Skip to contentSkip to conversation

Fix Warnining in arch/x86/kernel/signal.c

PATCH Subrata Modak 2 participants 6 messages May 14, 2009 – May 18, 2009 x86

lore
Share & tools
lore
#0 lore raw

Hello Hiroshi-san,

Quoted 33 lines · On Thu, 2009-05-14 at 09:24 +0900, Hiroshi Shimamoto wrote: H. Peter Anvin…
Quoted 24 lines
Quoted 23 lines · Ingo Molnar
Quoted 23 lines
Quoted 11 lines

Review

Quoted 11 lines

if (!access_ok(VERIFY_READ, frame, sizeof(*frame)))
goto badframe;

(quoted diff) +4 -3
-       if (__get_user(set.sig[0], &frame->sc.oldmask) || (_NSIG_WORDS > 1
-               && __copy_from_user(&set.sig[1], &frame->extramask,
-                                   sizeof(frame->extramask))))
+
+        if ( (__copy_from_user(&set.sig[1], &frame->extramask,
+                sizeof(frame->extramask)) && _NSIG_WORDS > 1) || 
+                __get_user(set.sig[0], &frame->sc.oldmask))
                goto badframe;
Quoted 2 lines

I'm not sure why this eliminates that warning.
set.sig[0] may not be initialized too, if __copy_from_user() failed.

True, but only when either or both of __copy_from_user() and (_NSIG_WORDS > 1) fails. But in all instances set.sig[1] gets initialized.

Quoted 6 lines

I don't have enough time to look at this right now, sorry.

Another question, __copy_from_user() will be called even if _NSIG_WORDS is less than 2, perhaps it never occurs.
I think, to check _NSIG_WORDS > 1 before calling __copy_from_user() is better.

Quoted 8 lines
Quoted 5 lines
Quoted 2 lines

Fine. Let Ingo/Thomas/Peter decide whether they would like this fix or drop it.

If you get the Acked-by from Hiroshi-san it looks good to me. He modified this code last.

This seriously looks wrong to me. If _NSIG_WORDS == 1, then calling __copy_from_user here is a serious error.

Quoted 8 lines

Right. If _NSIG_WORDS is 1, sigset_t set has only sig[0], writing to set.sig[1] means stack corruption.

Subrata, could you try like this?
if ((_NSIG_WORDS > 1 && __copy_from_user(&set.sig[1], ...) ||

      __get_user(set.sig[0], ...))

How about now ? Thanks for pointing that out. My mistake ;-)

To: Hiroshi Shimamoto <[email protected]>,
Cc: H. Peter Anvin <[email protected]>,
Cc: Ingo Molnar <[email protected]>,
Cc: [email protected],
Cc: Sachin P Sant <[email protected]>,
Cc: Andi Kleen <[email protected]>,
Cc: Andi Kleen <[email protected]>,
Cc: Linux Kernel <[email protected]>,
Cc: Balbir Singh <[email protected]>,
Cc: Thomas Gleixner <[email protected]>,
Cc: Ingo Molnar <[email protected]>
Subject: Re:[PATCH] Fix Warnining in arch/x86/kernel/signal.c
---

arch/x86/kernel/signal.c +3 -3
@@ -576,9 +576,9 @@ unsigned long sys_sigreturn(struct pt_re
 
 	if (!access_ok(VERIFY_READ, frame, sizeof(*frame)))
 		goto badframe;
-	if (__get_user(set.sig[0], &frame->sc.oldmask) || (_NSIG_WORDS > 1
-		&& __copy_from_user(&set.sig[1], &frame->extramask,
-				    sizeof(frame->extramask))))
+	if ((_NSIG_WORDS > 1 && __copy_from_user(&set.sig[1],
+		&frame->extramask, sizeof(frame->extramask))) ||
+		__get_user(set.sig[0], &frame->sc.oldmask))
 		goto badframe;
 
 	sigdelsetmask(&set, ~_BLOCKABLE);

---
Regards--
Subrata

Quoted 7 lines

I wonder whether gcc really complains about the case of __get_user(set.sig[0], ...) failure.
Why, the case which sig[0] initialized and sig[1] uninitialized is NG and the case which sig[0] uninitialized and sig[1] initialized is OK.

Thanks,
Hiroshi

#1 lore raw
Quoted 44 lines · Subrata Modak

Hello Hiroshi-san,

Quoted 29 lines · On Thu, 2009-05-14 at 09:24 +0900, Hiroshi Shimamoto wrote: H. Peter Anvin…
Quoted 29 lines
Quoted 28 lines · Ingo Molnar
Quoted 17 lines
Quoted 11 lines

arch/x86/kernel/signal.c unsigned long sys_sigreturn(struct pt_re · line 582

Quoted 11 lines

if (!access_ok(VERIFY_READ, frame, sizeof(*frame)))
goto badframe;

arch/x86/kernel/signal.c +4 -3
-       if (__get_user(set.sig[0], &frame->sc.oldmask) || (_NSIG_WORDS > 1
-               && __copy_from_user(&set.sig[1], &frame->extramask,
-                                   sizeof(frame->extramask))))
+
+        if ( (__copy_from_user(&set.sig[1], &frame->extramask,
+                sizeof(frame->extramask)) && _NSIG_WORDS > 1) || 
+                __get_user(set.sig[0], &frame->sc.oldmask))
                goto badframe;
Quoted 2 lines

I'm not sure why this eliminates that warning.
set.sig[0] may not be initialized too, if __copy_from_user() failed.

True, but only when either or both of __copy_from_user() and (_NSIG_WORDS > 1) fails. But in all instances set.sig[1] gets initialized.

Quoted 8 lines
Quoted 6 lines

I don't have enough time to look at this right now, sorry.

Another question, __copy_from_user() will be called even if _NSIG_WORDS is less than 2, perhaps it never occurs.
I think, to check _NSIG_WORDS > 1 before calling __copy_from_user() is better.

Fine. Let Ingo/Thomas/Peter decide whether they would like this fix or drop it.

If you get the Acked-by from Hiroshi-san it looks good to me. He modified this code last.

Quoted 10 lines
Quoted 2 lines

This seriously looks wrong to me. If _NSIG_WORDS == 1, then calling __copy_from_user here is a serious error.

Right. If _NSIG_WORDS is 1, sigset_t set has only sig[0], writing to set.sig[1] means stack corruption.

Subrata, could you try like this?
if ((_NSIG_WORDS > 1 && __copy_from_user(&set.sig[1], ...) ||

      __get_user(set.sig[0], ...))
Quoted 4 lines

How about now ? Thanks for pointing that out. My mistake ;-)

To: Hiroshi Shimamoto <[email protected]>,

looks good to me.
BTW who writes the description?

Quoted 41 lines

Cc: H. Peter Anvin <[email protected]>,
Cc: Ingo Molnar <[email protected]>,
Cc: [email protected],
Cc: Sachin P Sant <[email protected]>,
Cc: Andi Kleen <[email protected]>,
Cc: Andi Kleen <[email protected]>,
Cc: Linux Kernel <[email protected]>,
Cc: Balbir Singh <[email protected]>,
Cc: Thomas Gleixner <[email protected]>,
Cc: Ingo Molnar <[email protected]>
Subject: Re:[PATCH] Fix Warnining in arch/x86/kernel/signal.c
---

arch/x86/kernel/signal.c +3 -4
@@ -576,9 +576,9 @@ unsigned long sys_sigreturn(struct pt_re
 
 	if (!access_ok(VERIFY_READ, frame, sizeof(*frame)))
 		goto badframe;
-	if (__get_user(set.sig[0], &frame->sc.oldmask) || (_NSIG_WORDS > 1
-		&& __copy_from_user(&set.sig[1], &frame->extramask,
-				    sizeof(frame->extramask))))
+	if ((_NSIG_WORDS > 1 && __copy_from_user(&set.sig[1],
+		&frame->extramask, sizeof(frame->extramask))) ||
+		__get_user(set.sig[0], &frame->sc.oldmask))
 		goto badframe;
 
 	sigdelsetmask(&set, ~_BLOCKABLE);
 
---

Regards--
Subrata

Quoted 7 lines

I wonder whether gcc really complains about the case of __get_user(set.sig[0], ...) failure.
Why, the case which sig[0] initialized and sig[1] uninitialized is NG and the case which sig[0] uninitialized and sig[1] initialized is OK.

