ARC futex fixes
Subsystems (2)
cover ARC futex fixes
Hi Thomas/Peter,
Of late I've been debugging a seeming lost wakeup when running a specific EEMBC Multibench workload (4M-check -w4) on a quad core HS38 config on FPGA CONFIG_SMP, CONFIG_PREEMPT.
I've yet to nail that issue down, but in the process found some deficinecies in ARC futex backend. Can you please take a quick look and shout if there's something wrong there.
The futex backend for ARC, despite it's age, only recently started getting real testing due to recent switch to NPTL based userland (uClibc).
Thx,
-Vineet
Vineet Gupta (4):
ARC: add barriers to futex code
ARC: futex cosmetics
ARC: make futex_atomic_cmpxchg_inatomic() return bimodal
ARC: Enable HAVE_FUTEX_CMPXCHG
arch/arc/Kconfig | 1 + arch/arc/include/asm/futex.h | 44 +++++++++++++++++++++++++------------------- 2 files changed, 26 insertions(+), 19 deletions(-)
Signature · 1 line
1.9.1
PATCH 1/4 ARC: add barriers to futex code
The atomic ops on futex need to provide the full barrier just like regular atomics in kernel.
Also remove pagefault_enable/disable in futex_atomic_cmpxchg_inatomic() as core code already does that
--- arch/arc/include/asm/futex.h | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-)
arch/arc/include/asm/futex.h +8 -4
@@ -20,6 +20,7 @@#define __futex_atomic_op (insn, ret, oldval, uaddr, oparg)\ \ + smp_mb(); \ __asm__ __volatile__( \ "1: llock %1, [%2] \n" \ insn "\n" \
@@ -40,12 +41,14 @@ \ : "=&r" (ret), "=&r" (oldval) \ : "r" (uaddr), "r" (oparg), "ir" (-EFAULT) \ - : "cc", "memory") + : "cc", "memory"); \ +smp_mb (); \#else /* !CONFIG_ARC_HAS_LLSC */ #define __futex_atomic_op (insn, ret, oldval, uaddr, oparg)\ \ + smp_mb(); \ __asm__ __volatile__( \ "1: ld %1, [%2] \n" \ insn "\n" \
@@ -65,7 +68,8 @@ \ : "=&r" (ret), "=&r" (oldval) \ : "r" (uaddr), "r" (oparg), "ir" (-EFAULT) \ - : "cc", "memory") + : "cc", "memory"); \ +smp_mb (); \#endif
@@ -151,7 +155,7 @@ futex_atomic_cmpxchg_inatomic(u32 *uval, u32 __user *uaddr, u32 oldval,if (!access_ok (VERIFY_WRITE , uaddr,sizeof (int )))return -EFAULT ; -pagefault_disable (); +smp_mb (); __asm__ __volatile__(#ifdef CONFIG_ARC_HAS_LLSC
@@ -178,7 +182,7 @@ futex_atomic_cmpxchg_inatomic(u32 *uval, u32 __user *uaddr, u32 oldval, : "r"(oldval), "r"(newval), "r"(uaddr), "ir"(-EFAULT) : "cc", "memory"); -pagefault_enable (); +smp_mb (); *uval = val;return val;
Signature · 1 line
1.9.1
Re: [PATCH 1/4] ARC: add barriers to futex code
Quoted 28 lines
The atomic ops on futex need to provide the full barrier just like regular atomics in kernel.
