[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <qhj7tqo35lo3fhl7ne5yl4pw6kp45c6owc75f4cj2gyxot2ldo@aizhulohwmmu>
Date: Wed, 3 Dec 2025 11:17:53 -0800
From: Josh Poimboeuf <jpoimboe@...nel.org>
To: Linus Torvalds <torvalds@...ux-foundation.org>
Cc: Ingo Molnar <mingo@...nel.org>, linux-kernel@...r.kernel.org,
linux-tip-commits@...r.kernel.org, x86@...nel.org
Subject: Re: [tip: objtool/urgent] objtool: Consolidate annotation macros
On Wed, Dec 03, 2025 at 10:41:30AM -0800, Linus Torvalds wrote:
> On Wed, 3 Dec 2025 at 09:46, Josh Poimboeuf <jpoimboe@...nel.org> wrote:
> >
> > So that mergeable thing is the only way to convince the toolchains to
> > allow setting the section entsize
>
> Ugh. What a horrid reason. That *should* be trivially done by just
> having a linker script setting, but if one doesn't exist...
Believe me, I looked and there's surprisingly no way to do that from
linker script.
> What are the actual entry sizes, though? Because maybe we could just
> use alignment instead - which is trivially settable in the linker
> script.
That could work. At least .altinstructions (14 bytes) and __ex_table
(12 bytes) aren't power of two.
We could certainly pad them, That does increase memory usage though. On
my defconfig kernel that would add about ~11k.
--
Josh
Powered by blists - more mailing lists