Re: [PATCH v3 00/11] Performance fixes for 9p filesystem

[Eric Van Hensbergen]

Hi Christian, thanks for the feedback -- will dig in and see if I can
find what's gone south here.  Clearly my approach to writeback without
writeback_fid didn't cover all the corner cases and thats the cause of
the fault.  Can I get a better idea of how to reproduce - you booted
with a root 9p file system, and then tried to build...what?

Performance degradation is interesting, runs counter to the
unit-testing and benchmarking I did, but I didn't do something as
logical as a build to check -- need to tease apart whether this is a
read problem, a write problem...or both.  My intuition is that its on
the write side, but as part of going through the code I made the cache
code a lot more pessimistic so its possible I inadvertently killed an
optimistic optimization.

Finally, just to clarify, the panic you had at the end happened with
readahead?  Seems interesting because clearly it thought it was
writing back something that it shouldn't have been writing back (since
writeback caches weren't enabled).   I'm thinking something was marked
as dirty even though the underlying system just wrote-through the
change and so the writeback isn't actually required.  This may also be
an indicator of the performance issue if we are actually writing
through the data in addition to an unnecessary write-back (which I
also worry is writing back bad data in the second case).

Can you give me an idea of what the other misbehaviors were?

      -eric

On Thu, Feb 2, 2023 at 5:27 AM Christian Schoenebeck
<linux_oss@xxxxxxxxxxxxx> wrote:
>
> On Tuesday, January 24, 2023 3:38:23 AM CET Eric Van Hensbergen wrote:
> > This is the third version of a patch series which adds a number
> > of features to improve read/write performance in the 9p filesystem.
> > Mostly it focuses on fixing caching to help utilize the recently
> > increased MSIZE limits and also fixes some problematic behavior
> > within the writeback code.
> >
> > All together, these show roughly 10x speed increases on simple
> > file transfers.  Future patch sets will improve cache consistency
> > and directory caching.
> >
> > These patches are also available on github:
> > https://github.com/v9fs/linux/tree/ericvh/for-next
> > and on kernel.org:
> > https://git.kernel.org/pub/scm/linux/kernel/git/ericvh/v9fs.git
> >
> > Tested against qemu, cpu, and diod with fsx, dbench, and some
> > simple benchmarks.
>
> Looks like this needs more work.
>
> I only had a glimpse on your patches yet, but made some tests by doing
> compilations on guest on top of a 9p root fs [1], msize=500k. Under that
> scenario:
>
> * loose: this is suprisingly the only mode where I can see some performance
> increase, over "loose" on master it compiled ~5% faster, but I also got some
> misbehaviours on guest.
>
> * writeahead: no performance results, as compilation already aborts when
> trying to detect a compiler. I had to restore a previous snapshot to repair
> things after this test.
>
> * readahead: significant performance drop. In comparison to "loose" on master
> compilation takes 6 times longer with "readahead". There are some severe
> misbehaviours on guest as well, and after boot I get this warning:
>
> [    5.782846] folio expected an open fid inode->i_private=0000000000000000
> [    5.786641] WARNING: CPU: 0 PID: 321 at fs/9p/vfs_addr.c:174 v9fs_vfs_write_folio_locked (fs/9p/vfs_addr.c:174 (discriminator 3)) 9p
> [    5.792496] Modules linked in: ppdev(E) bochs(E) sg(E) drm_vram_helper(E) joydev(E) evdev(E) drm_kms_helper(E) serio_raw(E) drm_ttm_helper(E) pcsp)
> [    5.816806] CPU: 0 PID: 321 Comm: chown Tainted: G            E      6.2.0-rc6+ #61
> [    5.821694] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.16.1-0-g3208b098f51a-prebuilt.qemu.org 04/01/2014
> [    5.827362] RIP: 0010:v9fs_vfs_write_folio_locked (fs/9p/vfs_addr.c:174 (discriminator 3)) 9p
>
> Code starting with the faulting instruction
> ===========================================
>         ...
> [    5.835360] RSP: 0018:ffffc900007d3a38 EFLAGS: 00010282
> [    5.836982] RAX: 0000000000000000 RBX: ffff888106c86680 RCX: 0000000000000000
> [    5.838877] RDX: 0000000000000001 RSI: ffffffff821eb1e6 RDI: 00000000ffffffff
> [    5.841179] RBP: ffffea0004279300 R08: 0000000000000000 R09: 00000000ffffefff
> [    5.843039] R10: ffffc900007d38e8 R11: ffffffff824bede8 R12: 0000000000000000
> [    5.844850] R13: 00000000ffffffea R14: 0000000000000014 R15: 0000000000000014
> [    5.846366] FS:  00007fd0fc4a0580(0000) GS:ffff88842fc00000(0000) knlGS:0000000000000000
> [    5.848250] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> [    5.849386] CR2: 00007fd0fc38f4f0 CR3: 0000000100302000 CR4: 00000000000006f0
> [    5.850824] Call Trace:
> [    5.851622]  <TASK>
> [    5.852052] v9fs_vfs_writepage (fs/9p/vfs_addr.c:207) 9p
> [    5.852841] __writepage (mm/page-writeback.c:2537)
> [    5.853438] write_cache_pages (mm/page-writeback.c:2473)
> [    5.854205] ? __pfx___writepage (mm/page-writeback.c:2535)
> [    5.855309] ? delete_node (lib/radix-tree.c:575)
> [    5.856122] ? radix_tree_delete_item (lib/radix-tree.c:1432)
> [    5.857101] do_writepages (mm/page-writeback.c:2564 mm/page-writeback.c:2552 mm/page-writeback.c:2583)
> [    5.857954] ? radix_tree_delete_item (lib/radix-tree.c:1432)
> [    5.859103] filemap_fdatawrite_wbc (mm/filemap.c:389 mm/filemap.c:378)
> [    5.860043] __filemap_fdatawrite_range (mm/filemap.c:422)
> [    5.861050] filemap_write_and_wait_range (mm/filemap.c:682 mm/filemap.c:665)
> [    5.862132] v9fs_vfs_setattr_dotl (./include/linux/pagemap.h:60 fs/9p/vfs_inode_dotl.c:583) 9p
> [    5.863312] notify_change (fs/attr.c:486)
> [    5.864043] ? chown_common (fs/open.c:736)
> [    5.864793] chown_common (fs/open.c:736)
> [    5.865538] ? preempt_count_add (./include/linux/ftrace.h:977 kernel/sched/core.c:5737 kernel/sched/core.c:5734 kernel/sched/core.c:5762)
> [    5.866420] do_fchownat (fs/open.c:768)
> [    5.867419] __x64_sys_fchownat (fs/open.c:782 fs/open.c:779 fs/open.c:779)
> [    5.868319] do_syscall_64 (arch/x86/entry/common.c:50 arch/x86/entry/common.c:80)
> [    5.869116] entry_SYSCALL_64_after_hwframe (arch/x86/entry/entry_64.S:120)
> [    5.871008] RIP: 0033:0x7fd0fc3b7b9a
>
> Best regards,
> Christian Schoenebeck
>
> [1] https://wiki.qemu.org/Documentation/9p_root_fs
>
>
>