Also remove pagefault_enable/disable in
futex_atomic_cmpxchg_inatomic()as core code already does thatCc David Hildenbrand <[email protected]>Cc Peter Zijlstra (Intel) <[email protected]>Cc Thomas Gleixner <[email protected]>Cc Michel Lespinasse <[email protected]>Signed-off-by Vineet Gupta <[email protected]>--- arch/arc/include/asm/futex.h | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-)arch/arc/include/asm/futex.h +1
-0@@ -20,6 +20,7 @@#define __futex_atomic_op (insn, ret, oldval, uaddr, oparg)\ \ + smp_mb(); \ __asm__ __volatile__( \ "1: llock %1, [%2] \n" \ insn "\n" \
arch/arc/include/asm/futex.h line 45
Quoted 6 lines
arch/arc/include/asm/futex.h +2
-1@@ -40,12 +41,14 @@ \ : "=&r" (ret), "=&r" (oldval) \ : "r" (uaddr), "r" (oparg), "ir" (-EFAULT) \ - : "cc", "memory") + : "cc", "memory"); \ +smp_mb (); \
I think you should drop the ;
Quoted 10 lines
arch/arc/include/asm/futex.h +1
-0#else /* !CONFIG_ARC_HAS_LLSC */ #define __futex_atomic_op (insn, ret, oldval, uaddr, oparg)\ \ + smp_mb(); \ __asm__ __volatile__( \ "1: ld %1, [%2] \n" \ insn "\n" \
arch/arc/include/asm/futex.h line 72
Quoted 6 lines
arch/arc/include/asm/futex.h +2
-1@@ -65,7 +68,8 @@ \ : "=&r" (ret), "=&r" (oldval) \ : "r" (uaddr), "r" (oparg), "ir" (-EFAULT) \ - : "cc", "memory") + : "cc", "memory"); \ +smp_mb (); \
dito
Quoted 13 lines
arch/arc/include/asm/futex.h +1
-1#endif @@ -151,7 +155,7 @@ futex_atomic_cmpxchg_inatomic(u32 *uval, u32 __user *uaddr, u32 oldval,if (!access_ok (VERIFY_WRITE , uaddr,sizeof (int )))return -EFAULT ; -pagefault_disable (); +smp_mb (); __asm__ __volatile__(#ifdef CONFIG_ARC_HAS_LLSC
arch/arc/include/asm/futex.h futex_atomic_cmpxchg_inatomic(u32 *uval, u32 __user *uaddr, u32 oldval, · line 188
Quoted 8 lines
arch/arc/include/asm/futex.h +1
-1@@ -178,7 +182,7 @@ futex_atomic_cmpxchg_inatomic(u32 *uval, u32 __user *uaddr, u32 oldval, : "r"(oldval), "r"(newval), "r"(uaddr), "ir"(-EFAULT) : "cc", "memory"); -pagefault_enable (); +smp_mb (); *uval = val;return val;
Looks like pagefault_() magic is only required for futex_atomic_op_inuser. So this should be fine (and arc seems to be the only arch left that has in in _inatomic).
Not sure if you want to change the comment:
/* Compare-xchg with pagefaults disabled.
* Notes:
* -Best-Effort: Exchg happens only if compare succeeds.
Maybe something like "Compare-xchg: pagefaults have to be disabled by the caller"
Looks sane to me.
David
Re: [PATCH 1/4] ARC: add barriers to futex code
Quoted 21 lines · David Hildenbrand
Quoted 21 lines
The atomic ops on futex need to provide the full barrier just like regular atomics in kernel.
Also remove pagefault_enable/disable in
futex_atomic_cmpxchg_inatomic()as core code already does thatCc David Hildenbrand <[email protected]>Cc Peter Zijlstra (Intel) <[email protected]>Cc Thomas Gleixner <[email protected]>Cc Michel Lespinasse <[email protected]>Signed-off-by Vineet Gupta <[email protected]>--- arch/arc/include/asm/futex.h | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-)arch/arc/include/asm/futex.h +0
-0@@ -20,6 +20,7 @@
Quoted 14 lines
arch/arc/include/asm/futex.h line 45
Quoted 13 lines
arch/arc/include/asm/futex.h +3
-1#define __futex_atomic_op (insn, ret, oldval, uaddr, oparg)\ \ + smp_mb(); \ __asm__ __volatile__( \ "1: llock %1, [%2] \n" \ insn "\n" \@@ -40,12 +41,14 @@ \ : "=&r" (ret), "=&r" (oldval) \ : "r" (uaddr), "r" (oparg), "ir" (-EFAULT) \ - : "cc", "memory") + : "cc", "memory"); \ +smp_mb (); \I think you should drop the ;
OK sure !
Quoted 17 lines
Quoted 10 lines
arch/arc/include/asm/futex.h +1
-0#else /* !CONFIG_ARC_HAS_LLSC */ #define __futex_atomic_op (insn, ret, oldval, uaddr, oparg)\ \ + smp_mb(); \ __asm__ __volatile__( \ "1: ld %1, [%2] \n" \ insn "\n" \arch/arc/include/asm/futex.h line 72
Quoted 6 lines
arch/arc/include/asm/futex.h +2
-1@@ -65,7 +68,8 @@ \ : "=&r" (ret), "=&r" (oldval) \ : "r" (uaddr), "r" (oparg), "ir" (-EFAULT) \ - : "cc", "memory") + : "cc", "memory"); \ +smp_mb (); \dito
OK !
