[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <a984d174-0842-407f-8af9-ed1c506e442b@intel.com>
Date: Thu, 13 Jun 2024 12:43:35 -0700
From: Dave Hansen <dave.hansen@...el.com>
To: Naresh Kamboju <naresh.kamboju@...aro.org>, X86 ML <x86@...nel.org>,
open list <linux-kernel@...r.kernel.org>, lkft-triage@...ts.linaro.org,
Linux Regressions <regressions@...ts.linux.dev>
Cc: Thomas Gleixner <tglx@...utronix.de>, Borislav Petkov <bp@...en8.de>,
Dave Hansen <dave.hansen@...ux.intel.com>,
Peter Zijlstra <peterz@...radead.org>, Arnd Bergmann <arnd@...db.de>,
Dan Carpenter <dan.carpenter@...aro.org>,
Anders Roxell <anders.roxell@...aro.org>
Subject: Re: Bad FPU state detected at restore_fpregs_from_fpstate,
reinitializing FPU registers. WARNING: at arch/x86/mm/extable.c:127
fixup_exception
Does this fix it?
I think moving the explicit 'struct fpu' out of task_struct took the
knowledge away from the compiler on how to keep the XSAVE buffer
aligned. Once that happened, we ended up with unaligned XSAVE
operations and bad things happened.
Also, open-coding "task + sizeof(*task)" in three different places seems
suboptimal.
View attachment "debug-32bit-fpregs.patch" of type "text/x-patch" (2502 bytes)
Powered by blists - more mailing lists