[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <7165b24e-0628-c273-f1e2-67a75358991f@kernel.org>
Date: Thu, 10 Jun 2021 11:54:50 -0700
From: Nathan Chancellor <nathan@...nel.org>
To: Peter Zijlstra <peterz@...radead.org>
Cc: x86@...nel.org, jpoimboe@...hat.com, jbaron@...mai.com,
rostedt@...dmis.org, ardb@...nel.org, linux-kernel@...r.kernel.org,
samitolvanen@...gle.com, ndesaulniers@...gle.com,
clang-built-linux@...glegroups.com
Subject: Re: [PATCH 01/13] objtool: Rewrite hashtable sizing
On 6/10/2021 11:43 AM, Peter Zijlstra wrote:
> On Thu, Jun 10, 2021 at 11:14:51AM -0700, Nathan Chancellor wrote:
>
>> This patch as commit 25cf0d8aa2a3 ("objtool: Rewrite hashtable sizing")
>> in -tip causes a massive compile time regression with allmodconfig +
>> ThinLTO.
>
> Moo... the allyesconfig builds I used it on were much faster, but that
> was on regular GCC vmlinux.o after linking.
>
>> Adding Sami because I am not sure why this patch would have much of an impact
>> in relation to LTO. https://git.kernel.org/tip/25cf0d8aa2a3 is the patch in
>> question.
>>
>> If I can provide any further information or help debug, please let me know.
>>
>> If you are interested in reproducing this locally, you will need a
>> fairly recent LLVM stack (I used the stable release/12.x branch) and to
>> cherry-pick commit 976aac5f8829 ("kcsan: Fix debugfs initcall return
>> type") to fix an unrelated build failure. My script [2] can build a
>> self-contained toolchain fairly quickly if you cannot get one from your
>> package manager. A command like below will speed up the build a bit:
>
> Would something like llvm-13 from Debian be good enough?
>
> $ clang-13 --version
> Debian clang version 13.0.0-++20210418105309+a0898f0cecc7-1~exp1
> Target: x86_64-pc-linux-gnu
> Thread model: posix
> InstalledDir: /usr/bin
>
Yes, that would work. That is what we use in our CI.
Looks like Sami gave a reply that explains it.
Cheers,
Nathan
Powered by blists - more mailing lists