Quoted 25 lines
Quoted 13 lines
arch/arc/include/asm/futex.h +1
-1#endif @@ -151,7 +155,7 @@ futex_atomic_cmpxchg_inatomic(u32 *uval, u32 __user *uaddr, u32 oldval,if (!access_ok (VERIFY_WRITE , uaddr,sizeof (int )))return -EFAULT ; -pagefault_disable (); +smp_mb (); __asm__ __volatile__(#ifdef CONFIG_ARC_HAS_LLSC arch/arc/include/asm/futex.h futex_atomic_cmpxchg_inatomic(u32 *uval, u32 __user *uaddr, u32 oldval, · line 188
Quoted 8 lines
arch/arc/include/asm/futex.h +1
-1@@ -178,7 +182,7 @@ futex_atomic_cmpxchg_inatomic(u32 *uval, u32 __user *uaddr, u32 oldval, : "r"(oldval), "r"(newval), "r"(uaddr), "ir"(-EFAULT) : "cc", "memory"); -pagefault_enable (); +smp_mb (); *uval = val;return val;Looks like
pagefault_()magic is only required forfutex_atomic_op_inuser. So this should be fine (and arc seems to be the only arch left that has in in_inatomic).
Quoted 8 lines
Not sure if you want to change the comment:
/* Compare-xchg with pagefaults disabled.
* Notes:
* -Best-Effort: Exchg happens only if compare succeeds.Maybe something like "Compare-xchg: pagefaults have to be disabled by the caller"
Will do !
Quoted 1 line
Looks sane to me.
Thx for the quick review David. It seems ARC also needs the preempt disable magic for !LLSC config. I'll send a separate patch to that effect.
Thx,
-Vineet
Quoted 4 lines
David
Re: [PATCH 1/4] ARC: add barriers to futex code
In reply to Vineet GuptaQuoted 5 lines · Vineet Gupta
The atomic ops on futex need to provide the full barrier just like regular atomics in kernel.
Also remove pagefault_enable/disable in
futex_atomic_cmpxchg_inatomic()as core code already does that
Urgh, and of course tglx just left for holidays :-)
Quoted 10 lines
+++
b/arch/arc/include/asm/futex.harch/arc/include/asm/futex.h +1
-0@@ -20,6 +20,7 @@#define __futex_atomic_op (insn, ret, oldval, uaddr, oparg)\ \ + smp_mb(); \ __asm__ __volatile__( \ "1: llock %1, [%2] \n" \ insn "\n" \
arch/arc/include/asm/futex.h line 46
Quoted 7 lines
arch/arc/include/asm/futex.h +2
-1@@ -40,12 +41,14 @@ \ : "=&r" (ret), "=&r" (oldval) \ : "r" (uaddr), "r" (oparg), "ir" (-EFAULT) \ - : "cc", "memory") + : "cc", "memory"); \ +smp_mb (); \
So:
- alhpa: only has the first smp_mb(), suggesting RELEASE
- arm: only has the first smp_mb(), suggesting RELEASE
- arm64: has store-release + smp_mb(), suggesting full barriers
- MIPS: has LLSC_MB after, suggesting ACQUIRE
- powerpc: lwsync before, sync after, full barrier
x86 is of course boring and fully ordered
Looking at the usage site of futex_atomic_op_inuser(), that's in futex_wake_op() which might suggest RELEASE is indeed sufficient.
Which leaves me puzzled on MIPS, but what do I know.
At the very least this patch isn't wrong, fully ordered is sufficient.
Re: [PATCH 1/4] ARC: add barriers to futex code
Quoted 8 lines · Peter Zijlstra
Quoted 5 lines · Vineet Gupta
The atomic ops on futex need to provide the full barrier just like regular atomics in kernel.
Also remove pagefault_enable/disable in
futex_atomic_cmpxchg_inatomic()as core code already does thatUrgh, and of course tglx just left for holidays :-)
Damn, he's really missing out on this!
