h8300 build failures ijn linux-next

PATCH In review Guenter Roeck 5 participants 13 messages Aug 24, 2015 – Aug 25, 2015 RCU

lore

Share & tools

lore

Guenter Roeck

author

2015-08-24 18:34 UTC #0 lore raw

Hi,

In linux-next as of today (0824), all h8300 builds fail for me with an internal compiler error.

Building h8300:allnoconfig ... failed
--------------
Error log:

In file included from include/linux/rcupdate.h:429:0,
from include/linux/rcusync.h:5,
from kernel/rcu/sync.c:1:
include/linux/rcutiny.h: In function ‘rcu_barrier_sched’:
include/linux/rcutiny.h:55:20: internal compiler error: Segmentation fault
static inline void rcu_barrier_sched(void)

I tried both gcc 5.1 and 5.2, built with the patch set available to me.

Is there a more recent compiler / patch set available that might fix the problem ?

Thanks,
Guenter

Yoshinori Sato

2015-08-25 05:56 UTC #1 lore raw

Quoted 16 lines · Guenter Roeck

Hi,
In linux-next as of today (0824), all h8300 builds fail for me with an internal compiler error.
Building h8300:allnoconfig ... failed
--------------
Error log:
In file included from include/linux/rcupdate.h:429:0,
from include/linux/rcusync.h:5,
from kernel/rcu/sync.c:1:
include/linux/rcutiny.h: In function ‘rcu_barrier_sched’:
include/linux/rcutiny.h:55:20: internal compiler error: Segmentation fault
static inline void rcu_barrier_sched(void)

Quoted 6 lines

                    ^
I tried both gcc 5.1 and 5.2, built with the patch set available to me.
Is there a more recent compiler / patch set available that might fix the problem ?

No.
I tried gcc 6.0. It got same error.
It looks optimization problem on h8300 target.
I will investigate it.

Quoted 2 lines

Thanks,
Guenter

Signature · 2 lines

Yoshinori Sato
<[email protected]>

Mark Salter

2015-08-25 14:27 UTC #2 lore raw

Quoted 24 lines · Yoshinori Sato

Quoted 22 lines · Guenter Roeck
Hi,
In linux-next as of today (0824), all h8300 builds fail for me with an internal
compiler error.
Building h8300:allnoconfig ... failed
--------------
Error log:
In file included from include/linux/rcupdate.h:429:0,
from include/linux/rcusync.h:5,
from kernel/rcu/sync.c:1:
include/linux/rcutiny.h: In function ‘rcu_barrier_sched’:
include/linux/rcutiny.h:55:20: internal compiler error: Segmentation fault
static inline void rcu_barrier_sched(void)
                    ^
I tried both gcc 5.1 and 5.2, built with the patch set available to me.

Quoted 8 lines

Quoted 3 lines
Is there a more recent compiler / patch set available that might fix the problem ?
No.
I tried gcc 6.0. It got same error.
It looks optimization problem on h8300 target.
I will investigate it.

I get the same ICE with c6x gcc 5.1

Quoted 3 lines

Quoted 2 lines
Thanks,
Guenter

Guenter Roeck

author

2015-08-25 15:03 UTC #3 lore raw

Re: h8300 build failures in linux-next (in rcu code, gcc 5.x related, not limited to h8300)

Hi Mark,

Quoted 29 lines · Mark Salter

Quoted 28 lines · Yoshinori Sato
Quoted 25 lines · Guenter Roeck
Hi,
In linux-next as of today (0824), all h8300 builds fail for me with an internal
compiler error.
Building h8300:allnoconfig ... failed
--------------
Error log:
In file included from include/linux/rcupdate.h:429:0,
from include/linux/rcusync.h:5,
from kernel/rcu/sync.c:1:
include/linux/rcutiny.h: In function ‘rcu_barrier_sched’:
include/linux/rcutiny.h:55:20: internal compiler error: Segmentation fault
static inline void rcu_barrier_sched(void)
                     ^
I tried both gcc 5.1 and 5.2, built with the patch set available to me.
Is there a more recent compiler / patch set available that might fix the problem ?

