linux-kernel - Re: [PATCH 0/7] bcachefs compiler warning fixes for 32-bit

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]

Message-ID: <20230912193642.sdlwon4jta4ob5ac@moria.home.lan>
Date:   Tue, 12 Sep 2023 15:36:42 -0400
From:   Kent Overstreet <kent.overstreet@...ux.dev>
To:     Nathan Chancellor <nathan@...nel.org>
Cc:     bfoster@...hat.com, linux-bcachefs@...r.kernel.org,
        linux-kernel@...r.kernel.org, llvm@...ts.linux.dev,
        patches@...ts.linux.dev
Subject: Re: [PATCH 0/7] bcachefs compiler warning fixes for 32-bit

On Tue, Sep 12, 2023 at 12:15:37PM -0700, Nathan Chancellor wrote:
> Hi all,
> 
> This is a series of fixes for warnings that I now see from bcachefs when
> building my test matrix with LLVM in -next, mostly from 32-bit
> architectures. Most of the patches should be uncontroversial; the
> min_t/max_t changes are probably the worst ones.

Those patches all look fine, thanks

> I still see several instances of -Wframe-larger-than when building for
> 32-bit ARM (I am sure they will show up on other 32-bit architectures as
> well), which I am not entirely sure how to tackle. It looks like the
> majority of the instances are just due to large structures on the stack
> in combination with some inlining resulting in some spills, so it seems
> like moving to a heap allocation for some of those would the right fix
> but I know some maintainers would rather fix them in their own way,
> hence just the report.

I just started seeing these again as well. The biggest single object
that most of them have on their stack is btree_trans, but initializing
that has to do a heap allocation for the btree_paths array - so we might
as well switch to heap allocating the entire thing.

That will help.