diff mbox series

entry: Fix boot for !CONFIG_GENERIC_ENTRY

Message ID 87a6v8qd9p.fsf_-_@collabora.com (mailing list archive)
State New, archived
Headers show
Series entry: Fix boot for !CONFIG_GENERIC_ENTRY | expand

Commit Message

Gabriel Krisman Bertazi Nov. 23, 2020, 3:54 p.m. UTC
Gabriel Krisman Bertazi <krisman@collabora.com> writes:

> Jann Horn <jannh@google.com> writes:
>> As part of fixing this, it might be a good idea to put "enum
>> syscall_work_bit" behind a "#ifdef CONFIG_GENERIC_ENTRY" to avoid
>> future accidents like this?
>
> Hi Jan, Arnd,
>
> That is correct.  This is a copy pasta mistake.  My apologies.  I didn't
> have a !GENERIC_ENTRY device to test, but just the ifdef would have
> caught it.

I have patched it as suggested.  Tested on qemu for arm32 and on bare
metal for x86-64.

Once again, my apologies for the mistake.

-- >8 --
Subject: [PATCH] entry: Fix boot for !CONFIG_GENERIC_ENTRY

A copy-pasta mistake tries to set SYSCALL_WORK flags instead of TIF
flags for !CONFIG_GENERIC_ENTRY.  Also, add safeguards to catch this at
compilation time.

Reported-by: Naresh Kamboju <naresh.kamboju@linaro.org>
Suggested-by: Jann Horn <jannh@google.com>
Signed-off-by: Gabriel Krisman Bertazi <krisman@collabora.com>
---
 include/linux/thread_info.h | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

Comments

Kees Cook Nov. 24, 2020, 9:45 p.m. UTC | #1
On Mon, Nov 23, 2020 at 10:54:58AM -0500, Gabriel Krisman Bertazi wrote:
> Gabriel Krisman Bertazi <krisman@collabora.com> writes:
> 
> > Jann Horn <jannh@google.com> writes:
> >> As part of fixing this, it might be a good idea to put "enum
> >> syscall_work_bit" behind a "#ifdef CONFIG_GENERIC_ENTRY" to avoid
> >> future accidents like this?
> >
> > Hi Jan, Arnd,
> >
> > That is correct.  This is a copy pasta mistake.  My apologies.  I didn't
> > have a !GENERIC_ENTRY device to test, but just the ifdef would have
> > caught it.
> 
> I have patched it as suggested.  Tested on qemu for arm32 and on bare
> metal for x86-64.
> 
> Once again, my apologies for the mistake.
> 
> -- >8 --
> Subject: [PATCH] entry: Fix boot for !CONFIG_GENERIC_ENTRY
> 
> A copy-pasta mistake tries to set SYSCALL_WORK flags instead of TIF
> flags for !CONFIG_GENERIC_ENTRY.  Also, add safeguards to catch this at
> compilation time.
> 
> Reported-by: Naresh Kamboju <naresh.kamboju@linaro.org>
> Suggested-by: Jann Horn <jannh@google.com>
> Signed-off-by: Gabriel Krisman Bertazi <krisman@collabora.com>

Thanks for getting this fixed!

3136b93c3fb2 ("entry: Expose helpers to migrate TIF to SYSCALL_WORK flags")
Reviewed-by: Kees Cook <keescook@chromium.org>
Naresh Kamboju Nov. 25, 2020, 3:36 a.m. UTC | #2
On Wed, 25 Nov 2020 at 03:15, Kees Cook <keescook@chromium.org> wrote:
>
> On Mon, Nov 23, 2020 at 10:54:58AM -0500, Gabriel Krisman Bertazi wrote:
> > Gabriel Krisman Bertazi <krisman@collabora.com> writes:
> >
> > > Jann Horn <jannh@google.com> writes:
> > >> As part of fixing this, it might be a good idea to put "enum
> > >> syscall_work_bit" behind a "#ifdef CONFIG_GENERIC_ENTRY" to avoid
> > >> future accidents like this?
> > >
> > > Hi Jan, Arnd,
> > >
> > > That is correct.  This is a copy pasta mistake.  My apologies.  I didn't
> > > have a !GENERIC_ENTRY device to test, but just the ifdef would have
> > > caught it.
> >
> > I have patched it as suggested.  Tested on qemu for arm32 and on bare
> > metal for x86-64.
> >
> > Once again, my apologies for the mistake.
> >
> > -- >8 --
> > Subject: [PATCH] entry: Fix boot for !CONFIG_GENERIC_ENTRY
> >
> > A copy-pasta mistake tries to set SYSCALL_WORK flags instead of TIF
> > flags for !CONFIG_GENERIC_ENTRY.  Also, add safeguards to catch this at
> > compilation time.

This patch tested on arm64, arm, x86_64 and i386
and the reported issue got fixed.

> >
> > Reported-by: Naresh Kamboju <naresh.kamboju@linaro.org>
> > Suggested-by: Jann Horn <jannh@google.com>
> > Signed-off-by: Gabriel Krisman Bertazi <krisman@collabora.com>
>
> Thanks for getting this fixed!
>
> 3136b93c3fb2 ("entry: Expose helpers to migrate TIF to SYSCALL_WORK flags")
> Reviewed-by: Kees Cook <keescook@chromium.org>

Tested-by: Naresh Kamboju <naresh.kamboju@linaro.org>

- Naresh
diff mbox series

Patch

diff --git a/include/linux/thread_info.h b/include/linux/thread_info.h
index 6a597fd5d351..45ad3176e2fa 100644
--- a/include/linux/thread_info.h
+++ b/include/linux/thread_info.h
@@ -35,6 +35,7 @@  enum {
 	GOOD_STACK,
 };
 
+#ifdef CONFIG_GENERIC_ENTRY
 enum syscall_work_bit {
 	SYSCALL_WORK_BIT_SECCOMP,
 	SYSCALL_WORK_BIT_SYSCALL_TRACEPOINT,
@@ -48,6 +49,7 @@  enum syscall_work_bit {
 #define SYSCALL_WORK_SYSCALL_TRACE	BIT(SYSCALL_WORK_BIT_SYSCALL_TRACE)
 #define SYSCALL_WORK_SYSCALL_EMU	BIT(SYSCALL_WORK_BIT_SYSCALL_EMU)
 #define SYSCALL_WORK_SYSCALL_AUDIT	BIT(SYSCALL_WORK_BIT_SYSCALL_AUDIT)
+#endif
 
 #include <asm/thread_info.h>
 
@@ -127,11 +129,11 @@  static inline int test_ti_thread_flag(struct thread_info *ti, int flag)
 	clear_bit(SYSCALL_WORK_BIT_##fl, &task_thread_info(t)->syscall_work)
 #else
 #define set_syscall_work(fl) \
-	set_ti_thread_flag(current_thread_info(), SYSCALL_WORK_##fl)
+	set_ti_thread_flag(current_thread_info(), TIF_##fl)
 #define test_syscall_work(fl) \
-	test_ti_thread_flag(current_thread_info(), SYSCALL_WORK_##fl)
+	test_ti_thread_flag(current_thread_info(), TIF_##fl)
 #define clear_syscall_work(fl) \
-	clear_ti_thread_flag(current_thread_info(), SYSCALL_WORK_##fl)
+	clear_ti_thread_flag(current_thread_info(), TIF_##fl)
 
 #define set_task_syscall_work(t, fl) \
 	set_ti_thread_flag(task_thread_info(t), TIF_##fl)