Thanks,
Hiroshi

#2 lore raw
In reply to Subrata Modak
Quoted 38 lines · Subrata Modak

Hello Hiroshi-san,

Quoted 34 lines · On Thu, 2009-05-14 at 09:24 +0900, Hiroshi Shimamoto wrote: H. Peter Anvin…
Quoted 24 lines
Quoted 23 lines · Ingo Molnar
Quoted 23 lines
Quoted 11 lines

arch/x86/kernel/signal.c unsigned long sys_sigreturn(struct pt_re · line 582

Quoted 11 lines

if (!access_ok(VERIFY_READ, frame, sizeof(*frame)))
goto badframe;

arch/x86/kernel/signal.c +4 -3
-       if (__get_user(set.sig[0], &frame->sc.oldmask) || (_NSIG_WORDS > 1
-               && __copy_from_user(&set.sig[1], &frame->extramask,
-                                   sizeof(frame->extramask))))
+
+        if ( (__copy_from_user(&set.sig[1], &frame->extramask,
+                sizeof(frame->extramask)) && _NSIG_WORDS > 1) || 
+                __get_user(set.sig[0], &frame->sc.oldmask))
                goto badframe;
Quoted 2 lines

I'm not sure why this eliminates that warning.
set.sig[0] may not be initialized too, if __copy_from_user() failed.

True, but only when either or both of __copy_from_user() and (_NSIG_WORDS > 1) fails. But in all instances set.sig[1] gets initialized.

Quoted 6 lines

I don't have enough time to look at this right now, sorry.

Another question, __copy_from_user() will be called even if _NSIG_WORDS is less than 2, perhaps it never occurs.
I think, to check _NSIG_WORDS > 1 before calling __copy_from_user() is better.

Quoted 7 lines
Quoted 5 lines
Quoted 2 lines

Fine. Let Ingo/Thomas/Peter decide whether they would like this fix or drop it.

If you get the Acked-by from Hiroshi-san it looks good to me. He modified this code last.

This seriously looks wrong to me. If _NSIG_WORDS == 1, then calling __copy_from_user here is a serious error.

Right. If _NSIG_WORDS is 1, sigset_t set has only sig[0], writing to set.sig[1] means stack corruption.

Quoted 7 lines
Quoted 5 lines

Subrata, could you try like this?
if ((_NSIG_WORDS > 1 && __copy_from_user(&set.sig[1], ...) ||

      __get_user(set.sig[0], ...))

How about now ? Thanks for pointing that out. My mistake ;-)

Hi Subrata, I have a question.
Have you tried to compile on x86_64 whether the compiler claims the similar code in sys32_sigreturn() in arch/x86/ia32/ia32_signal.c?

Thanks,
Hiroshi

#3 lore raw
Quoted 45 lines · Hiroshi Shimamoto
Quoted 44 lines · Subrata Modak

Hello Hiroshi-san,

Quoted 39 lines · On Thu, 2009-05-14 at 09:24 +0900, Hiroshi Shimamoto wrote: H. Peter Anvin…
Quoted 24 lines
Quoted 23 lines · Ingo Molnar
Quoted 23 lines
Quoted 11 lines

arch/x86/kernel/signal.c unsigned long sys_sigreturn(struct pt_re · line 582

Quoted 11 lines

if (!access_ok(VERIFY_READ, frame, sizeof(*frame)))
goto badframe;

arch/x86/kernel/signal.c +4 -3
-       if (__get_user(set.sig[0], &frame->sc.oldmask) || (_NSIG_WORDS > 1
-               && __copy_from_user(&set.sig[1], &frame->extramask,
-                                   sizeof(frame->extramask))))
+
+        if ( (__copy_from_user(&set.sig[1], &frame->extramask,
+                sizeof(frame->extramask)) && _NSIG_WORDS > 1) || 
+                __get_user(set.sig[0], &frame->sc.oldmask))
                goto badframe;
Quoted 2 lines

I'm not sure why this eliminates that warning.
set.sig[0] may not be initialized too, if __copy_from_user() failed.

