[PATCH] fix for_each_cpu() again for NR_CPUS=1

Jan Beulich posted 1 patch 1 week, 6 days ago
Test gitlab-ci failed
Patches applied successfully (tree, apply log)
git fetch https://gitlab.com/xen-project/patchew/xen tags/patchew/4d993e67-2bdf-9ac3-f78e-daf279642de1@suse.com

[PATCH] fix for_each_cpu() again for NR_CPUS=1

Posted by Jan Beulich 1 week, 6 days ago
Unfortunately aa50f45332f1 ("xen: fix for_each_cpu when NR_CPUS=1") has
caused quite a bit of fallout with gcc10, e.g. (there are at least two
more similar ones, and I didn't bother trying to find them all):

In file included from .../xen/include/xen/config.h:13,
                 from <command-line>:
core_parking.c: In function ‘core_parking_power’:
.../xen/include/asm/percpu.h:12:51: error: array subscript 1 is above array bounds of ‘long unsigned int[1]’ [-Werror=array-bounds]
   12 |     (*RELOC_HIDE(&per_cpu__##var, __per_cpu_offset[cpu]))
.../xen/include/xen/compiler.h:141:29: note: in definition of macro ‘RELOC_HIDE’
  141 |     (typeof(ptr)) (__ptr + (off)); })
      |                             ^~~
core_parking.c:133:39: note: in expansion of macro ‘per_cpu’
  133 |             core_tmp = cpumask_weight(per_cpu(cpu_core_mask, cpu));
      |                                       ^~~~~~~
In file included from .../xen/include/xen/percpu.h:4,
                 from .../xen/include/asm/msr.h:7,
                 from .../xen/include/asm/time.h:5,
                 from .../xen/include/xen/time.h:76,
                 from .../xen/include/xen/spinlock.h:4,
                 from .../xen/include/xen/cpu.h:5,
                 from core_parking.c:19:
.../xen/include/asm/percpu.h:6:22: note: while referencing ‘__per_cpu_offset’
    6 | extern unsigned long __per_cpu_offset[NR_CPUS];
      |                      ^~~~~~~~~~~~~~~~

One of the further errors even went as far as claiming that an array
index (range) of [0, 0] was outside the bounds of a [1] array, so
something fishy is pretty clearly going on there.

The compiler apparently wants to be able to see that the loop isn't
really a loop in order to avoid triggering such warnings, yet what
exactly makes it consider the loop exit condition constant and within
the [0, 1] range isn't obvious - using ((mask)->bits[0] & 1) instead of
cpumask_test_cpu() for example did _not_ help.

Re-instate a special form of for_each_cpu(), experimentally "proven" to
avoid the diagnostics.

Signed-off-by: Jan Beulich <jbeulich@suse.com>

--- a/xen/include/xen/cpumask.h
+++ b/xen/include/xen/cpumask.h
@@ -368,10 +368,15 @@ static inline void free_cpumask_var(cpum
 #define FREE_CPUMASK_VAR(m) free_cpumask_var(m)
 #endif
 
+#if NR_CPUS > 1
 #define for_each_cpu(cpu, mask)			\
 	for ((cpu) = cpumask_first(mask);	\
 	     (cpu) < nr_cpu_ids;		\
 	     (cpu) = cpumask_next(cpu, mask))
+#else /* NR_CPUS == 1 */
+#define for_each_cpu(cpu, mask)			\
+	for ((cpu) = 0; (cpu) < cpumask_test_cpu(0, mask); ++(cpu))
+#endif /* NR_CPUS */
 
 /*
  * The following particular system cpumasks and operations manage

Revert NR_CPUS=1 fix from 4.15 (was: Re: [PATCH] fix for_each_cpu() again for NR_CPUS=1)

Posted by Roger Pau Monné 1 week, 5 days ago
On Wed, Mar 31, 2021 at 04:52:47PM +0200, Jan Beulich wrote:
> Unfortunately aa50f45332f1 ("xen: fix for_each_cpu when NR_CPUS=1") has
> caused quite a bit of fallout with gcc10, e.g. (there are at least two
> more similar ones, and I didn't bother trying to find them all):
> 
> In file included from .../xen/include/xen/config.h:13,
>                  from <command-line>:
> core_parking.c: In function ‘core_parking_power’:
> .../xen/include/asm/percpu.h:12:51: error: array subscript 1 is above array bounds of ‘long unsigned int[1]’ [-Werror=array-bounds]
>    12 |     (*RELOC_HIDE(&per_cpu__##var, __per_cpu_offset[cpu]))
> .../xen/include/xen/compiler.h:141:29: note: in definition of macro ‘RELOC_HIDE’
>   141 |     (typeof(ptr)) (__ptr + (off)); })
>       |                             ^~~
> core_parking.c:133:39: note: in expansion of macro ‘per_cpu’
>   133 |             core_tmp = cpumask_weight(per_cpu(cpu_core_mask, cpu));
>       |                                       ^~~~~~~
> In file included from .../xen/include/xen/percpu.h:4,
>                  from .../xen/include/asm/msr.h:7,
>                  from .../xen/include/asm/time.h:5,
>                  from .../xen/include/xen/time.h:76,
>                  from .../xen/include/xen/spinlock.h:4,
>                  from .../xen/include/xen/cpu.h:5,
>                  from core_parking.c:19:
> .../xen/include/asm/percpu.h:6:22: note: while referencing ‘__per_cpu_offset’
>     6 | extern unsigned long __per_cpu_offset[NR_CPUS];
>       |                      ^~~~~~~~~~~~~~~~

