linux-kernel - Re: [PATCH v2 06/12] gfs2: Replace the repetitive bio chaining code patterns

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <CANubcdVAitTW_aBqwxC=TV77rg_iie0uX54_qEtMCjgdN+zeig@mail.gmail.com>
Date: Sun, 7 Dec 2025 18:03:53 +0800
From: Stephen Zhang <starzhangzsd@...il.com>
To: Andreas Gruenbacher <agruenba@...hat.com>, sashal@...nel.org
Cc: Johannes.Thumshirn@....com, hch@...radead.org, ming.lei@...hat.com, 
	Gao Xiang <hsiangkao@...ux.alibaba.com>, linux-block@...r.kernel.org, 
	linux-bcache@...r.kernel.org, nvdimm@...ts.linux.dev, 
	virtualization@...ts.linux.dev, linux-nvme@...ts.infradead.org, 
	gfs2@...ts.linux.dev, ntfs3@...ts.linux.dev, linux-xfs@...r.kernel.org, 
	linux-kernel@...r.kernel.org, zhangshida@...inos.cn
Subject: Re: [PATCH v2 06/12] gfs2: Replace the repetitive bio chaining code patterns

Andreas Gruenbacher <agruenba@...hat.com> 于2025年12月5日周五 16:55写道：
>
> On Fri, Dec 5, 2025 at 8:46 AM Stephen Zhang <starzhangzsd@...il.com> wrote:
> > Andreas Gruenbacher <agruenba@...hat.com> 于2025年12月4日周四 17:37写道：
> > >
> > > On Mon, Dec 1, 2025 at 11:31 AM Andreas Gruenbacher <agruenba@...hat.com> wrote:
> > > > On Sat, Nov 29, 2025 at 3:48 AM Stephen Zhang <starzhangzsd@...il.com> wrote:
> > > > > This one should also be dropped because the 'prev' and 'new' are in
> > > > > the wrong order.
> > > >
> > > > Ouch. Thanks for pointing this out.
> > >
> > > Linus has merged the fix for this bug now, so this patch can be
> > > updated / re-added.
> > >
> >
> > Thank you for the update. I'm not clear on what specifically has been
> > merged or how to verify it.
> > Could you please clarify which fix was merged,
>
> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=8a157e0a0aa5
> "gfs2: Fix use of bio_chain"
>
>

Thank you for the detailed clarification. Here is a polite and
professional rephrasing of your message:

---

[WARNING]
Hello,

I may not have expressed myself clearly, and you might have
misunderstood my point.

In the original code, the real end I/O handler (`gfs2_end_log_read`)
was placed at the end of the chained bio list, while the newer
`bio_chain_endio` was placed earlier. With `bio_chain(new, prev)`,
the chain looked like:

`bio1 → bio2 → bio3`
`bio_chain_endio → bio_chain_endio → gfs2_end_log_read`

This ensured the actual handler (`gfs2_end_log_read`) was triggered
at the end of the chain.

However, after the fix changed the order to `bio_chain(prev, new)`,
the chain now looks like:

`bio1 → bio2 → bio3`
`gfs2_end_log_read → bio_chain_endio → bio_chain_endio`

This seems to place `gfs2_end_log_read` at the beginning rather
than the end, potentially preventing it from being executed as intended.

I hope I misunderstand the gfs2 code logic, and your fix may still be
correct. However, given how quickly the change was made and ported
back, I wanted to highlight this concern in case the original behavior
was intentional.

Thank you for your attention to this matter.

Best regards,
Shida

> > and if I should now resubmit the cleanup patches?
> >
> > Thanks,
> > Shida
> >
> > > Thanks,
> > > Andreas
> > >
> >
>