linux-ext4.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: "Ritesh Harjani (IBM)" <ritesh.list@gmail.com>
To: Andreas Dilger <adilger@dilger.ca>
Cc: Theodore Ts'o <tytso@mit.edu>,
	linux-ext4@vger.kernel.org,
	Harshad Shirwadkar <harshadshirwadkar@gmail.com>,
	Wang Shilong <wshilong@ddn.com>,
	Andreas Dilger <adilger.kernel@dilger.ca>, Li Xi <lixi@ddn.com>,
	Saranya Muruganandam <saranyamohan@google.com>
Subject: Re: [RFCv1 12/72] libext2fs: dupfs: Add fs clone & merge api
Date: Sat, 19 Nov 2022 10:32:29 +0530	[thread overview]
Message-ID: <20221119050229.xznagyl2v6bmqgvt@riteshh-domain> (raw)
In-Reply-To: <78178FB1-8B01-4956-BDBE-751251E2D58A@dilger.ca>

On 22/11/18 12:46PM, Andreas Dilger wrote:
> On Nov 7, 2022, at 05:23, Ritesh Harjani (IBM) <ritesh.list@gmail.com> wrote:
> > 
> > From: Saranya Muruganandam <saranyamohan@google.com>
> > 
> > This patch mainly adds "parent" & "clone_flags" member in ext2_filsys struct
> > for enabling multi-threading. Based on what CLONE flags will be passed from
> > the client of libext2fs down to ext2fs_clone_fs(), those structures/bitmaps will
> > be cloned (thread-aware child copy) and rest will be shared with the parent fs.
> > 
> > The same flags will also help to merge those cloned bitmap structures back into
> > the parent bitmaps when ext2fs_merge_fs() will be called with childfs struct.
> > 
> > Signed-off-by: Saranya Muruganandam <saranyamohan@google.com>
> > Signed-off-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
> > ---
> > lib/ext2fs/dupfs.c  | 183 ++++++++++++++++++++++++++++++++++++++++++++
> > lib/ext2fs/ext2fs.h |  23 ++++++
> > 2 files changed, 206 insertions(+)
> > 
> > diff --git a/lib/ext2fs/dupfs.c b/lib/ext2fs/dupfs.c
> > index 02721e1a..ecc57cf7 100644
> > --- a/lib/ext2fs/dupfs.c
> > +++ b/lib/ext2fs/dupfs.c
> > @@ -14,8 +14,12 @@
> > #if HAVE_UNISTD_H
> > #include <unistd.h>
> > #endif
> > +#if HAVE_PTHREAD_H
> > +#include <pthread.h>
> > +#endif
> > #include <time.h>
> > #include <string.h>
> > +#include <assert.h>
> > 
> > #include "ext2_fs.h"
> > #include "ext2fsP.h"
> > @@ -120,3 +124,182 @@ errout:
> > 
> > }
> > 
> > +#ifdef HAVE_PTHREAD
> > +errcode_t ext2fs_clone_fs(ext2_filsys fs, ext2_filsys *dest, unsigned int flags)
> > +{
> > +    errcode_t retval;
> > +    ext2_filsys childfs;
> > +
> > +    EXT2_CHECK_MAGIC(fs, EXT2_ET_MAGIC_EXT2FS_FILSYS);
> > +
> > +    retval = ext2fs_get_mem(sizeof(struct struct_ext2_filsys), &childfs);
> > +    if (retval)
> > +        return retval;
> > +
> > +    /* make an exact copy implying lists and memory structures are shared */
> > +    memcpy(childfs, fs, sizeof(struct struct_ext2_filsys));
> > +    childfs->inode_map = NULL;
> > +    childfs->block_map = NULL;
> > +    childfs->badblocks = NULL;
> > +    childfs->dblist = NULL;
> > +
> > +    pthread_mutex_lock(&fs->refcount_mutex);
> > +    fs->refcount++;
> > +    pthread_mutex_unlock(&fs->refcount_mutex);
> 
> The locking her doesn't make sense. Why is the mutex only protecting the "refcount",
> and not the rest of the fields in fs?
> 
> It would make more sense to hold the mutex over the whole copy process, from up before
> the memcpy(), until down after the last structure is cloned. Otherwise, once fs has
> been cloned once and given to another thread it would be possible for them to
> change these structures. 