Quoted 7 lines

Quoted 4 lines
No.
I tried gcc 6.0. It got same error.
It looks optimization problem on h8300 target.
I will investigate it.
I get the same ICE with c6x gcc 5.1

Interesting. h8300 was the only gcc 5+ compiler I had used so far. Just for fun, I built an alpha toolchain with gcc 5.1. Same problem there. So it looks like it isn't a h8300 problem after all, it is a gcc problem.

Updated subject line, and copied rcu maintainers.

Guenter

Oleg Nesterov

2015-08-25 16:14 UTC #4 lore raw

Re: h8300 build failures in linux-next (in rcu code, gcc 5.x related, not limited to h8300)

Quoted 38 lines · Guenter Roeck

Hi Mark,
Quoted 34 lines · Mark Salter
Quoted 32 lines · Yoshinori Sato
Quoted 18 lines · Guenter Roeck
Hi,
In linux-next as of today (0824), all h8300 builds fail for me with an internal
compiler error.
Building h8300:allnoconfig ... failed
--------------
Error log:
In file included from include/linux/rcupdate.h:429:0,
from include/linux/rcusync.h:5,
from kernel/rcu/sync.c:1:
include/linux/rcutiny.h: In function ‘rcu_barrier_sched’:
include/linux/rcutiny.h:55:20: internal compiler error: Segmentation fault
static inline void rcu_barrier_sched(void)
Quoted 7 lines
                     ^
I tried both gcc 5.1 and 5.2, built with the patch set available to me.
Is there a more recent compiler / patch set available that might fix the problem ?
No.
I tried gcc 6.0. It got same error.
It looks optimization problem on h8300 target.
I will investigate it.

Quoted 8 lines

Quoted 2 lines
I get the same ICE with c6x gcc 5.1
Interesting. h8300 was the only gcc 5+ compiler I had used so far. Just for fun, I built an alpha toolchain with gcc 5.1. Same problem there. So it looks like it isn't a h8300 problem after all, it is a gcc problem.
Updated subject line, and copied rcu maintainers.

Damn. This was caused by my patch, I need to react, and I do not know what to say ;)

OK, this is gcc bug. But we probably need the workaround, and I don't have a h8300 machine so I can't test the fix.

Could you please try to compile kernel/rcu/sync.o with the patch below? Just to know how "stable" this bug is. I expect the compilation should fail again, rcutiny.h:rcu_barrier() is "inline" too.

Oleg.

kernel/rcu/sync.c +0 -1

@@ -24,7 +24,6 @@ static const struct {
 	[RCU_SCHED_SYNC] = {
 		.sync = synchronize_sched,
 		.call = call_rcu_sched,
-		.wait = rcu_barrier_sched,
 		__INIT_HELD(rcu_read_lock_sched_held)
 	},
 	[RCU_BH_SYNC] = {

Oleg Nesterov

2015-08-25 16:25 UTC #5 lore raw

Re: h8300 build failures in linux-next (in rcu code, gcc 5.x related, not limited to h8300)

Quoted 7 lines · Oleg Nesterov

OK, this is gcc bug. But we probably need the workaround, and I don't have a h8300 machine so I can't test the fix.
Could you please try to compile kernel/rcu/sync.o with the patch below? Just to know how "stable" this bug is. I expect the compilation should fail again, rcutiny.h:rcu_barrier() is "inline" too.

Wait, I reproduced the same problem on x86. Thanks.

Oleg.

