++++++++++++++++++++++++++++++++++++---
4 files changed, 122 insertions(+), 12 deletions(-)
diff --git a/e2fsck/e2fsck.c b/e2fsck/e2fsck.c
index 1e295e3e..a5150dab 100644
--- a/e2fsck/e2fsck.c
+++ b/e2fsck/e2fsck.c
@@ -187,6 +187,9 @@ errcode_t e2fsck_reset_context(e2fsck_t ctx)
ctx->fs_fragmented_dir = 0;
ctx->large_files = 0;
ctx->large_dirs = 0;
+#ifdef HAVE_PTHREAD
+ ctx->fs_need_locking = 0;
+#endif
for (i=0; i < MAX_EXTENT_DEPTH_COUNT; i++)
ctx->extent_depth_count[i] = 0;
diff --git a/e2fsck/e2fsck.h b/e2fsck/e2fsck.h
index e3276924..01bd9d01 100644
--- a/e2fsck/e2fsck.h
+++ b/e2fsck/e2fsck.h
@@ -488,8 +488,9 @@ struct e2fsck_struct {
#ifdef HAVE_PTHREAD
__u32 fs_num_threads;
+ int fs_need_locking;
/* serialize fix operation for multiple threads */
- pthread_mutex_t fs_fix_mutex;
+ pthread_rwlock_t fs_fix_rwlock;
/* protect block_found_map, block_dup_map */
pthread_rwlock_t fs_block_map_rwlock;
#endif
@@ -553,6 +554,8 @@ extern int e2fsck_strnlen(const char * s, int count);
extern void e2fsck_pass1(e2fsck_t ctx);
extern void e2fsck_pass1_dupblocks(e2fsck_t ctx, char *block_buf);
+extern void e2fsck_pass1_check_lock(e2fsck_t ctx);
+extern void e2fsck_pass1_check_unlock(e2fsck_t ctx);
extern void e2fsck_pass2(e2fsck_t ctx);
extern void e2fsck_pass3(e2fsck_t ctx);
extern void e2fsck_pass4(e2fsck_t ctx);
diff --git a/e2fsck/pass1.c b/e2fsck/pass1.c
index a5dc6e44..29333acf 100644
--- a/e2fsck/pass1.c
+++ b/e2fsck/pass1.c
@@ -991,8 +991,10 @@ static void finish_processing_inode(e2fsck_t ctx, ext2_ino_t ino,
#define FINISH_INODE_LOOP(ctx, ino, pctx, failed_csum) \
do { \
finish_processing_inode((ctx), (ino), (pctx), (failed_csum)); \
- if ((ctx)->flags & E2F_FLAG_ABORT) \
+ if ((ctx)->flags & E2F_FLAG_ABORT) { \
+ e2fsck_pass1_check_unlock(ctx); \
return; \
+ } \
} while (0)
static int could_be_block_map(ext2_filsys fs, struct ext2_inode *inode)
@@ -1374,8 +1376,12 @@ static errcode_t e2fsck_pass1_prepare(e2fsck_t ctx)
ext2fs_mark_block_bitmap2(ctx->block_found_map,
fs->super->s_mmp_block);
#ifdef HAVE_PTHREAD
- pthread_mutex_init(&ctx->fs_fix_mutex, NULL);
+ pthread_rwlock_init(&ctx->fs_fix_rwlock, NULL);
pthread_rwlock_init(&ctx->fs_block_map_rwlock, NULL);
+ if (ctx->fs_num_threads > 1)
+ ctx->fs_need_locking = 1;
+ else
+ ctx->fs_need_locking = 0;
#endif
return 0;
@@ -1387,6 +1393,10 @@ static void e2fsck_pass1_post(e2fsck_t ctx)
ext2_filsys fs = ctx->fs;
char *block_buf;
+#ifdef HAVE_PTHREAD
+ ctx->fs_need_locking = 0;
+#endif
+
if (e2fsck_should_abort(ctx))
return;
@@ -1662,6 +1672,7 @@ void e2fsck_pass1_run(e2fsck_t ctx)
#endif
while (1) {
+ e2fsck_pass1_check_lock(ctx);
if (ino % (fs->super->s_inodes_per_group * 4) == 1) {
if (e2fsck_mmp_update(fs))
fatal_error(ctx, 0);
@@ -1672,8 +1683,10 @@ void e2fsck_pass1_run(e2fsck_t ctx)
if (ino > ino_threshold)
pass1_readahead(ctx, &ra_group, &ino_threshold);
ehandler_operation(old_op);
- if (e2fsck_should_abort(ctx))
+ if (e2fsck_should_abort(ctx)) {
+ e2fsck_pass1_check_unlock(ctx);
goto endit;
+ }
if (pctx.errcode == EXT2_ET_BAD_BLOCK_IN_INODE_TABLE) {
/*
* If badblocks says badblocks is bad, offer to clear
@@ -1694,27 +1707,45 @@ void e2fsck_pass1_run(e2fsck_t ctx)
fix_problem(ctx, PR_1_ISCAN_ERROR,
&pctx);
ctx->flags |= E2F_FLAG_ABORT;
+ e2fsck_pass1_check_unlock(ctx);
+ goto endit;
} else
ctx->flags |= E2F_FLAG_RESTART;
- goto endit;
+ err = ext2fs_inode_scan_goto_blockgroup(scan,
+ 0);
+ if (err) {
+ fix_problem(ctx, PR_1_ISCAN_ERROR,
+ &pctx);
+ ctx->flags |= E2F_FLAG_ABORT;
+ e2fsck_pass1_check_unlock(ctx);
+ goto endit;
+ }
+ e2fsck_pass1_check_unlock(ctx);
+ continue;
}
if (!ctx->inode_bb_map)
alloc_bb_map(ctx);
ext2fs_mark_inode_bitmap2(ctx->inode_bb_map, ino);
ext2fs_mark_inode_bitmap2(ctx->inode_used_map, ino);
+ e2fsck_pass1_check_unlock(ctx);
continue;
}
- if (pctx.errcode == EXT2_ET_SCAN_FINISHED)
+ if (pctx.errcode == EXT2_ET_SCAN_FINISHED) {
+ e2fsck_pass1_check_unlock(ctx);
break;
+ }
if (pctx.errcode &&
pctx.errcode != EXT2_ET_INODE_CSUM_INVALID &&
pctx.errcode != EXT2_ET_INODE_IS_GARBAGE) {
fix_problem(ctx, PR_1_ISCAN_ERROR, &pctx);
ctx->flags |= E2F_FLAG_ABORT;
+ e2fsck_pass1_check_unlock(ctx);
goto endit;
}
- if (!ino)
+ if (!ino) {
+ e2fsck_pass1_check_unlock(ctx);
break;
+ }
#ifdef HAVE_PTHREAD
if (ctx->global_ctx)
ctx->thread_info.et_inode_number++;
@@ -1767,6 +1798,7 @@ void e2fsck_pass1_run(e2fsck_t ctx)
pctx.num = inode->i_links_count;
fix_problem(ctx, PR_1_ICOUNT_STORE, &pctx);
ctx->flags |= E2F_FLAG_ABORT;
+ e2fsck_pass1_check_unlock(ctx);
goto endit;
}
} else if ((ino >= EXT2_FIRST_INODE(fs->super)) &&
@@ -1781,6 +1813,7 @@ void e2fsck_pass1_run(e2fsck_t ctx)
}
}
FINISH_INODE_LOOP(ctx, ino, &pctx, failed_csum);
+ e2fsck_pass1_check_unlock(ctx);
continue;
}
@@ -1801,6 +1834,7 @@ void e2fsck_pass1_run(e2fsck_t ctx)
&pctx);
if (res < 0) {
/* skip FINISH_INODE_LOOP */
+ e2fsck_pass1_check_unlock(ctx);
continue;
}
}
@@ -1822,6 +1856,7 @@ void e2fsck_pass1_run(e2fsck_t ctx)
} else if (fix_problem(ctx, PR_1_INLINE_DATA_SET, &pctx)) {
e2fsck_clear_inode(ctx, ino, inode, 0, "pass1");
/* skip FINISH_INODE_LOOP */
+ e2fsck_pass1_check_unlock(ctx);
continue;
}
}
@@ -1866,6 +1901,7 @@ void e2fsck_pass1_run(e2fsck_t ctx)
if (err) {
pctx.errcode = err;
ctx->flags |= E2F_FLAG_ABORT;
+ e2fsck_pass1_check_unlock(ctx);
goto endit;
}
inode->i_flags &= ~EXT4_INLINE_DATA_FL;
@@ -1880,6 +1916,7 @@ void e2fsck_pass1_run(e2fsck_t ctx)
/* Some other kind of non-xattr error? */
pctx.errcode = err;
ctx->flags |= E2F_FLAG_ABORT;
+ e2fsck_pass1_check_unlock(ctx);
goto endit;
}
}
@@ -1917,6 +1954,7 @@ void e2fsck_pass1_run(e2fsck_t ctx)
ext2fs_mark_inode_bitmap2(ctx->inode_used_map,
ino);
/* skip FINISH_INODE_LOOP */
+ e2fsck_pass1_check_unlock(ctx);
continue;
}
}
@@ -1980,6 +2018,7 @@ void e2fsck_pass1_run(e2fsck_t ctx)
pctx.num = 4;
fix_problem(ctx, PR_1_ALLOCATE_BBITMAP_ERROR, &pctx);
ctx->flags |= E2F_FLAG_ABORT;
+ e2fsck_pass1_check_unlock(ctx);
goto endit;
}
pb.ino = EXT2_BAD_INO;
@@ -1997,16 +2036,19 @@ void e2fsck_pass1_run(e2fsck_t ctx)
if (pctx.errcode) {
fix_problem(ctx, PR_1_BLOCK_ITERATE, &pctx);
ctx->flags |= E2F_FLAG_ABORT;
+ e2fsck_pass1_check_unlock(ctx);
goto endit;
}
if (pb.bbcheck)
if (!fix_problem(ctx, PR_1_BBINODE_BAD_METABLOCK_PROMPT, &pctx)) {
ctx->flags |= E2F_FLAG_ABORT;
+ e2fsck_pass1_check_unlock(ctx);
goto endit;
}
ext2fs_mark_inode_bitmap2(ctx->inode_used_map, ino);
clear_problem_context(&pctx);
FINISH_INODE_LOOP(ctx, ino, &pctx, failed_csum);
+ e2fsck_pass1_check_unlock(ctx);
continue;
} else if (ino == EXT2_ROOT_INO) {
/*
@@ -2048,6 +2090,7 @@ void e2fsck_pass1_run(e2fsck_t ctx)
}
check_blocks(ctx, &pctx, block_buf, NULL);
FINISH_INODE_LOOP(ctx, ino, &pctx, failed_csum);
+ e2fsck_pass1_check_unlock(ctx);
continue;
}
if ((inode->i_links_count ||
@@ -2075,6 +2118,7 @@ void e2fsck_pass1_run(e2fsck_t ctx)
}
check_blocks(ctx, &pctx, block_buf, NULL);
FINISH_INODE_LOOP(ctx, ino, &pctx, failed_csum);
+ e2fsck_pass1_check_unlock(ctx);
continue;
}
if ((inode->i_links_count ||
@@ -2101,6 +2145,7 @@ void e2fsck_pass1_run(e2fsck_t ctx)
}
check_blocks(ctx, &pctx, block_buf, NULL);
FINISH_INODE_LOOP(ctx, ino, &pctx, failed_csum);
+ e2fsck_pass1_check_unlock(ctx);
continue;
}
if ((inode->i_links_count ||
@@ -2139,11 +2184,13 @@ void e2fsck_pass1_run(e2fsck_t ctx)
}
check_blocks(ctx, &pctx, block_buf, NULL);
FINISH_INODE_LOOP(ctx, ino, &pctx, failed_csum);
+ e2fsck_pass1_check_unlock(ctx);
continue;
}
if (!inode->i_links_count) {
FINISH_INODE_LOOP(ctx, ino, &pctx, failed_csum);
+ e2fsck_pass1_check_unlock(ctx);
continue;
}
/*
@@ -2253,12 +2300,14 @@ void e2fsck_pass1_run(e2fsck_t ctx)
ctx->fs_symlinks_count++;
if (inode->i_flags & EXT4_INLINE_DATA_FL) {
FINISH_INODE_LOOP(ctx, ino, &pctx, failed_csum);
+ e2fsck_pass1_check_unlock(ctx);
continue;
} else if (ext2fs_is_fast_symlink(inode)) {
ctx->fs_fast_symlinks_count++;
check_blocks(ctx, &pctx, block_buf,
&ea_ibody_quota);
FINISH_INODE_LOOP(ctx, ino, &pctx, failed_csum);
+ e2fsck_pass1_check_unlock(ctx);
continue;
}
}
@@ -2306,16 +2355,21 @@ void e2fsck_pass1_run(e2fsck_t ctx)
FINISH_INODE_LOOP(ctx, ino, &pctx, failed_csum);
- if (ctx->flags & E2F_FLAG_SIGNAL_MASK)
+ if (e2fsck_should_abort(ctx)) {
+ e2fsck_pass1_check_unlock(ctx);
goto endit;
+ }
if (process_inode_count >= ctx->process_inode_size) {
process_inodes(ctx, block_buf, inodes_to_process,
&process_inode_count);
- if (e2fsck_should_abort(ctx))
+ if (e2fsck_should_abort(ctx)) {
+ e2fsck_pass1_check_unlock(ctx);
goto endit;
+ }
}
+ e2fsck_pass1_check_unlock(ctx);
}
process_inodes(ctx, block_buf, inodes_to_process,
&process_inode_count);
diff --git a/e2fsck/util.c b/e2fsck/util.c
index 5714576a..b7c1e7a5 100644
--- a/e2fsck/util.c
+++ b/e2fsck/util.c
@@ -82,7 +82,8 @@ void fatal_error(e2fsck_t ctx, const char *msg)
}
out:
ctx->flags |= E2F_FLAG_ABORT;
- if (ctx->flags & E2F_FLAG_SETJMP_OK)
+ if (!(ctx->options & E2F_OPT_MULTITHREAD) &&
+ ctx->flags & E2F_FLAG_SETJMP_OK)
longjmp(ctx->abort_loc, 1);
if (ctx->logf)
fprintf(ctx->logf, "Exit status: %d\n", exit_value);
@@ -580,38 +581,79 @@ void e2fsck_read_inode_full(e2fsck_t ctx, unsigned long ino,
if (!global_ctx) \
global_ctx = ctx; \
+/**
+ * before we hold write lock, read lock should
+ * has been held.