Ok, so there are two things here. Let's first discuss it from the e2fsck perspective. 
ext2fs_clone_fs() is called in the thread prepare for each thread
sequentially. Also what we are really trying to do here is copy the ext2_filsys
source fs and then all threads will work on their cloned "childfs".

So we don't require any locking, because even when threads will start running,
they will only modify their childfs.

Now.. let's see it from the libext2fs API perspective. So will there be any need
for locking for any other client which might use ext2fs_clone_fs()?
Because the general pthread programming would be like :

start():
	for (t = 0; t < num_threads; t++) {
		clone_contexts() -> ext2fs_clone_fs()
		pthread_run()
		<..>
	}

finish():
	for (t = 0; t < num_threads; t++) {
		pthread_join()
		merge_contexts() -> ext2fs_merge_fs()
	}

So why we don't require any locking here, right? 
Unless ofcourse start() phase can also be called in parallel. 
But that won't be true no.

> 
> > +    if ((flags & EXT2FS_CLONE_INODE) && fs->inode_map) {
> > +        retval = ext2fs_copy_bitmap(fs->inode_map, &childfs->inode_map);
> > +        if (retval)
> > +            return retval;
> > +        childfs->inode_map->fs = childfs;
> > +    }
> > +
> > +    if ((flags & EXT2FS_CLONE_BLOCK) && fs->block_map) {
> > +        retval = ext2fs_copy_bitmap(fs->block_map, &childfs->block_map);
> > +        if (retval)
> > +            return retval;
> > +        childfs->block_map->fs = childfs;
> > +    }
> > +
> > +    if ((flags & EXT2FS_CLONE_BADBLOCKS) && fs->badblocks) {
> > +        retval = ext2fs_badblocks_copy(fs->badblocks, &childfs->badblocks);
> > +        if (retval)
> > +            return retval;
> > +    }
> > +
> > +    if ((flags & EXT2FS_CLONE_DBLIST) && fs->dblist) {
> > +        retval = ext2fs_copy_dblist(fs->dblist, &childfs->dblist);
> > +        if (retval)
> > +            return retval;
> > +        childfs->dblist->fs = childfs;
> > +    }
> > +
> > +    /* icache when NULL will be rebuilt if needed */
> > +    childfs->icache = NULL;
> 
> This should be up where the other NULL values are set. 

ok.

> 
> > +
> > +    childfs->clone_flags = flags;
> > +    childfs->parent = fs;
> > +    *dest = childfs;
> > +
> > +    return 0;
> > +}
> > +
> > +errcode_t ext2fs_merge_fs(ext2_filsys *thread_fs)
> > +{
> > +    ext2_filsys fs = *thread_fs;
> > +    errcode_t retval = 0;
> > +    ext2_filsys dest = fs->parent;
> > +    ext2_filsys src = fs;
> > +    unsigned int flags = fs->clone_flags;
> > +    struct ext2_inode_cache *icache;
> > +    io_channel dest_io;
> > +    io_channel dest_image_io;
> > +    ext2fs_inode_bitmap inode_map;
> > +    ext2fs_block_bitmap block_map;
> > +    ext2_badblocks_list badblocks;
> > +    ext2_dblist dblist;
> > +    void *priv_data;
> > +    int fsflags;
> > +
> > +    pthread_mutex_lock(&fs->refcount_mutex);
> > +    fs->refcount--;
> > +    assert(fs->refcount >= 0);
> > +    pthread_mutex_unlock(&fs->refcount_mutex);
> 
> Same here. The mutex should be held over the whole copy process. 