Yoshinori Sato

2015-08-25 16:52 UTC #6 lore raw

Re: h8300 build failures in linux-next (in rcu code, gcc 5.x related, not limited to h8300)

In reply to Guenter Roeck

Quoted 41 lines · Guenter Roeck

Hi Mark,
Quoted 29 lines · Mark Salter
Quoted 28 lines · Yoshinori Sato
Quoted 25 lines · Guenter Roeck
Hi,
In linux-next as of today (0824), all h8300 builds fail for me with an internal
compiler error.
Building h8300:allnoconfig ... failed
--------------
Error log:
In file included from include/linux/rcupdate.h:429:0,
from include/linux/rcusync.h:5,
from kernel/rcu/sync.c:1:
include/linux/rcutiny.h: In function ‘rcu_barrier_sched’:
include/linux/rcutiny.h:55:20: internal compiler error: Segmentation fault
static inline void rcu_barrier_sched(void)
                     ^
I tried both gcc 5.1 and 5.2, built with the patch set available to me.
Is there a more recent compiler / patch set available that might fix the problem ?
Quoted 7 lines
Quoted 4 lines
No.
I tried gcc 6.0. It got same error.
It looks optimization problem on h8300 target.
I will investigate it.
I get the same ICE with c6x gcc 5.1

Quoted 8 lines

Interesting. h8300 was the only gcc 5+ compiler I had used so far. Just for fun, I built an alpha toolchain with gcc 5.1. Same problem there. So it looks like it isn't a h8300 problem after all, it is a gcc problem.
Updated subject line, and copied rcu maintainers.
Guenter

Yes.
gcc bug #67055.
Already fixed in trunk.

Signature · 2 lines

Yoshinori Sato
<[email protected]>

Oleg Nesterov

2015-08-25 17:36 UTC #7 lore raw

Re: h8300 build failures in linux-next (in rcu code, gcc 5.x related, not limited to h8300)

Quoted 4 lines · Yoshinori Sato

Yes.
gcc bug #67055.
Already fixed in trunk.

Yes, thanks a lot.

Paul, it seems that gcc actually dislikes your ec90a194a "rcu:
Create a synchronize_rcu_mult()" commit ;) rcu/sync.c was just lucky enough to trigger the problem.

I'll try to make the fix today...

Oleg.

Oleg Nesterov

2015-08-25 18:05 UTC #8 lore raw

Re: h8300 build failures in linux-next (in rcu code, gcc 5.x related, not limited to h8300)

Quoted 7 lines · Oleg Nesterov

Quoted 4 lines · Yoshinori Sato
Yes.
gcc bug #67055.
Already fixed in trunk.

Quoted 7 lines

Yes, thanks a lot.
Paul, it seems that gcc actually dislikes your ec90a194a "rcu:
Create a synchronize_rcu_mult()" commit ;) rcu/sync.c was just lucky enough to trigger the problem.
I'll try to make the fix today...

Mark, Guenter, any chance you can check if the patch below helps? It does on x86.

Oleg.

include/linux/rcupdate.h +5 -6

@@ -230,12 +230,11 @@ void __wait_rcu_gp(bool checktiny, int n, call_rcu_func_t *crcu_array,
 		   struct rcu_synchronize *rs_array);
 
 #define _wait_rcu_gp(checktiny, ...) \
-do { \
-	call_rcu_func_t __crcu_array[] = { __VA_ARGS__ }; \
-	const int __n = ARRAY_SIZE(__crcu_array); \
-	struct rcu_synchronize __rs_array[__n]; \
-	\
-	__wait_rcu_gp(checktiny, __n, __crcu_array, __rs_array); \
+do {									\
+	call_rcu_func_t __crcu_array[] = { __VA_ARGS__ }; 		\
+	struct rcu_synchronize __rs_array[ARRAY_SIZE(__crcu_array)];	\
+	__wait_rcu_gp(checktiny, ARRAY_SIZE(__crcu_array),		\
+			__crcu_array, __rs_array); 			\
 } while (0)
 
 #define wait_rcu_gp(...) _wait_rcu_gp(false, __VA_ARGS__)

Mark Salter

2015-08-25 18:16 UTC #9 lore raw

Re: h8300 build failures in linux-next (in rcu code, gcc 5.x related, not limited to h8300)

Quoted 10 lines · Oleg Nesterov

Quoted 9 lines · Oleg Nesterov
Quoted 4 lines · Yoshinori Sato
Yes.
gcc bug #67055.
Already fixed in trunk.
Yes, thanks a lot.

Quoted 8 lines

Quoted 5 lines
Paul, it seems that gcc actually dislikes your ec90a194a "rcu:
Create a synchronize_rcu_mult()" commit ;) rcu/sync.c was just lucky enough to trigger the problem.
I'll try to make the fix today...
Mark, Guenter, any chance you can check if the patch below helps? It does on x86.

Works for c6x.

Quoted 30 lines

Oleg.

include/linux/rcupdate.h +3 -6

@@ -230,12 +230,11 @@ void __wait_rcu_gp(bool checktiny, int n, 
 call_rcu_func_t *crcu_array,
 		   struct rcu_synchronize *rs_array);
 
 #define _wait_rcu_gp(checktiny, ...) \
-do { \
-	call_rcu_func_t __crcu_array[] = { __VA_ARGS__ }; \
-	const int __n = ARRAY_SIZE(__crcu_array); \
-	struct rcu_synchronize __rs_array[__n]; \
-	\
-	__wait_rcu_gp(checktiny, __n, __crcu_array, __rs_array); \
+do {									
 \
+	call_rcu_func_t __crcu_array[] = { __VA_ARGS__ }; 		
 \
+	struct rcu_synchronize __rs_array[ARRAY_SIZE(__crcu_array)];

include/linux/rcupdate.h +1 -0

+	__wait_rcu_gp(checktiny, ARRAY_SIZE(__crcu_array),

include/linux/rcupdate.h +1 -0

+			__crcu_array, __rs_array);

\
} while (0)

#define wait_rcu_gp(...) _wait_rcu_gp(false, __VA_ARGS__)

Oleg Nesterov

2015-08-25 18:47 UTC #10 lore raw

PATCH 1/1 rcu: change _wait_rcu_gp() to work around gcc 67055 bug

In reply to Guenter Roeck

The code like

	const int n = const-expr;
	whatever_t array[n];

in inline function can confuse gcc, see

	https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67055#c13

This leads to

In file included from include/linux/rcupdate.h:429:0,
from include/linux/rcusync.h:5,
from kernel/rcu/sync.c:1:
include/linux/rcutiny.h: In function 'rcu_barrier_sched':
include/linux/rcutiny.h:55:20: internal compiler error: Segmentation fault
static inline void rcu_barrier_sched(void)

build failure after ec90a194a "rcu: Create a synchronize_rcu_mult()" and 114b7fd4b "rcu: Create rcu_sync infrastructure" meet each other.

Reported-and-tested-by Mark Salter <[email protected]>

Reported-by Guenter Roeck <[email protected]>

Signed-off-by Oleg Nesterov <[email protected]>

---
 include/linux/rcupdate.h | 11 +++++------
 1 file changed, 5 insertions(+), 6 deletions(-)

include/linux/rcupdate.h +5 -6

index ff47651..c353c2b 100644
@@ -230,12 +230,11 @@ void __wait_rcu_gp(bool checktiny, int n, call_rcu_func_t *crcu_array,
 		   struct rcu_synchronize *rs_array);
 
 #define _wait_rcu_gp(checktiny, ...) \
-do { \
-	call_rcu_func_t __crcu_array[] = { __VA_ARGS__ }; \
-	const int __n = ARRAY_SIZE(__crcu_array); \
-	struct rcu_synchronize __rs_array[__n]; \
-	\
-	__wait_rcu_gp(checktiny, __n, __crcu_array, __rs_array); \
+do {									\
+	call_rcu_func_t __crcu_array[] = { __VA_ARGS__ }; 		\
+	struct rcu_synchronize __rs_array[ARRAY_SIZE(__crcu_array)];	\
+	__wait_rcu_gp(checktiny, ARRAY_SIZE(__crcu_array),		\
+			__crcu_array, __rs_array); 			\
 } while (0)
 