+ */
void e2fsck_pass1_fix_lock(e2fsck_t ctx)
{
+ int err;
+
+ if (!ctx->fs_need_locking)
+ return;
+
e2fsck_get_lock_context(ctx);
- pthread_mutex_lock(&global_ctx->fs_fix_mutex);
+ err = pthread_rwlock_trywrlock(&global_ctx->fs_fix_rwlock);
+ assert(err != 0);
+ pthread_rwlock_unlock(&global_ctx->fs_fix_rwlock);
+ pthread_rwlock_wrlock(&global_ctx->fs_fix_rwlock);
}
void e2fsck_pass1_fix_unlock(e2fsck_t ctx)
{
+ if (!ctx->fs_need_locking)
+ return;
e2fsck_get_lock_context(ctx);
- pthread_mutex_unlock(&global_ctx->fs_fix_mutex);
+ /* unlock write lock */
+ pthread_rwlock_unlock(&global_ctx->fs_fix_rwlock);
+ /* get read lock again */
+ pthread_rwlock_rdlock(&global_ctx->fs_fix_rwlock);
+}
+
+void e2fsck_pass1_check_lock(e2fsck_t ctx)
+{
+ if (!ctx->fs_need_locking)
+ return;
+ e2fsck_get_lock_context(ctx);
+ pthread_rwlock_rdlock(&global_ctx->fs_fix_rwlock);
+}
+
+void e2fsck_pass1_check_unlock(e2fsck_t ctx)
+{
+ if (!ctx->fs_need_locking)
+ return;
+ e2fsck_get_lock_context(ctx);
+ pthread_rwlock_unlock(&global_ctx->fs_fix_rwlock);
}
void e2fsck_pass1_block_map_w_lock(e2fsck_t ctx)
{
+ if (!ctx->fs_need_locking)
+ return;
e2fsck_get_lock_context(ctx);
pthread_rwlock_wrlock(&global_ctx->fs_block_map_rwlock);
}
void e2fsck_pass1_block_map_w_unlock(e2fsck_t ctx)
{
+ if (!ctx->fs_need_locking)
+ return;
e2fsck_get_lock_context(ctx);
pthread_rwlock_unlock(&global_ctx->fs_block_map_rwlock);
}
void e2fsck_pass1_block_map_r_lock(e2fsck_t ctx)
{
+ if (!ctx->fs_need_locking)
+ return;
e2fsck_get_lock_context(ctx);
pthread_rwlock_rdlock(&global_ctx->fs_block_map_rwlock);
}
void e2fsck_pass1_block_map_r_unlock(e2fsck_t ctx)
{
+ if (!ctx->fs_need_locking)
+ return;
e2fsck_get_lock_context(ctx);
pthread_rwlock_unlock(&global_ctx->fs_block_map_rwlock);
}
@@ -624,6 +666,14 @@ void e2fsck_pass1_fix_lock(e2fsck_t ctx)
void e2fsck_pass1_fix_unlock(e2fsck_t ctx)
{
+}
+void e2fsck_pass1_check_lock(e2fsck_t ctx)
+{
+
+}
+void e2fsck_pass1_check_unlock(e2fsck_t ctx)
+{
+
}
void e2fsck_pass1_block_map_w_lock(e2fsck_t ctx)
{
--
2.37.3
Powered by blists - more mailing lists