Re: [PATCH v4 4/4] perf test amd ibs: Add sample period unit test

[Arnaldo Carvalho de Melo <acme@kernel.org>]

On Tue, Apr 29, 2025 at 05:55:13PM -0300, Arnaldo Carvalho de Melo wrote:
> On Tue, Apr 29, 2025 at 03:59:38AM +0000, Ravi Bangoria wrote:
> > IBS Fetch and IBS Op PMUs has various constraints on supported sample
> > periods. Add perf unit tests to test those.
> > 
> > Running it in parallel with other tests causes intermittent failures.
> > Mark it exclusive to force it to run sequentially. Sample output on a
> > Zen5 machine:
> 
> I've applied the series and will test it now, but found some problems
> when building in some non-glibc systems, namely the use of PAGE_SIZE,
> that is used in libc headers, even in glibc, its just that in glibc we
> happen not to include that header where PAGE_SIZE gets redefined:
> 
> ⬢ [acme@toolbx perf-tools-next]$ grep PAGE_SIZE /usr/include/sys/*.h
> /usr/include/sys/user.h:#define PAGE_SIZE		(1UL << PAGE_SHIFT)
> /usr/include/sys/user.h:#define PAGE_MASK		(~(PAGE_SIZE-1))
> /usr/include/sys/user.h:#define NBPG			PAGE_SIZE
> ⬢ [acme@toolbx perf-tools-next]$
> 
> So I folded the following patch, see if it is acceptable and please ack.
> 
> Thanks for respining it!

Another issue when building with clang on musl:

arch/x86/tests/amd-ibs-period.c:81:3: error: no matching function for call to 'memcpy'
                memcpy(func, insn1, sizeof(insn1));
                ^~~~~~
/usr/include/string.h:27:7: note: candidate function not viable: no known conversion from 'int (*)(void)' to 'void *' for 1st argument
void *memcpy (void *__restrict, const void *__restrict, size_t);
      ^
/usr/include/fortify/string.h:40:27: note: candidate function not viable: no known conversion from 'int (*)(void)' to 'void *const' for 1st argument
_FORTIFY_FN(memcpy) void *memcpy(void * _FORTIFY_POS0 __od,
                          ^
arch/x86/tests/amd-ibs-period.c:87:3: error: no matching function for call to 'memcpy'
                memcpy(func, insn2, sizeof(insn2));
                ^~~~~~
/usr/include/string.h:27:7: note: candidate function not viable: no known conversion from 'int (*)(void)' to 'void *' for 1st argument
void *memcpy (void *__restrict, const void *__restrict, size_t);
      ^
/usr/include/fortify/string.h:40:27: note: candidate function not viable: no known conversion from 'int (*)(void)' to 'void *const' for 1st argument
_FORTIFY_FN(memcpy) void *memcpy(void * _FORTIFY_POS0 __od,
                          ^
2 errors generated.
  CC      /tmp/build/perf/ui/browsers/header.o
  CC      /tmp/build/perf/arch/x86/util/mem-events.o

Adding the patch below cures it, still need to test on a Zen 5 system.

These issues were just in the regression test.

- Arnaldo

diff --git a/tools/perf/arch/x86/tests/amd-ibs-period.c b/tools/perf/arch/x86/tests/amd-ibs-period.c
index 946b0a377554fb81..a198434da9b5c4a1 100644
--- a/tools/perf/arch/x86/tests/amd-ibs-period.c
+++ b/tools/perf/arch/x86/tests/amd-ibs-period.c
@@ -78,13 +78,13 @@ static int dummy_workload_1(unsigned long count)
 	else if (count > 10000000)
 		count = 10000000;
 	while (count--) {
-		memcpy(func, insn1, sizeof(insn1));
+		memcpy((void *)func, insn1, sizeof(insn1));
 		if (func() != 1) {
 			pr_debug("ERROR insn1\n");
 			ret = -1;
 			goto out;
 		}
-		memcpy(func, insn2, sizeof(insn2));
+		memcpy((void *)func, insn2, sizeof(insn2));
 		if (func() != 2) {
 			pr_debug("ERROR insn2\n");
 			ret = -1;