Quoted 21 lines
Quoted 10 lines
+++
b/arch/arc/include/asm/futex.harch/arc/include/asm/futex.h +1
-0@@ -20,6 +20,7 @@#define __futex_atomic_op (insn, ret, oldval, uaddr, oparg)\ \ + smp_mb(); \ __asm__ __volatile__( \ "1: llock %1, [%2] \n" \ insn "\n" \arch/arc/include/asm/futex.h line 46
Quoted 7 lines
arch/arc/include/asm/futex.h +2
-1@@ -40,12 +41,14 @@ \ : "=&r" (ret), "=&r" (oldval) \ : "r" (uaddr), "r" (oparg), "ir" (-EFAULT) \ - : "cc", "memory") + : "cc", "memory"); \ +smp_mb (); \So:
Quoted 3 lines
- alhpa: only has the first
smp_mb(), suggesting RELEASE
- arm: only has the firstsmp_mb(), suggesting RELEASE
- arm64: has store-release +smp_mb(), suggesting full barriers
I'd be ok relaxing that to smp_mb() but I don't think I'm brave enough to go all the way to an STLXR. You can lose SC if you combine explicit barrier instructions with the acquire/release instructions and I dread to think what userspace is doing...
Quoted 1 line
- MIPS: has
LLSC_MBafter, suggesting ACQUIRE
Yikes, so there's a fun semantic difference there. Maybe we should go look at glibc (which only uses one of the futex ops in pthread_cond_wait iirc).
Quoted 10 lines
- powerpc: lwsync before, sync after, full barrier
x86 is of course boring and fully ordered
Looking at the usage site of
futex_atomic_op_inuser(), that's infutex_wake_op()which might suggest RELEASE is indeed sufficient.Which leaves me puzzled on MIPS, but what do I know.
At the very least this patch isn't wrong, fully ordered is sufficient.
Agreed.
Will
Re: [PATCH 1/4] ARC: add barriers to futex code
In reply to Peter ZijlstraQuoted 36 lines · Peter Zijlstra
Quoted 5 lines · Vineet Gupta
The atomic ops on futex need to provide the full barrier just like regular atomics in kernel.
Also remove pagefault_enable/disable in
futex_atomic_cmpxchg_inatomic()as core code already does thatUrgh, and of course tglx just left for holidays :-)
Quoted 10 lines
+++
b/arch/arc/include/asm/futex.h(quoted diff) +1
-0@@ -20,6 +20,7 @@ #define __futex_atomic_op(insn, ret, oldval, uaddr, oparg)\ \ + smp_mb(); \ __asm__ __volatile__( \ "1: llock %1, [%2] \n" \ insn "\n" \Review line 46
Quoted 7 lines
(quoted diff) +2
-1@@ -40,12 +41,14 @@ \ : "=&r" (ret), "=&r" (oldval) \ : "r" (uaddr), "r" (oparg), "ir" (-EFAULT) \ - : "cc", "memory") + : "cc", "memory"); \ + smp_mb(); \So:
- alhpa: only has the first
smp_mb(), suggesting RELEASE
- arm: only has the firstsmp_mb(), suggesting RELEASE
- arm64: has store-release +smp_mb(), suggesting full barriers
- MIPS: hasLLSC_MBafter, suggesting ACQUIRE
- powerpc: lwsync before, sync after, full barrier
Quoted 6 lines
x86 is of course boring and fully ordered
Looking at the usage site of
futex_atomic_op_inuser(), that's infutex_wake_op()which might suggest RELEASE is indeed sufficient.Which leaves me puzzled on MIPS, but what do I know.
So I _think_ the MIPS code is broken. The mips futex code is from 2006 and the mips smp_mb__before_llsc bits are from 2010, so its well possible this was missed.
Ralf, David, did I miss the obvious or does the below patch make sense?
---
Subject: MIPS: Fix __futex_atomic_op() for WEAK_REORDERING_BEYOND_LLSC
Current __futex_atomic_op() doesn't have smp_mb__before_llsc(), meaning it would allow prior load/stores to escape out and re-order against the ll/sc itself.
While the exact requirements on __futex_atomic_op() are a little unclear at the moment it must be either RELEASE or fully ordered, without smp_mb__before_llsc() the MIPS code is neither.
Therefore add smp_mb__before_llsc().