At this point, should be consider reverting the original fix from the
4.15 branch, so that we don't release something that's build broken
with gcc 10?

Roger.

Revert NR_CPUS=1 fix from 4.15 (was: Re: [PATCH] fix for_each_cpu() again for NR_CPUS=1)

Posted by Ian Jackson 1 week ago
Roger Pau Monné writes ("Revert NR_CPUS=1 fix from 4.15 (was: Re: [PATCH] fix for_each_cpu() again for NR_CPUS=1)"):
> At this point, should be consider reverting the original fix from the
> 4.15 branch, so that we don't release something that's build broken
> with gcc 10?

Yes.  I think so.

Release-Acked-by: Ian Jackson <iwj@xenproject.org>

But please leave it to me to commit it.  I will do so at or after
around 14:00 UK time (13:00 UTC) today unless someone objects.

Thanks,
Ian.

Re: Revert NR_CPUS=1 fix from 4.15 (was: Re: [PATCH] fix for_each_cpu() again for NR_CPUS=1)

Posted by Jan Beulich 1 week, 5 days ago
On 01.04.2021 11:00, Roger Pau Monné wrote:
> On Wed, Mar 31, 2021 at 04:52:47PM +0200, Jan Beulich wrote:
>> Unfortunately aa50f45332f1 ("xen: fix for_each_cpu when NR_CPUS=1") has
>> caused quite a bit of fallout with gcc10, e.g. (there are at least two
>> more similar ones, and I didn't bother trying to find them all):
>>
>> In file included from .../xen/include/xen/config.h:13,
>>                  from <command-line>:
>> core_parking.c: In function ‘core_parking_power’:
>> .../xen/include/asm/percpu.h:12:51: error: array subscript 1 is above array bounds of ‘long unsigned int[1]’ [-Werror=array-bounds]
>>    12 |     (*RELOC_HIDE(&per_cpu__##var, __per_cpu_offset[cpu]))
>> .../xen/include/xen/compiler.h:141:29: note: in definition of macro ‘RELOC_HIDE’
>>   141 |     (typeof(ptr)) (__ptr + (off)); })
>>       |                             ^~~
>> core_parking.c:133:39: note: in expansion of macro ‘per_cpu’
>>   133 |             core_tmp = cpumask_weight(per_cpu(cpu_core_mask, cpu));
>>       |                                       ^~~~~~~
>> In file included from .../xen/include/xen/percpu.h:4,
>>                  from .../xen/include/asm/msr.h:7,
>>                  from .../xen/include/asm/time.h:5,
>>                  from .../xen/include/xen/time.h:76,
>>                  from .../xen/include/xen/spinlock.h:4,
>>                  from .../xen/include/xen/cpu.h:5,
>>                  from core_parking.c:19:
>> .../xen/include/asm/percpu.h:6:22: note: while referencing ‘__per_cpu_offset’
>>     6 | extern unsigned long __per_cpu_offset[NR_CPUS];
>>       |                      ^~~~~~~~~~~~~~~~
> 
> At this point, should be consider reverting the original fix from the
> 4.15 branch, so that we don't release something that's build broken
> with gcc 10?

Well, I didn't propose reverting (or taking this fix) because I think
build breakage is better than runtime breakage. But in the end, Ian,
it's up to you.

Jan

Re: Revert NR_CPUS=1 fix from 4.15 (was: Re: [PATCH] fix for_each_cpu() again for NR_CPUS=1)

Posted by Roger Pau Monné 1 week, 5 days ago
On Thu, Apr 01, 2021 at 11:26:03AM +0200, Jan Beulich wrote:
> On 01.04.2021 11:00, Roger Pau Monné wrote:
> > On Wed, Mar 31, 2021 at 04:52:47PM +0200, Jan Beulich wrote:
> >> Unfortunately aa50f45332f1 ("xen: fix for_each_cpu when NR_CPUS=1") has
> >> caused quite a bit of fallout with gcc10, e.g. (there are at least two
> >> more similar ones, and I didn't bother trying to find them all):
> >>
> >> In file included from .../xen/include/xen/config.h:13,
> >>                  from <command-line>:
> >> core_parking.c: In function ‘core_parking_power’:
> >> .../xen/include/asm/percpu.h:12:51: error: array subscript 1 is above array bounds of ‘long unsigned int[1]’ [-Werror=array-bounds]
> >>    12 |     (*RELOC_HIDE(&per_cpu__##var, __per_cpu_offset[cpu]))
> >> .../xen/include/xen/compiler.h:141:29: note: in definition of macro ‘RELOC_HIDE’
> >>   141 |     (typeof(ptr)) (__ptr + (off)); })
> >>       |                             ^~~
> >> core_parking.c:133:39: note: in expansion of macro ‘per_cpu’
> >>   133 |             core_tmp = cpumask_weight(per_cpu(cpu_core_mask, cpu));
> >>       |                                       ^~~~~~~
> >> In file included from .../xen/include/xen/percpu.h:4,
> >>                  from .../xen/include/asm/msr.h:7,
> >>                  from .../xen/include/asm/time.h:5,
> >>                  from .../xen/include/xen/time.h:76,
> >>                  from .../xen/include/xen/spinlock.h:4,
> >>                  from .../xen/include/xen/cpu.h:5,
> >>                  from core_parking.c:19:
> >> .../xen/include/asm/percpu.h:6:22: note: while referencing ‘__per_cpu_offset’
> >>     6 | extern unsigned long __per_cpu_offset[NR_CPUS];
> >>       |                      ^~~~~~~~~~~~~~~~
> > 
> > At this point, should be consider reverting the original fix from the
> > 4.15 branch, so that we don't release something that's build broken
> > with gcc 10?
> 
> Well, I didn't propose reverting (or taking this fix) because I think
> build breakage is better than runtime breakage. But in the end, Ian,
> it's up to you.

Oh, right, sorry. The build issue only happens with NR_CPUS=1, in
which case I agree, there's no need to do anything in 4.15 IMO.

Sorry for bothering.

Roger.

Re: Revert NR_CPUS=1 fix from 4.15 (was: Re: [PATCH] fix for_each_cpu() again for NR_CPUS=1)

Posted by Ian Jackson 1 week ago
Roger Pau Monné writes ("Re: Revert NR_CPUS=1 fix from 4.15 (was: Re: [PATCH] fix for_each_cpu() again for NR_CPUS=1)"):
> On Thu, Apr 01, 2021 at 11:26:03AM +0200, Jan Beulich wrote:
> > Well, I didn't propose reverting (or taking this fix) because I think
> > build breakage is better than runtime breakage. But in the end, Ian,
> > it's up to you.
> 
> Oh, right, sorry. The build issue only happens with NR_CPUS=1, in
> which case I agree, there's no need to do anything in 4.15 IMO.

Oh.  Right.  I had the impression that the build breakage broke other
configurations too.

Since you're saying that's not the case, please disregard my earlier
mail.

Ian.

Re: [PATCH] fix for_each_cpu() again for NR_CPUS=1

Posted by Dario Faggioli 1 week, 6 days ago
On Wed, 2021-03-31 at 16:52 +0200, Jan Beulich wrote:
> Unfortunately aa50f45332f1 ("xen: fix for_each_cpu when NR_CPUS=1")
> has
> caused quite a bit of fallout with gcc10, e.g. (there are at least
> two
> more similar ones, and I didn't bother trying to find them all):
> 
Oh, wow... Sorry about that. I was sure I had checked (and with gcc10),
but clearly I'm wrong.

> [...]
> 
> Re-instate a special form of for_each_cpu(), experimentally "proven"
> to
> avoid the diagnostics.
> 
> Signed-off-by: Jan Beulich <jbeulich@suse.com>
> 
Reviewed-by: Dario Faggioli <dfaggioli@suse.com>

Thanks and Regards
-- 
Dario Faggioli, Ph.D
http://about.me/dario.faggioli
Virtualization Software Engineer
SUSE Labs, SUSE https://www.suse.com/
-------------------------------------------------------------------
<<This happens because _I_ choose it to happen!>> (Raistlin Majere)

Re: [PATCH] fix for_each_cpu() again for NR_CPUS=1

Posted by Jan Beulich 1 week, 6 days ago
On 31.03.2021 18:55, Dario Faggioli wrote:
> On Wed, 2021-03-31 at 16:52 +0200, Jan Beulich wrote:
>> Unfortunately aa50f45332f1 ("xen: fix for_each_cpu when NR_CPUS=1")
>> has
>> caused quite a bit of fallout with gcc10, e.g. (there are at least
>> two
>> more similar ones, and I didn't bother trying to find them all):
>>
> Oh, wow... Sorry about that. I was sure I had checked (and with gcc10),
> but clearly I'm wrong.

Perhaps you did try a debug build, while I was seeing the issues in
a non-debug one?

>> Re-instate a special form of for_each_cpu(), experimentally "proven"
>> to
>> avoid the diagnostics.
>>
>> Signed-off-by: Jan Beulich <jbeulich@suse.com>
>>
> Reviewed-by: Dario Faggioli <dfaggioli@suse.com>

Thanks.

Jan