From: Mark Rutland <mark.rutland@arm.com>
To: Will Deacon <will@kernel.org>
Cc: linux-arm-kernel@lists.infradead.org, broonie@kernel.org,
catalin.marinas@arm.com, daniel.kiss@arm.com,
david.spickett@arm.com, luis.machado@arm.com, maz@kernel.org,
richard.sandiford@arm.com, sander.desmalen@arm.com,
tabba@google.com, tamas.petz@arm.com, tkjos@google.com,
yury.khrustalev@arm.com
Subject: Re: [PATCH 03/20] arm64/fpsimd: signal: Clear PSTATE.SM when restoring FPSIMD frame only
Date: Wed, 7 May 2025 15:01:50 +0100 [thread overview]
Message-ID: <aBtnzhmqfdbs7hhT@J2N7QTR9R3> (raw)
In-Reply-To: <20250507124644.GA2227@willie-the-truck>
On Wed, May 07, 2025 at 01:46:45PM +0100, Will Deacon wrote:
> On Tue, May 06, 2025 at 04:25:06PM +0100, Mark Rutland wrote:
> > diff --git a/arch/arm64/kernel/signal.c b/arch/arm64/kernel/signal.c
> > index 48d3c0129dade..fdce1b856f498 100644
> > --- a/arch/arm64/kernel/signal.c
> > +++ b/arch/arm64/kernel/signal.c
> > @@ -280,6 +280,7 @@ static int restore_fpsimd_context(struct user_ctxs *user)
> > __get_user_error(fpsimd.fpcr, &(user->fpsimd->fpcr), err);
> >
> > clear_thread_flag(TIF_SVE);
> > + current->thread.svcr &= ~SVCR_SM_MASK;
> > current->thread.fp_type = FP_STATE_FPSIMD;
>
> Hmm, I think we're preemptible here so do we need some compiler barriers
> to make sure that the context-switching code doesn't see these fields in
> an inconsistent state?
We avoid that problem by ensuring that the task's FPSIMD/SVE/SME state isn't
live on any CPU for the duration of the signal code, such that it is safe to
manipulate the saved state in preemptible context. Details below.
That was handled in one of the earlier patches in the arm64 for-next/sme-fixes
branch, queded as commit:
929fa99b1215966f ("arm64/fpsimd: signal: Always save+flush state early")
As of that commit, in setup_rt_frame() we call
fpsimd_save_and_flush_current_state(), which does:
get_cpu_fpsimd_context();
fpsimd_save_user_state();
fpsimd_flush_task_state(current);
put_cpu_fpsimd_context();
That ensures (in a premption-safe way) that the tasks state has been saved
to memory, TIF_FOREIGN_FPSTATE is set, and that
task->thread.fpsimd_cpu==NR_CPUS. Thus on context switch:
* When switching away from this task, fpsimd_thread_switch() sees
TIF_FOREIGN_FPSTATE is set, and doesn't manipulate the task's saved state.
* When switching to this task, fpsimd_thread_switch() doesn't find any existing
state bound on the CPU, and sets TIF_FOREIGN_FPSTATE, without manipulating
the task's saved state.
The task's state will only be relaoded onto the CPU when returning to
userspace, via fpsimd_restore_current_state().
Note that we changed fpsimd_update_current_state() to modify the saved state
without reloading it onto the CPU.
Mark.
