linux-kernel - Re: [GIT PULL] Ceph updates for 5.20-rc1

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

Date:   Thu, 11 Aug 2022 14:29:34 -0700
From:   Linus Torvalds <torvalds@...ux-foundation.org>
To:     Ilya Dryomov <idryomov@...il.com>
Cc:     ceph-devel@...r.kernel.org, linux-kernel@...r.kernel.org,
        Jeff Layton <jlayton@...nel.org>,
        Al Viro <viro@...iv.linux.org.uk>,
        Matthew Wilcox <willy@...radead.org>
Subject: Re: [GIT PULL] Ceph updates for 5.20-rc1

On Thu, Aug 11, 2022 at 1:56 PM Ilya Dryomov <idryomov@...il.com> wrote:
>
> On Thu, Aug 11, 2022 at 10:04 PM Linus Torvalds
> <torvalds@...ux-foundation.org> wrote:
> >
> > What's the point of warning about bogus folios more than once? That's
> > a debug warning - if it hits even once, that's already "uhhuh,
> > something is bad". Showing the warning more than once is likely just
> > going to cause more problems, not give you more information.
>
> Xiubo and Jeff used it to track down some issues between netfs library
> and folio code that have been randomly plaguing our automated tests for
> a couple of releases.  We already knew that there were issues in that
> area and the actual occurrences mattered.  This was done in cooperation
> with Willy and, since he was involved and this is a no-impact change,
> I didn't think twice.

I don't mind the warning.

I mind the "more than ONCE" part.

If it's a "this shouldn't happen, but if it ever happens I want to
know about it" situation, then the ONCE variant should be what you
want.

And that variant already existed, and adding a new and inferior macro
seems to just have been pointless.

As to dcache issues - if you really don't get an ack from Al, you can
at least make me aware of it before-hand. That's one of the files that
I at least personally care about, and while I would much prefer an ack
from Al for anything that touches it, at least I'll likely be less
unhappy about changes if I was made aware of them ahead of time,
instead of seeing a pull request that suddenly changes that file.

               Linus