True, but only when either or both of __copy_from_user() and (_NSIG_WORDS > 1) fails. But in all instances set.sig[1] gets initialized.

Quoted 6 lines

I don't have enough time to look at this right now, sorry.

Another question, __copy_from_user() will be called even if _NSIG_WORDS is less than 2, perhaps it never occurs.
I think, to check _NSIG_WORDS > 1 before calling __copy_from_user() is better.

Quoted 7 lines
Quoted 5 lines
Quoted 2 lines

Fine. Let Ingo/Thomas/Peter decide whether they would like this fix or drop it.

If you get the Acked-by from Hiroshi-san it looks good to me. He modified this code last.

This seriously looks wrong to me. If _NSIG_WORDS == 1, then calling __copy_from_user here is a serious error.

Right. If _NSIG_WORDS is 1, sigset_t set has only sig[0], writing to set.sig[1] means stack corruption.

Subrata, could you try like this?
if ((_NSIG_WORDS > 1 && __copy_from_user(&set.sig[1], ...) ||

      __get_user(set.sig[0], ...))
Quoted 5 lines
Quoted 1 line

How about now ? Thanks for pointing that out. My mistake ;-)

Hi Subrata, I have a question.
Have you tried to compile on x86_64 whether the compiler claims the similar code in sys32_sigreturn() in arch/x86/ia32/ia32_signal.c?

Oops. No, the compiler does not complain here. It simply compiles fine.

So, do you want to take a different view for the patch against arch/x86/kernel/signal.c, or, i would resend it with the following things fixed:

Quoted 5 lines

looks good to me.
BTW who writes the description?

Regards--
Subrata

Quoted 3 lines

Thanks,
Hiroshi

#4 lore raw
Quoted 50 lines · Subrata Modak
Quoted 44 lines · Hiroshi Shimamoto
Quoted 43 lines · Subrata Modak

Hello Hiroshi-san,

Quoted 39 lines · On Thu, 2009-05-14 at 09:24 +0900, Hiroshi Shimamoto wrote: H. Peter Anvin…
Quoted 24 lines
Quoted 23 lines · Ingo Molnar
Quoted 23 lines
Quoted 11 lines

arch/x86/kernel/signal.c unsigned long sys_sigreturn(struct pt_re · line 582

Quoted 11 lines

if (!access_ok(VERIFY_READ, frame, sizeof(*frame)))
goto badframe;

arch/x86/kernel/signal.c +4 -3
-       if (__get_user(set.sig[0], &frame->sc.oldmask) || (_NSIG_WORDS > 1
-               && __copy_from_user(&set.sig[1], &frame->extramask,
-                                   sizeof(frame->extramask))))
+
+        if ( (__copy_from_user(&set.sig[1], &frame->extramask,
+                sizeof(frame->extramask)) && _NSIG_WORDS > 1) || 
+                __get_user(set.sig[0], &frame->sc.oldmask))
                goto badframe;
Quoted 2 lines

I'm not sure why this eliminates that warning.
set.sig[0] may not be initialized too, if __copy_from_user() failed.

True, but only when either or both of __copy_from_user() and (_NSIG_WORDS > 1) fails. But in all instances set.sig[1] gets initialized.

Quoted 6 lines

I don't have enough time to look at this right now, sorry.

Another question, __copy_from_user() will be called even if _NSIG_WORDS is less than 2, perhaps it never occurs.
I think, to check _NSIG_WORDS > 1 before calling __copy_from_user() is better.

Quoted 7 lines
Quoted 5 lines
Quoted 2 lines

Fine. Let Ingo/Thomas/Peter decide whether they would like this fix or drop it.

If you get the Acked-by from Hiroshi-san it looks good to me. He modified this code last.

This seriously looks wrong to me. If _NSIG_WORDS == 1, then calling __copy_from_user here is a serious error.

Right. If _NSIG_WORDS is 1, sigset_t set has only sig[0], writing to set.sig[1] means stack corruption.

Subrata, could you try like this?
if ((_NSIG_WORDS > 1 && __copy_from_user(&set.sig[1], ...) ||

      __get_user(set.sig[0], ...))
Quoted 4 lines
Quoted 1 line

How about now ? Thanks for pointing that out. My mistake ;-)

Hi Subrata, I have a question.
Have you tried to compile on x86_64 whether the compiler claims the similar code in sys32_sigreturn() in arch/x86/ia32/ia32_signal.c?

Quoted 5 lines

Oops. No, the compiler does not complain here. It simply compiles fine.

So, do you want to take a different view for the patch against arch/x86/kernel/signal.c, or, i would resend it with the following things fixed:

If you don't think this fix is urgent, could you please check whether that warning is false positive on 32bit or gcc for 64bit has an issue not to complain against ia32 part?

I think arch/x86/kernel/signal.c and arch/x86/ia32/ia32_signal.c should have the same code.

Thanks,
Hiroshi

#5 lore raw
Quoted 53 lines · Hiroshi Shimamoto
Quoted 52 lines · Subrata Modak
Quoted 44 lines · Hiroshi Shimamoto
Quoted 43 lines · Subrata Modak

Hello Hiroshi-san,

Quoted 39 lines · On Thu, 2009-05-14 at 09:24 +0900, Hiroshi Shimamoto wrote: H. Peter Anvin…
Quoted 24 lines
Quoted 23 lines · Ingo Molnar
Quoted 23 lines
Quoted 11 lines

arch/x86/kernel/signal.c unsigned long sys_sigreturn(struct pt_re · line 582

Quoted 11 lines

if (!access_ok(VERIFY_READ, frame, sizeof(*frame)))
goto badframe;

arch/x86/kernel/signal.c +4 -3
-       if (__get_user(set.sig[0], &frame->sc.oldmask) || (_NSIG_WORDS > 1
-               && __copy_from_user(&set.sig[1], &frame->extramask,
-                                   sizeof(frame->extramask))))
+
+        if ( (__copy_from_user(&set.sig[1], &frame->extramask,
+                sizeof(frame->extramask)) && _NSIG_WORDS > 1) || 
+                __get_user(set.sig[0], &frame->sc.oldmask))
                goto badframe;
Quoted 2 lines

I'm not sure why this eliminates that warning.
set.sig[0] may not be initialized too, if __copy_from_user() failed.

True, but only when either or both of __copy_from_user() and (_NSIG_WORDS > 1) fails. But in all instances set.sig[1] gets initialized.

Quoted 6 lines

I don't have enough time to look at this right now, sorry.

Another question, __copy_from_user() will be called even if _NSIG_WORDS is less than 2, perhaps it never occurs.
I think, to check _NSIG_WORDS > 1 before calling __copy_from_user() is better.

Quoted 7 lines
Quoted 5 lines
Quoted 2 lines

Fine. Let Ingo/Thomas/Peter decide whether they would like this fix or drop it.

If you get the Acked-by from Hiroshi-san it looks good to me. He modified this code last.

This seriously looks wrong to me. If _NSIG_WORDS == 1, then calling __copy_from_user here is a serious error.

Right. If _NSIG_WORDS is 1, sigset_t set has only sig[0], writing to set.sig[1] means stack corruption.

Subrata, could you try like this?
if ((_NSIG_WORDS > 1 && __copy_from_user(&set.sig[1], ...) ||

      __get_user(set.sig[0], ...))
Quoted 4 lines
Quoted 1 line

How about now ? Thanks for pointing that out. My mistake ;-)

Hi Subrata, I have a question.
Have you tried to compile on x86_64 whether the compiler claims the similar code in sys32_sigreturn() in arch/x86/ia32/ia32_signal.c?

Oops. No, the compiler does not complain here. It simply compiles fine.

Quoted 8 lines
Quoted 3 lines

So, do you want to take a different view for the patch against arch/x86/kernel/signal.c, or, i would resend it with the following things fixed:

If you don't think this fix is urgent, could you please check whether that warning is false positive on 32bit or gcc for 64bit has an issue not to complain against ia32 part?

I will try to find out.

Quoted 2 lines

I think arch/x86/kernel/signal.c and arch/x86/ia32/ia32_signal.c should have the same code.

Sure. Not a problem. Please drop the patch for the time being.

Regards--
Subrata

Quoted 3 lines

Thanks,
Hiroshi