Tarantool development patches archive
 help / color / mirror / Atom feed
From: Sergey Kaplun via Tarantool-patches <tarantool-patches@dev.tarantool.org>
To: Sergey Bronnikov <sergeyb@tarantool.org>
Cc: tarantool-patches@dev.tarantool.org
Subject: Re: [Tarantool-patches] [PATCH luajit 2/2] Different fix for partial snapshot restore due to stack overflow.
Date: Tue, 10 Jun 2025 19:22:32 +0300	[thread overview]
Message-ID: <aEhbyNWyhdmvmsYt@root> (raw)
In-Reply-To: <9d1c7b94-b18d-4290-b017-27ac9e4f6ef2@tarantool.org>

Hi, Sergey!
Thanks for the review!
Please, consider my answer below.

On 10.06.25, Sergey Bronnikov wrote:
> Hello, Sergey!
> 
> Thanks for the patch! See a comment below.
> 
> Sergey
> 
> On 6/10/25 13:28, Sergey Kaplun wrote:
> > From: Mike Pall <mike>
> >
> > Reported by Junlong Li. Fixed by Peter Cawley.
> >
> > (cherry picked from commit 86e7123bb1782a5f200ba5e83b8c4f3fbad4f7bc)
> >
> > This patch is a follow-up to the previous commit, which leads to a dirty
> > read of the pseudo-valid PC set for the cframe on snapshot restoration.
> > To avoid these dirty reads, this patch sets the PC to the outer frame
> > as well before possible error throwing.
> >
> > Sergey Kaplun:
> > * added the description and the test for the problem
> >
> > Part of tarantool/tarantool#11278
> > ---
> >   src/lj_snap.c                                 |  3 +-
> >   src/lj_trace.c                                |  4 +-
> >   ...-1196-stack-overflow-snap-restore.test.lua | 65 +++++++++++++++++++
> >   3 files changed, 68 insertions(+), 4 deletions(-)
> >   create mode 100644 test/tarantool-tests/lj-1196-stack-overflow-snap-restore.test.lua
> >
> > diff --git a/src/lj_snap.c b/src/lj_snap.c
> > index 8d7bd868..4cfae579 100644
> > --- a/src/lj_snap.c
> > +++ b/src/lj_snap.c

<snipped>

> > diff --git a/test/tarantool-tests/lj-1196-stack-overflow-snap-restore.test.lua b/test/tarantool-tests/lj-1196-stack-overflow-snap-restore.test.lua
> > new file mode 100644
> > index 00000000..942d1f82
> > --- /dev/null
> > +++ b/test/tarantool-tests/lj-1196-stack-overflow-snap-restore.test.lua
> > @@ -0,0 +1,65 @@
> > +local tap = require('tap')
> > +
> > +-- Test file to demonstrate LuaJIT dirty reads after stack
> > +-- overflow during restoration from the snapshot.
> > +-- The test fails before the patch under Valgrind.
> 
> Please specify valgrind option that is required for reproducing the bug.

This just mean the default valgrind settings (without any flags).

> 
> Cannot reproduce with command line below with reverted patch:
> 
> VALGRIND_OPTS="--leak-check=no --malloc-fill=0x00 --free-fill=0x00" 
> ctest -V -R test/tarantool-tests/lj-1196-partial-snap-restore.test.lua -V

This doesn't fail due to --leadk-check=no flag.
Also, I suppose this should fail under ASan.

> 
> > +--
> > +-- luacheck: push no max_comment_line_length
> > +--
> > +-- See also:https://github.com/LuaJIT/LuaJIT/issues/1196,
> > +--https://www.freelists.org/post/luajit/Invalid-read-found-by-valgrind.
> > +--
> > +-- luacheck: pop
> > +

<snipped>

> > +test:done(true)

-- 
Best regards,
Sergey Kaplun

  reply	other threads:[~2025-06-10 16:22 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-06-10 10:28 [Tarantool-patches] [PATCH luajit 0/2] Snapshot restore with " Sergey Kaplun via Tarantool-patches
2025-06-10 10:28 ` [Tarantool-patches] [PATCH luajit 1/2] Handle partial snapshot restore due to " Sergey Kaplun via Tarantool-patches
2025-06-10 15:29   ` Sergey Bronnikov via Tarantool-patches
2025-06-10 16:03     ` Sergey Kaplun via Tarantool-patches
2025-06-10 10:28 ` [Tarantool-patches] [PATCH luajit 2/2] Different fix for " Sergey Kaplun via Tarantool-patches
2025-06-10 16:15   ` Sergey Bronnikov via Tarantool-patches
2025-06-10 16:22     ` Sergey Kaplun via Tarantool-patches [this message]
2025-06-11  9:36       ` Sergey Bronnikov via Tarantool-patches

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=aEhbyNWyhdmvmsYt@root \
    --to=tarantool-patches@dev.tarantool.org \
    --cc=sergeyb@tarantool.org \
    --cc=skaplun@tarantool.org \
    --subject='Re: [Tarantool-patches] [PATCH luajit 2/2] Different fix for partial snapshot restore due to stack overflow.' \
    /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

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox