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 RESEND -tip v2 2/7] x86/insn-eval: Fix signedness bug in segment selector handling
Date: Tue, 28 Apr 2026 18:03:36 +0200 [thread overview]
Message-ID: <20260428160443.3593331-2-ubizjak@gmail.com> (raw)
In-Reply-To: <20260428160443.3593331-1-ubizjak@gmail.com>
get_segment_selector() returns a short, while callers use the pattern:
short sel = get_segment_selector(...);
if (sel < 0)
return -ERR;
Segment selectors are 16-bit values, but storing them in a signed
16-bit type means values with the MSB set (>= 0x8000) become negative.
This causes valid selectors to be misinterpreted as errors by the
'sel < 0' check.
Change get_segment_selector() to return int and update all call sites
to use 'int sel' to avoid unintended sign extension and keep error
handling via negative return values correct.
Additionally, remove the explicit & 0xffff masking when reading
segment registers. The compiler already zero-extends
unsigned 16-bit values when loading them into a
wider type, so the masking is redundant.
With this change, valid segment selectors are no longer
confused with error returns.
Signed-off-by: Uros Bizjak <ubizjak@gmail.com>
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>
---
arch/x86/lib/insn-eval.c | 28 ++++++++++++++--------------
1 file changed, 14 insertions(+), 14 deletions(-)
diff --git a/arch/x86/lib/insn-eval.c b/arch/x86/lib/insn-eval.c
index e03eeec55cfe..b8847ce0b282 100644
--- a/arch/x86/lib/insn-eval.c
+++ b/arch/x86/lib/insn-eval.c
@@ -340,18 +340,18 @@ static int resolve_seg_reg(struct insn *insn, struct pt_regs *regs, int regoff)
*
* -EINVAL on error.
*/
-static short get_segment_selector(struct pt_regs *regs, int seg_reg_idx)
+static int get_segment_selector(struct pt_regs *regs, int seg_reg_idx)
{
- unsigned short sel;
+ unsigned int sel;
#ifdef CONFIG_X86_64
switch (seg_reg_idx) {
case INAT_SEG_REG_IGNORE:
return 0;
case INAT_SEG_REG_CS:
- return (unsigned short)(regs->cs & 0xffff);
+ return regs->cs;
case INAT_SEG_REG_SS:
- return (unsigned short)(regs->ss & 0xffff);
+ return regs->ss;
case INAT_SEG_REG_DS:
savesegment(ds, sel);
return sel;
@@ -373,9 +373,9 @@ static short get_segment_selector(struct pt_regs *regs, int seg_reg_idx)
if (v8086_mode(regs)) {
switch (seg_reg_idx) {
case INAT_SEG_REG_CS:
- return (unsigned short)(regs->cs & 0xffff);
+ return regs->cs;
case INAT_SEG_REG_SS:
- return (unsigned short)(regs->ss & 0xffff);
+ return regs->ss;
case INAT_SEG_REG_DS:
return vm86regs->ds;
case INAT_SEG_REG_ES:
@@ -392,15 +392,15 @@ static short get_segment_selector(struct pt_regs *regs, int seg_reg_idx)
switch (seg_reg_idx) {
case INAT_SEG_REG_CS:
- return (unsigned short)(regs->cs & 0xffff);
+ return regs->cs;
case INAT_SEG_REG_SS:
- return (unsigned short)(regs->ss & 0xffff);
+ return regs->ss;
case INAT_SEG_REG_DS:
- return (unsigned short)(regs->ds & 0xffff);
+ return regs->ds;
case INAT_SEG_REG_ES:
- return (unsigned short)(regs->es & 0xffff);
+ return regs->es;
case INAT_SEG_REG_FS:
- return (unsigned short)(regs->fs & 0xffff);
+ return regs->fs;
case INAT_SEG_REG_GS:
savesegment(gs, sel);
return sel;
@@ -688,7 +688,7 @@ static bool get_desc(struct desc_struct *out, unsigned short sel)
unsigned long insn_get_seg_base(struct pt_regs *regs, int seg_reg_idx)
{
struct desc_struct desc;
- short sel;
+ int sel;
sel = get_segment_selector(regs, seg_reg_idx);
if (sel < 0)
@@ -756,7 +756,7 @@ static unsigned long get_seg_limit(struct pt_regs *regs, int seg_reg_idx)
{
struct desc_struct desc;
unsigned long limit;
- short sel;
+ int sel;
sel = get_segment_selector(regs, seg_reg_idx);
if (sel < 0)
@@ -803,7 +803,7 @@ static unsigned long get_seg_limit(struct pt_regs *regs, int seg_reg_idx)
int insn_get_code_seg_params(struct pt_regs *regs)
{
struct desc_struct desc;
- short sel;
+ int sel;
if (v8086_mode(regs))
/* Address and operand size are both 16-bit. */
--
2.53.0
next prev parent reply other threads:[~2026-04-28 16:05 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-04-28 16:03 [PATCH RESEND -tip v2 1/7] x86_32/segment: Always return correctly zero-extended values from savesegment_*() Uros Bizjak
2026-04-28 16:03 ` Uros Bizjak [this message]
2026-04-28 16:03 ` [PATCH RESEND -tip v2 3/7] x86/ptrace: Use savesegment() in get_segment_reg() instead of inline asm Uros Bizjak
2026-04-28 16:03 ` [PATCH RESEND -tip v2 4/7] x86/ptrace: Use unsigned short for segment values in get_segment_reg() Uros Bizjak
2026-04-28 16:03 ` [PATCH RESEND -tip v2 5/7] x86/kexec: store segment registers directly to memory in crash_setup_regs() Uros Bizjak
2026-04-28 16:03 ` [PATCH RESEND -tip v2 6/7] x86/segment: Introduce savesegment_mem16() helper to write segment selectors to memory Uros Bizjak
2026-04-28 16:03 ` [PATCH RESEND -tip v2 7/7] x86/process: Use savesegment_mem16() when saving segment selectors 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=20260428160443.3593331-2-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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox