arm: cacheflush: get rid of restarting block

-----Original Message-----
From: Vladimir Murzin [mailto:vladimir.murzin at arm.com]
Sent: Wednesday, November 26, 2014 12:19 AM
To: linux-arm-kernel at lists.infradead.org; linux at arm.linux.org.uk
Cc: chanho.min at lge.com; will.deacon at arm.com; stable at vger.kernel.org
Subject: [PATCH] arm: cacheflush: get rid of restarting block

We cannot restart cacheflush safely if a process provides user-defined signal handler and signal is pending. In this case -EINTR is returned and it is expected that process re-invokes syscall. However, there are a few problems with that:
* looks like nobody bothers checking return value from cacheflush
* but if it did, we don't provide the restart address for that, so the process has to use the same range again
* ...and again, what might lead to looping forever

So, remove cacheflush restarting code and terminate cache flushing as early as fatal signal is pending.

Cc stable at vger.kernel.org # 3.12+

Reported-by Chanho Min <[email protected]>

Signed-off-by Vladimir Murzin <[email protected]>

Acked-by Will Deacon <[email protected]>

---
Hi,

There is my proposal for the problem reported in [1]. Since that thread has not been progressing much, I'm addressing the problem in separate patch.

Comments?

[1] https://lkml.org/lkml/2014/11/13/29

 arch/arm/include/asm/thread_info.h |   11 -----------
 arch/arm/kernel/traps.c            |   31 ++-----------------------------
 2 files changed, 2 insertions(+), 40 deletions(-)

arch/arm/include/asm/thread_info.h +0 -11

index fc44d37..ce73ab6 100644
@@ -44,16 +44,6 @@ struct cpu_context_save {
 	__u32	extra[2];		/* Xscale 'acc' register, etc */
 };
 
-struct arm_restart_block {
-	union {
-		/* For user cache flushing */
-		struct {
-			unsigned long start;
-			unsigned long end;
-		} cache;
-	};
-};
-
 /*
  * low level task data that entry.S needs immediate access to.
  * __switch_to() assumes cpu_context follows immediately after cpu_domain.

@@ -79,7 +69,6 @@ struct thread_info {
 	unsigned long		thumbee_state;	/* ThumbEE Handler Base register */
 #endif
 	struct restart_block	restart_block;
-	struct arm_restart_block	arm_restart_block;
 };
 
 #define INIT_THREAD_INFO(tsk)						\

arch/arm/kernel/traps.c +2 -30

index 0c8b108..8dd03ee 100644
@@ -533,8 +533,6 @@ static int bad_syscall(int n, struct pt_regs *regs)
 	return regs->ARM_r0;
 }
 
-static long do_cache_op_restart(struct restart_block *);
-
 static inline int
 __do_cache_op(unsigned long start, unsigned long end)
 {

@@ -543,24 +541,8 @@ __do_cache_op(unsigned long start, unsigned long end)
 	do {
 		unsigned long chunk = min(PAGE_SIZE, end - start);
 
-		if (signal_pending(current)) {
-			struct thread_info *ti = current_thread_info();
-
-			ti->restart_block = (struct restart_block) {
-				.fn	= do_cache_op_restart,
-			};
-
-			ti->arm_restart_block = (struct arm_restart_block) {
-				{
-					.cache = {
-						.start	= start,
-						.end	= end,
-					},
-				},
-			};
-
-			return -ERESTART_RESTARTBLOCK;
-		}
+		if (fatal_signal_pending(current))
+			return 0;
 
 		ret = flush_cache_user_range(start, start + chunk);
 		if (ret)

@@ -573,15 +555,6 @@ __do_cache_op(unsigned long start, unsigned long end)
 	return 0;
 }
 
-static long do_cache_op_restart(struct restart_block *unused)
-{
-	struct arm_restart_block *restart_block;
-
-	restart_block = &current_thread_info()->arm_restart_block;
-	return __do_cache_op(restart_block->cache.start,
-			     restart_block->cache.end);
-}
-
 static inline int
 do_cache_op(unsigned long start, unsigned long end, int flags)
 {
--

1.7.9.5

Quoted 26 lines

-----Original Message-----
From: Vladimir Murzin [mailto:vladimir.murzin at arm.com]
Sent: Wednesday, November 26, 2014 12:19 AM
To: linux-arm-kernel at lists.infradead.org; linux at arm.linux.org.uk
Cc: chanho.min at lge.com; will.deacon at arm.com; stable at vger.kernel.org
Subject: [PATCH] arm: cacheflush: get rid of restarting block

We cannot restart cacheflush safely if a process provides user-defined signal handler and signal is pending. In this case -EINTR is returned and it is expected that process re-invokes syscall. However, there are a few problems with that:
* looks like nobody bothers checking return value from cacheflush
* but if it did, we don't provide the restart address for that, so the process has to use the same range again
* ...and again, what might lead to looping forever

So, remove cacheflush restarting code and terminate cache flushing as early as fatal signal is pending.

Cc stable at vger.kernel.org # 3.12+

Reported-by Chanho Min <[email protected]>

Signed-off-by Vladimir Murzin <[email protected]>

Acked-by Will Deacon <[email protected]>

---
Hi,

Quoted 5 lines

There is my proposal for the problem reported in [1]. Since that thread has not been progressing much, I'm addressing the problem in separate patch.

Comments?

Thanks, We hope this to be applied.

Quoted 30 lines · Chanho Min

Quoted 30 lines

-----Original Message-----
From: Vladimir Murzin [mailto:vladimir.murzin at arm.com]
Sent: Wednesday, November 26, 2014 12:19 AM
To: linux-arm-kernel at lists.infradead.org; linux at arm.linux.org.uk
Cc: chanho.min at lge.com; will.deacon at arm.com; stable at vger.kernel.org
Subject: [PATCH] arm: cacheflush: get rid of restarting block

We cannot restart cacheflush safely if a process provides user-defined signal handler and signal is pending. In this case -EINTR is returned and it is expected that process re-invokes syscall. However, there are a few problems with that:
* looks like nobody bothers checking return value from cacheflush
* but if it did, we don't provide the restart address for that, so the process has to use the same range again
* ...and again, what might lead to looping forever

So, remove cacheflush restarting code and terminate cache flushing as early as fatal signal is pending.

Cc stable at vger.kernel.org # 3.12+

Reported-by Chanho Min <[email protected]>

Signed-off-by Vladimir Murzin <[email protected]>

Acked-by Will Deacon <[email protected]>

---
Hi,

There is my proposal for the problem reported in [1]. Since that thread has not been progressing much, I'm addressing the problem in separate patch.

Quoted 2 lines

Quoted 1 line

Comments?

Thanks, We hope this to be applied.

Ok, good. In which case, can you put it into the patch system please, Vladimir?

Will

Quoted 32 lines · Will Deacon

Quoted 31 lines · Chanho Min

Quoted 31 lines

-----Original Message-----
From: Vladimir Murzin [mailto:vladimir.murzin at arm.com]
Sent: Wednesday, November 26, 2014 12:19 AM
To: linux-arm-kernel at lists.infradead.org; linux at arm.linux.org.uk
Cc: chanho.min at lge.com; will.deacon at arm.com; stable at vger.kernel.org
Subject: [PATCH] arm: cacheflush: get rid of restarting block

We cannot restart cacheflush safely if a process provides user-defined signal handler and signal is pending. In this case -EINTR is returned and it is expected that process re-invokes syscall. However, there are a few problems with that:
* looks like nobody bothers checking return value from cacheflush
* but if it did, we don't provide the restart address for that, so the process has to use the same range again
* ...and again, what might lead to looping forever

So, remove cacheflush restarting code and terminate cache flushing as early as fatal signal is pending.

Cc stable at vger.kernel.org # 3.12+

Reported-by Chanho Min <[email protected]>

Signed-off-by Vladimir Murzin <[email protected]>

Acked-by Will Deacon <[email protected]>

---
Hi,

There is my proposal for the problem reported in [1]. Since that thread has not been progressing much, I'm addressing the problem in separate patch.

Comments?

Quoted 4 lines

Quoted 1 line

Thanks, We hope this to be applied.

Ok, good. In which case, can you put it into the patch system please, Vladimir?

Already in patch system as 8226/1.

Vladimir

Quoted 5 lines

Will

-- IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.

ARM Limited, Registered office 110 Fulbourn Road, Cambridge CB1 9NJ, Registered in England & Wales, Company No: 2557590 ARM Holdings plc, Registered office 110 Fulbourn Road, Cambridge CB1 9NJ, Registered in England & Wales, Company No: 2548782

--
To unsubscribe from this list: send the line "unsubscribe stable" in the body of a message to majordomo at vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html