--- arch/mips/include/asm/futex.h | 2 ++ 1 file changed, 2 insertions(+)
arch/mips/include/asm/futex.h +2 -0
@@ -20,6 +20,8 @@#define __futex_atomic_op (insn, ret, oldval, uaddr, oparg) \ { \ + smp_mb__before_llsc(); \ + \ if (cpu_has_llsc && R10000_LLSC_WAR) { \ __asm__ __volatile__( \ " .set push \n" \
PATCH 2/4 ARC: futex cosmetics
In reply to Vineet Gupta--- arch/arc/include/asm/futex.h | 17 +++++++++-------- 1 file changed, 9 insertions(+), 8 deletions(-)
arch/arc/include/asm/futex.h +9 -8
@@ -94,6 +94,7 @@ static inline int futex_atomic_op_inuser(int encoded_op, u32 __user *uaddr)__futex_atomic_op ("mov %0, %3" , ret, oldval, uaddr, oparg);break ;case FUTEX_OP_ADD : +/* oldval = *uaddr; *uaddr += oparg ; ret = *uaddr */ __futex_atomic_op ("add %0, %1, %3" , ret, oldval, uaddr, oparg);break ;case FUTEX_OP_OR :
@@ -147,12 +148,12 @@ static inline int futex_atomic_op_inuser(int encoded_op, u32 __user *uaddr)* -user access r/w fails: return -EFAULT */ static inline int -futex_atomic_cmpxchg_inatomic (u32 *uval,u32 __user *uaddr,u32 oldval, -u32 newval) +futex_atomic_cmpxchg_inatomic (u32 *uval,u32 __user *uaddr,u32 expval, +u32 newval) { -u32 val; +u32 existval; -if (!access_ok (VERIFY_WRITE , uaddr,sizeof (int ))) +if (!access_ok (VERIFY_WRITE , uaddr,sizeof (u32)))return -EFAULT ;smp_mb ();
@@ -178,14 +179,14 @@ futex_atomic_cmpxchg_inatomic(u32 *uval, u32 __user *uaddr, u32 oldval," .word 1b, 4b \n" " .word 2b, 4b \n" " .previous\n" - : "=&r"(val) - : "r"(oldval), "r"(newval), "r"(uaddr), "ir"(-EFAULT) + : "=&r"(existval) + : "r"(expval), "r"(newval), "r"(uaddr), "ir"(-EFAULT) : "cc", "memory"); smp_mb(); - *uval = val; - return val; + *uval = existval; + return existval; }#endif
Signature · 1 line
1.9.1
PATCH 3/4 ARC: make futex_atomic_cmpxchg_inatomic() return bimodal
In reply to Vineet GuptaCallers of cmpxchg_futex_value_locked() in futex code expect bimodal return value:
!0 (essentially -EFAULT as failure)
0 (success)
Before this patch, the success return value was old value of futex, which could very well be non zero, causing caller to possibly take the failure path erroneously.
Fix that by returning 0 for success
(This fix was done back in 2011 for all upstream arches, which ARC obviously missed)
--- arch/arc/include/asm/futex.h | 19 ++++++++++--------- 1 file changed, 10 insertions(+), 9 deletions(-)
arch/arc/include/asm/futex.h +10 -9
@@ -151,6 +151,7 @@ static inline int futex_atomic_cmpxchg_inatomic(u32 *uval, u32 __user *uaddr, u32 expval, u32 newval) { +int ret =0 ;u32 existval;if (!access_ok (VERIFY_WRITE , uaddr,sizeof (u32)))
@@ -160,18 +161,18 @@ futex_atomic_cmpxchg_inatomic(u32 *uval, u32 __user *uaddr, u32 expval, __asm__ __volatile__( #ifdef CONFIG_ARC_HAS_LLSC - "1: llock %0, [%3] \n" - " brne %0, %1, 3f \n" - "2: scond %2, [%3] \n" + "1: llock %1, [%4] \n" + " brne %1, %2, 3f \n" + "2: scond %3, [%4] \n" " bnz 1b \n" #else - "1: ld %0, [%3] \n" - " brne %0, %1, 3f \n" - "2: st %2, [%3] \n" + "1: ld %1, [%4] \n" + " brne %1, %2, 3f \n" + "2: st %3, [%4] \n" #endif"3: \n" " .section .fixup,\"ax\" \n" -"4: mov %0, %4 \n" +"4: mov %0, %5 \n" " b 3b \n" " .previous \n" " .section __ex_table,\"a\" \n"