Again same discussion as above. We don't need to have a lock here, because the
ext2fs_merge_fs() will happen post pthread_join() for each thread.

-ritesh

> 
> Cheers, Andreas
> 
> 
> > +    icache = dest->icache;
> > +    dest_io = dest->io;
> > +    dest_image_io = dest->image_io;
> > +    inode_map = dest->inode_map;
> > +    block_map = dest->block_map;
> > +    badblocks = dest->badblocks;
> > +    dblist = dest->dblist;
> > +    priv_data = dest->priv_data;
> > +    fsflags = dest->flags;
> > +
> > +    memcpy(dest, src, sizeof(struct struct_ext2_filsys));
> > +
> > +    dest->io = dest_io;
> > +    dest->image_io = dest_image_io;
> > +    dest->icache = icache;
> > +    dest->inode_map = inode_map;
> > +    dest->block_map = block_map;
> > +    dest->badblocks = badblocks;
> > +    dest->dblist = dblist;
> > +    dest->priv_data = priv_data;
> > +    if (dest->dblist)
> > +        dest->dblist->fs = dest;
> > +    dest->flags = src->flags | fsflags;
> > +    if (!(src->flags & EXT2_FLAG_VALID) || !(dest->flags & EXT2_FLAG_VALID))
> > +        ext2fs_unmark_valid(dest);
> > +
> > +    if ((flags & EXT2FS_CLONE_INODE) && src->inode_map) {
> > +        if (dest->inode_map == NULL) {
> > +            dest->inode_map = src->inode_map;
> > +            src->inode_map = NULL;
> > +        } else {
> > +            retval = ext2fs_merge_bitmap(src->inode_map, dest->inode_map, NULL, NULL);
> > +            if (retval)
> > +                goto out;
> > +        }
> > +        dest->inode_map->fs = dest;
> > +    }
> > +
> > +    if ((flags & EXT2FS_CLONE_BLOCK) && src->block_map) {
> > +        if (dest->block_map == NULL) {
> > +            dest->block_map = src->block_map;
> > +            src->block_map = NULL;
> > +        } else {
> > +            retval = ext2fs_merge_bitmap(src->block_map, dest->block_map, NULL, NULL);
> > +            if (retval)
> > +                goto out;
> > +        }
> > +        dest->block_map->fs = dest;
> > +    }
> > +
> > +    if ((flags & EXT2FS_CLONE_BADBLOCKS) && src->badblocks) {
> > +        if (dest->badblocks == NULL)
> > +            retval = ext2fs_badblocks_copy(src->badblocks, &dest->badblocks);
> > +        else
> > +            retval = ext2fs_badblocks_merge(src->badblocks, dest->badblocks);
> > +        if (retval)
> > +            goto out;
> > +    }
> > +
> > +    if ((flags & EXT2FS_CLONE_DBLIST) && src->dblist) {
> > +        if (dest->dblist == NULL) {
> > +            dest->dblist = src->dblist;
> > +            src->dblist = NULL;
> > +        } else {
> > +            retval = ext2fs_merge_dblist(src->dblist, dest->dblist);
> > +            if (retval)
> > +                goto out;
> > +        }
> > +        dest->dblist->fs = dest;
> > +    }
> > +
> > +    if (src->icache) {
> > +        ext2fs_free_inode_cache(src->icache);
> > +        src->icache = NULL;
> > +    }
> > +
> > +out:
> > +    if (src->io)
> > +        io_channel_close(src->io);
> > +
> > +    if ((flags & EXT2FS_CLONE_INODE) && src->inode_map)
> > +        ext2fs_free_generic_bmap(src->inode_map);
> > +    if ((flags & EXT2FS_CLONE_BLOCK) && src->block_map)
> > +        ext2fs_free_generic_bmap(src->block_map);
> > +    if ((flags & EXT2FS_CLONE_BADBLOCKS) && src->badblocks)
> > +        ext2fs_badblocks_list_free(src->badblocks);
> > +    if ((flags & EXT2FS_CLONE_DBLIST) && src->dblist) {
> > +        ext2fs_free_dblist(src->dblist);
> > +        src->dblist = NULL;
> > +    }
> > +
> > +    ext2fs_free_mem(&src);
> > +    *thread_fs = NULL;
> > +
> > +    return retval;
> > +}
> > +#endif
> > diff --git a/lib/ext2fs/ext2fs.h b/lib/ext2fs/ext2fs.h
> > index 139a25fc..b1505f95 100644
> > --- a/lib/ext2fs/ext2fs.h
> > +++ b/lib/ext2fs/ext2fs.h
> > @@ -12,6 +12,10 @@
> > #ifndef _EXT2FS_EXT2FS_H
> > #define _EXT2FS_EXT2FS_H
> > 
> > +#ifdef HAVE_PTHREAD_H
> > +#include <pthread.h>
> > +#endif
> > +
> > #ifdef __GNUC__
> > #define EXT2FS_ATTR(x) __attribute__(x)
> > #else
> > @@ -331,6 +335,13 @@ struct struct_ext2_filsys {
> >    struct ext2fs_hashmap* block_sha_map;
> > 
> >    const struct ext2fs_nls_table *encoding;
> > +
> > +#ifdef HAVE_PTHREAD
> > +    struct struct_ext2_filsys *parent;
> > +    size_t refcount;
> > +    pthread_mutex_t refcount_mutex;
> > +    unsigned int clone_flags;
> > +#endif
> > };
> > 
> > #if EXT2_FLAT_INCLUDES
> > @@ -1057,6 +1068,18 @@ extern errcode_t ext2fs_move_blocks(ext2_filsys fs,
> > /* check_desc.c */
> > extern errcode_t ext2fs_check_desc(ext2_filsys fs);
> > 
> > +#ifdef HAVE_PTHREAD
> > +/* flags for ext2fs_clone_fs */
> > +#define EXT2FS_CLONE_BLOCK            0x0001
> > +#define EXT2FS_CLONE_INODE            0x0002
> > +#define EXT2FS_CLONE_BADBLOCKS        0x0004
> > +#define EXT2FS_CLONE_DBLIST            0x0008
> > +
> > +extern errcode_t ext2fs_clone_fs(ext2_filsys fs, ext2_filsys *dest,
> > +                                 unsigned int flags);
> > +extern errcode_t ext2fs_merge_fs(ext2_filsys *fs);
> > +#endif
> > +
> > /* closefs.c */
> > extern errcode_t ext2fs_close(ext2_filsys fs);
> > extern errcode_t ext2fs_close2(ext2_filsys fs, int flags);
> > -- 
> > 2.37.3
> > 