next prev parent reply other threads:[~2025-05-07 14:41 UTC|newest]
Thread overview: 44+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-05-06 15:25 [PATCH 00/20] arm64: FPSIMD/SVE/SME fixes + re-eanble SME Mark Rutland
2025-05-06 15:25 ` [PATCH 01/20] kselftest/arm64: fp-ptrace: Fix expected FPMR value when PSTATE.SM is changed Mark Rutland
2025-05-06 15:25 ` [PATCH 02/20] arm64/fpsimd: Do not discard modified SVE state Mark Rutland
2025-05-06 15:25 ` [PATCH 03/20] arm64/fpsimd: signal: Clear PSTATE.SM when restoring FPSIMD frame only Mark Rutland
2025-05-07 12:46 ` Will Deacon
2025-05-07 14:01 ` Mark Rutland [this message]
2025-05-07 14:39 ` Will Deacon
2025-05-06 15:25 ` [PATCH 04/20] arm64/fpsimd: signal: Mandate SVE payload for streaming-mode state Mark Rutland
2025-05-07 0:59 ` Mark Brown
2025-05-07 12:59 ` Will Deacon
2025-05-07 14:21 ` Mark Rutland
2025-05-07 14:29 ` Will Deacon
2025-05-07 15:02 ` Mark Rutland
2025-05-07 16:14 ` Will Deacon
2025-05-06 15:25 ` [PATCH 05/20] arm64/fpsimd: ptrace: Consistently handle partial writes to NT_ARM_(S)SVE Mark Rutland
2025-05-06 15:25 ` [PATCH 06/20] arm64/fpsimd: Clarify sve_sync_*() functions Mark Rutland
2025-05-06 15:25 ` [PATCH 07/20] arm64/fpsimd: Factor out {sve,sme}_state_size() helpers Mark Rutland
2025-05-06 15:25 ` [PATCH 08/20] arm64/fpsimd: Add task_smstop_sm() Mark Rutland
2025-05-06 15:25 ` [PATCH 09/20] arm64/fpsimd: signal: Use SMSTOP behaviour in setup_return() Mark Rutland
2025-05-06 15:25 ` [PATCH 10/20] arm64/fpsimd: Remove redundant task->mm check Mark Rutland
2025-05-06 15:25 ` [PATCH 11/20] arm64/fpsimd: Consistently preserve FPSIMD state during clone() Mark Rutland
2025-05-06 15:25 ` [PATCH 12/20] arm64/fpsimd: Clear PSTATE.SM " Mark Rutland
2025-05-06 15:25 ` [PATCH 13/20] arm64/fpsimd: Make clone() compatible with ZA lazy saving Mark Rutland
2025-05-07 14:58 ` Will Deacon
2025-05-07 15:22 ` Mark Rutland
2025-05-07 16:11 ` Will Deacon
2025-05-07 17:21 ` Mark Rutland
2025-05-07 15:57 ` Yury Khrustalev
2025-05-06 15:25 ` [PATCH 14/20] arm64/fpsimd: ptrace/prctl: Ensure VL changes do not resurrect stale data Mark Rutland
2025-05-06 15:25 ` [PATCH 15/20] arm64/fpsimd: ptrace/prctl: Ensure VL changes leave task in a valid state Mark Rutland
2025-05-07 16:12 ` Will Deacon
2025-05-07 17:10 ` Mark Rutland
2025-05-08 10:31 ` Will Deacon
2025-05-06 15:25 ` [PATCH 16/20] arm64/fpsimd: ptrace: Save task state before generating SVE header Mark Rutland
2025-05-06 15:25 ` [PATCH 17/20] arm64/fpsimd: ptrace: Do not present register data for inactive mode Mark Rutland
2025-05-06 15:25 ` [PATCH 18/20] arm64/fpsimd: ptrace: Mandate SVE payload for streaming-mode state Mark Rutland
2025-05-07 1:09 ` Mark Brown
2025-05-06 15:25 ` [PATCH 19/20] arm64/fpsimd: ptrace: Gracefully handle errors Mark Rutland
2025-05-07 16:32 ` Will Deacon
2025-05-08 12:12 ` Mark Rutland
2025-05-06 15:25 ` [PATCH 20/20] arm64/fpsimd: Allow CONFIG_ARM64_SME to be selected Mark Rutland
2025-05-07 1:48 ` [PATCH 00/20] arm64: FPSIMD/SVE/SME fixes + re-eanble SME Mark Brown
2025-05-07 9:56 ` Mark Rutland
2025-05-07 11:26 ` Mark Brown
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=aBtnzhmqfdbs7hhT@J2N7QTR9R3 \
--to=mark.rutland@arm.com \
--cc=broonie@kernel.org \
--cc=catalin.marinas@arm.com \
--cc=daniel.kiss@arm.com \
--cc=david.spickett@arm.com \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=luis.machado@arm.com \
--cc=maz@kernel.org \
--cc=richard.sandiford@arm.com \
--cc=sander.desmalen@arm.com \
--cc=tabba@google.com \
--cc=tamas.petz@arm.com \
--cc=tkjos@google.com \
--cc=will@kernel.org \
--cc=yury.khrustalev@arm.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox