All of lore.kernel.org
 help / color / mirror / Atom feed
From: Uros Bizjak <ubizjak@gmail.com>
To: x86@kernel.org, linux-kernel@vger.kernel.org
Cc: Uros Bizjak <ubizjak@gmail.com>,
	Thomas Gleixner <tglx@kernel.org>, Ingo Molnar <mingo@kernel.org>,
	Borislav Petkov <bp@alien8.de>,
	Dave Hansen <dave.hansen@linux.intel.com>,
	"H. Peter Anvin" <hpa@zytor.com>
Subject: [PATCH -tip 0/1] x86_32/segment: Always return correctly zero-extended values from savesegment_*()
Date: Tue, 14 Apr 2026 09:44:41 +0200	[thread overview]
Message-ID: <20260414074613.77330-1-ubizjak@gmail.com> (raw)

On 32-bit x86, reading segment registers into a 32-bit type can expose
undefined upper bits on older processors (e.g. Intel Quark X1000,
Pentium, and earlier), where bits 31:16 are not defined.

As a result, undefined upper bits may be stored in variables
(e.g. saved_fs and saved_gs in apm_32.c), structures (e.g. thread_struct.gs)
and arrays (e.g. elf_gregset_t[]). Some of these are later used, for
example in process_32.c::__switch_to():

	if (prev->gs | next->gs)
		loadsegment(gs, next->gs);

which results in unneccessary reloads.

Introduce __seg_return_t as an intermediate type for __savesegment_*():
u16 on CONFIG_X86_32 and unsigned long otherwise.
    
This ensures that __savesegment_*() returns correctly zero-extended
values in all cases, avoiding propagation of undefined high bits on
32-bit systems while preserving existing behavior on 64-bit.

A follow-up patch can remove the majority of zero-extensions by addressing
a long-standing XXX comment in thread_struct and reusing gsindex instead of
gs for 32-bit x86 targets as well.

Cc: Thomas Gleixner <tglx@kernel.org>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>

Uros Bizjak (1):
  x86_32/segment: Always return correctly zero-extended values from
    savesegment_*()

 arch/x86/include/asm/segment.h | 12 +++++++++++-
 1 file changed, 11 insertions(+), 1 deletion(-)

-- 
2.53.0


             reply	other threads:[~2026-04-14  7:46 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-04-14  7:44 Uros Bizjak [this message]
2026-04-14  7:44 ` [PATCH -tip 1/1] x86_32/segment: Always return correctly zero-extended values from savesegment_*() Uros Bizjak

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=20260414074613.77330-1-ubizjak@gmail.com \
    --to=ubizjak@gmail.com \
    --cc=bp@alien8.de \
    --cc=dave.hansen@linux.intel.com \
    --cc=hpa@zytor.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@kernel.org \
    --cc=tglx@kernel.org \
    --cc=x86@kernel.org \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.