@@ -179,14 +180,14 @@ futex_atomic_cmpxchg_inatomic(u32 *uval, u32 __user *uaddr, u32 expval," .word 1b, 4b \n" " .word 2b, 4b \n" " .previous\n" - : "=&r"(existval) + : "+&r"(ret), "=&r"(existval) : "r"(expval), "r"(newval), "r"(uaddr), "ir"(-EFAULT) : "cc", "memory"); smp_mb(); *uval = existval; - return existval; + return ret; }#endif
Signature · 1 line
1.9.1
PATCH 4/4 ARC: Enable HAVE_FUTEX_CMPXCHG
In reply to Vineet GuptaARC doesn't need the runtime detection of futex cmpxchg op
--- arch/arc/Kconfig | 1 + 1 file changed, 1 insertion(+)
arch/arc/Kconfig +1 -0
@@ -22,6 +22,7 @@ config ARC select GENERIC_SMP_IDLE_THREAD select HAVE_ARCH_KGDB select HAVE_ARCH_TRACEHOOK + select HAVE_FUTEX_CMPXCHG select HAVE_IOREMAP_PROT select HAVE_KPROBES select HAVE_KRETPROBES
Signature · 1 line
1.9.1
[PATCH 5/4] ARC: ensure futex ops are atomic in !LLSC config
In reply to Vineet GuptaW/o hardware assisted atomic r-m-w the best we can do is to disable preemption.
--- arch/arc/include/asm/futex.h | 12 ++++++++++++ 1 file changed, 12 insertions(+)
arch/arc/include/asm/futex.h +12 -0
@@ -87,6 +87,9 @@ static inline int futex_atomic_op_inuser(int encoded_op, u32 __user *uaddr)if (!access_ok (VERIFY_WRITE , uaddr,sizeof (int )))return -EFAULT ; +#ifndef CONFIG_ARC_HAS_LLSC +preempt_disable ();/* to guarantee atomic r-m-w of futex op */ +#endif pagefault_disable ();switch (op) {
@@ -111,6 +114,9 @@ static inline int futex_atomic_op_inuser(int encoded_op, u32 __user *uaddr) }pagefault_enable (); +#ifndef CONFIG_ARC_HAS_LLSC +preempt_enable (); +#endif if (!ret) {switch (cmp) {
@@ -153,6 +159,9 @@ futex_atomic_cmpxchg_inatomic(u32 *uval, u32 __user *uaddr, u32 expval,if (!access_ok (VERIFY_WRITE , uaddr,sizeof (u32)))return -EFAULT ; +#ifndef CONFIG_ARC_HAS_LLSC +preempt_disable ();/* to guarantee atomic r-m-w of futex op */ +#endif smp_mb (); __asm__ __volatile__(
@@ -182,6 +191,9 @@ futex_atomic_cmpxchg_inatomic(u32 *uval, u32 __user *uaddr, u32 expval,smp_mb (); +#ifndef CONFIG_ARC_HAS_LLSC +preempt_enable (); +#endif *uval = existval;return ret; }
Signature · 1 line
1.9.1
On Thursday 06 August 2015 06:05 PM, Vineet Gupta wrote:
Hi Thomas/Peter,
Of late I've been debugging a seeming lost wakeup when running a specific EEMBC Multibench workload (4M-check -w4) on a quad core HS38 config on FPGA CONFIG_SMP, CONFIG_PREEMPT.
I've yet to nail that issue down, but in the process found some deficinecies in ARC futex backend. Can you please take a quick look and shout if there's something wrong there.
The futex backend for ARC, despite it's age, only recently started getting real testing due to recent switch to NPTL based userland (uClibc).
Thx,
-Vineet
Vineet Gupta (4):
ARC: add barriers to futex code
ARC: futex cosmetics
ARC: make futex_atomic_cmpxchg_inatomic() return bimodal
ARC: Enable HAVE_FUTEX_CMPXCHG
arch/arc/Kconfig | 1 + arch/arc/include/asm/futex.h | 44 +++++++++++++++++++++++++------------------- 2 files changed, 26 insertions(+), 19 deletions(-)
Hi Thomas,
Could u please take a look at these patches when u get a chance ?
Thx,
-Vineet