 #define wait_rcu_gp(...) _wait_rcu_gp(false, __VA_ARGS__)

Signature · 1 line

2.4.3

Paul E. McKenney

2015-08-25 19:00 UTC #11 lore raw

Re: [PATCH] rcu: change _wait_rcu_gp() to work around gcc 67055 bug

Quoted 19 lines · Oleg Nesterov

The code like
	const int n = const-expr;
	whatever_t array[n];
in inline function can confuse gcc, see
	https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67055#c13
This leads to
In file included from include/linux/rcupdate.h:429:0,
from include/linux/rcusync.h:5,
from kernel/rcu/sync.c:1:
include/linux/rcutiny.h: In function 'rcu_barrier_sched':
include/linux/rcutiny.h:55:20: internal compiler error: Segmentation fault
static inline void rcu_barrier_sched(void)

Quoted 6 lines

build failure after ec90a194a "rcu: Create a synchronize_rcu_mult()" and 114b7fd4b "rcu: Create rcu_sync infrastructure" meet each other.
Reported-and-tested-by Mark Salter <[email protected]>
Reported-by Guenter Roeck <[email protected]>
Signed-off-by Oleg Nesterov <[email protected]>

Ouch!!! Queued for testing, in the absence of problems, will push into the upcoming merge window.

Thank you all for the testing and the fix!

Thanx, Paul

Quoted 30 lines

---
 include/linux/rcupdate.h | 11 +++++------
 1 file changed, 5 insertions(+), 6 deletions(-)

include/linux/rcupdate.h +5 -6

index ff47651..c353c2b 100644
@@ -230,12 +230,11 @@ void __wait_rcu_gp(bool checktiny, int n, call_rcu_func_t *crcu_array,
 		   struct rcu_synchronize *rs_array);
 
 #define _wait_rcu_gp(checktiny, ...) \
-do { \
-	call_rcu_func_t __crcu_array[] = { __VA_ARGS__ }; \
-	const int __n = ARRAY_SIZE(__crcu_array); \
-	struct rcu_synchronize __rs_array[__n]; \
-	\
-	__wait_rcu_gp(checktiny, __n, __crcu_array, __rs_array); \
+do {									\
+	call_rcu_func_t __crcu_array[] = { __VA_ARGS__ }; 		\
+	struct rcu_synchronize __rs_array[ARRAY_SIZE(__crcu_array)];	\
+	__wait_rcu_gp(checktiny, ARRAY_SIZE(__crcu_array),		\
+			__crcu_array, __rs_array); 			\
 } while (0)
 
 #define wait_rcu_gp(...) _wait_rcu_gp(false, __VA_ARGS__)

Signature · 1 line

2.4.3

Oleg Nesterov

2015-08-25 19:04 UTC #12 lore raw

Re: [PATCH] rcu: change _wait_rcu_gp() to work around gcc 67055 bug

In reply to Oleg Nesterov

Quoted 20 lines · Oleg Nesterov

The code like
	const int n = const-expr;
	whatever_t array[n];
in inline function can confuse gcc, see
	https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67055#c13
This leads to
In file included from include/linux/rcupdate.h:429:0,
from include/linux/rcusync.h:5,
from kernel/rcu/sync.c:1:
include/linux/rcutiny.h: In function 'rcu_barrier_sched':
include/linux/rcutiny.h:55:20: internal compiler error: Segmentation fault
static inline void rcu_barrier_sched(void)

Quoted 2 lines

build failure after ec90a194a "rcu: Create a synchronize_rcu_mult()" and 114b7fd4b "rcu: Create rcu_sync infrastructure" meet each other.

and just in case... make make kernel/rcu/rcutorture.o fails the same way if CONFIG_RCU_TORTURE_TEST && CONFIG_TINY_RCU. So we need a work- around even before/without rcu_sync.

Oleg.