  reply	other threads:[~2022-11-19  5:02 UTC|newest]

Thread overview: 104+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-11-07 12:20 [RFCv1 00/72] e2fsprogs: Parallel fsck support Ritesh Harjani (IBM)
2022-11-07 12:20 ` [RFCv1 01/72] e2fsck: Fix unbalanced mutex unlock for BOUNCE_MTX Ritesh Harjani (IBM)
2022-11-17 16:02   ` Theodore Ts'o
2022-11-17 18:45     ` Ritesh Harjani (IBM)
2022-11-18 10:34   ` Andreas Dilger
2022-11-18 11:37     ` Ritesh Harjani (IBM)
2022-11-18 13:20       ` Andreas Dilger
2022-11-19  3:46         ` Ritesh Harjani (IBM)
2023-01-24 16:40   ` Theodore Ts'o
2022-11-07 12:20 ` [RFCv1 02/72] gen_bitmaps: Fix ext2fs_compare_generic_bmap/bitmap logic Ritesh Harjani (IBM)
2022-11-23  5:04   ` Andreas Dilger
2023-01-24 16:59     ` Theodore Ts'o
2022-11-07 12:20 ` [RFCv1 03/72] blkmap64_ba: Add common helper for bits size calculation Ritesh Harjani (IBM)
2022-11-18 10:40   ` Andreas Dilger
2022-11-07 12:20 ` [RFCv1 04/72] badblocks: Remove unused badblocks_flags Ritesh Harjani (IBM)
2022-11-18 13:26   ` Andreas Dilger
2022-11-07 12:20 ` [RFCv1 05/72] badblocks: Add badblocks merge logic Ritesh Harjani (IBM)
2022-11-18 13:31   ` Andreas Dilger
2022-11-07 12:20 ` [RFCv1 06/72] dblist: add dblist " Ritesh Harjani (IBM)
2022-11-18 13:34   ` Andreas Dilger
2022-11-07 12:20 ` [RFCv1 07/72] libext2fs: Add rbtree bitmap " Ritesh Harjani (IBM)
2022-11-07 12:20 ` [RFCv1 08/72] libext2fs: Add bitmaps merge ops Ritesh Harjani (IBM)
2022-11-18 13:36   ` Andreas Dilger
2022-11-07 12:20 ` [RFCv1 09/72] libext2fs: Add flush cleanup API Ritesh Harjani (IBM)
2022-11-18 13:39   ` Andreas Dilger
2022-11-07 12:20 ` [RFCv1 10/72] libext2fs: merge icounts after thread finishes Ritesh Harjani (IBM)
2022-11-18 13:40   ` Andreas Dilger
2022-11-07 12:20 ` [RFCv1 11/72] libext2fs: merge quota context after threads finish Ritesh Harjani (IBM)
2022-11-18 13:42   ` Andreas Dilger
2022-11-07 12:21 ` [RFCv1 12/72] libext2fs: dupfs: Add fs clone & merge api Ritesh Harjani (IBM)
2022-11-18 19:46   ` Andreas Dilger
2022-11-19  5:02     ` Ritesh Harjani (IBM) [this message]
2022-11-07 12:21 ` [RFCv1 13/72] tst_badblocks: Add unit test to verify badblocks list " Ritesh Harjani (IBM)
2022-12-12 20:35   ` Andreas Dilger
2022-11-07 12:21 ` [RFCv1 14/72] tst_bitmaps_standalone: Add copy and merge bitmaps test Ritesh Harjani (IBM)
2022-12-12 20:40   ` Andreas Dilger
2022-12-14  5:12     ` Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 15/72] tst_bitmaps_pthread: Add merge bitmaps test using pthreads Ritesh Harjani (IBM)
2022-12-14 21:15   ` Andreas Dilger
2022-11-07 12:21 ` [RFCv1 16/72] tst_libext2fs_pthread: Add libext2fs merge/clone unit tests Ritesh Harjani (IBM)
2022-12-14 21:17   ` Andreas Dilger
2022-11-07 12:21 ` [RFCv1 17/72] libext2fs: Add support for ext2fs_test_block_bitmap_range2_valid() Ritesh Harjani (IBM)
2022-12-14 21:21   ` Andreas Dilger
2022-11-07 12:21 ` [RFCv1 18/72] libext2fs: Add support to get average group count Ritesh Harjani (IBM)
2022-12-14 21:24   ` Andreas Dilger
2022-11-07 12:21 ` [RFCv1 19/72] libext2fs: Misc fixes for struct_ext2_filsys Ritesh Harjani (IBM)
2022-12-14 21:22   ` Andreas Dilger
2022-11-07 12:21 ` [RFCv1 20/72] libext2fs: avoid too much memory allocation in case fs_num_threads Ritesh Harjani (IBM)
2022-11-18 13:37   ` Andreas Dilger
2022-11-07 12:21 ` [RFCv1 21/72] e2fsck: add -m option for multithread Ritesh Harjani (IBM)
2022-12-14 21:32   ` Andreas Dilger
2022-11-07 12:21 ` [RFCv1 22/72] e2fsck: copy context when using multi-thread fsck Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 23/72] e2fsck: create logs for multi-threads Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 24/72] e2fsck: configure one pfsck thread Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 25/72] e2fsck: Add e2fsck_pass1_thread_join return value Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 26/72] e2fsck: Use merge/clone apis of libext2fs Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 27/72] e2fsck: Add e2fsck_pass1_merge_bitmap() api Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 28/72] e2fsck: Add asserts in open_channel_fs Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 29/72] e2fsck: add start/end group for thread Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 30/72] e2fsck: split groups to different threads Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 31/72] e2fsck: print thread log properly Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 32/72] e2fsck: do not change global variables Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 33/72] e2fsck: optimize the inserting of dir_info_db Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 34/72] e2fsck: merge dir_info after thread finishes Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 35/72] e2fsck: rbtree bitmap for dir Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 36/72] e2fsck: merge icounts after thread finishes Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 37/72] e2fsck: add debug codes for multiple threads Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 38/72] e2fsck: merge counts after threads finish Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 39/72] e2fsck: merge dx_dir_info " Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 40/72] e2fsck: merge dirs_to_hash when " Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 41/72] e2fsck: merge context flags properly Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 42/72] e2fsck: merge quota context after threads finish Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 43/72] e2fsck: serialize fix operations Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 44/72] e2fsck: move some fixes out of parallel pthreads Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 45/72] e2fsck: split and merge invalid bitmaps Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 46/72] e2fsck: merge EA blocks properly Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 47/72] e2fsck: kickoff mutex lock for block found map Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 48/72] e2fsck: allow admin specify number of threads Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 49/72] e2fsck: adjust " Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 50/72] e2fsck: fix readahead for pfsck of pass1 Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 51/72] e2fsck: merge options after threads finish Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 52/72] e2fsck: reset lost_and_found " Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 53/72] e2fsck: merge extent depth count " Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 54/72] e2fsck: simplify e2fsck context merging codes Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 55/72] e2fsck: set E2F_FLAG_ALLOC_OK after threads Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 56/72] e2fsck: wait fix thread finish before checking Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 57/72] e2fsck: cleanup e2fsck_pass1_thread_join() Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 58/72] e2fsck: make default smallest RA size to 1M Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 59/72] e2fsck: update mmp block in one thread Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 60/72] e2fsck: reset @inodes_to_rebuild if restart Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 61/72] tests: add pfsck test Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 62/72] e2fsck: fix memory leaks with pfsck enabled Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 63/72] e2fsck: misc cleanups for pfsck Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 64/72] e2fsck: propagate number of threads Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 65/72] e2fsck: Annotating fields in e2fsck_struct Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 66/72] e2fsck: merge casefolded dir lists after thread finish Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 67/72] sec: support encrypted files handling in pfsck mode Ritesh Harjani (IBM)
2022-11-07 19:22   ` Eric Biggers
2022-11-07 12:21 ` [RFCv1 68/72] e2fsck: Fix io->align assert check Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 69/72] e2fsck: Fix double free of inodes_to_process Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 70/72] e2fsck: Fix and simplify update_mmp in case of pfsck Ritesh Harjani (IBM)
2022-11-07 12:21 ` [RFCv1 71/72] e2fsck: Make threads call log_out after pthread_join Ritesh Harjani (IBM)
2022-11-07 12:22 ` [RFCv1 72/72] tests/f_multithread: Fix f_multithread related tests Ritesh Harjani (IBM)
     [not found] ` <B4ED1C86-D3EC-4A0A-97B3-CFCB46617E1A@dilger.ca>
2022-11-19  5:39   ` [RFCv1 00/72] e2fsprogs: Parallel fsck support Ritesh Harjani

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20221119050229.xznagyl2v6bmqgvt@riteshh-domain \
    --to=ritesh.list@gmail.com \
    --cc=adilger.kernel@dilger.ca \
    --cc=adilger@dilger.ca \
    --cc=harshadshirwadkar@gmail.com \
    --cc=linux-ext4@vger.kernel.org \
    --cc=lixi@ddn.com \
    --cc=saranyamohan@google.com \
    --cc=tytso@mit.edu \
    --cc=wshilong@ddn.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).