Skip to content

Commit f8c74cf

Browse files
yghannamgregkh
authored andcommitted
x86/cpu/AMD: Fix cpu_llc_id for AMD Fam17h systems
commit b0b6e86 upstream. cpu_llc_id (Last Level Cache ID) derivation on AMD Fam17h has an underflow bug when extracting the socket_id value. It starts from 0 so subtracting 1 from it will result in an invalid value. This breaks scheduling topology later on since the cpu_llc_id will be incorrect. For example, the the cpu_llc_id of the *other* CPU in the loops in set_cpu_sibling_map() underflows and we're generating the funniest thread_siblings masks and then when I run 8 threads of nbench, they get spread around the LLC domains in a very strange pattern which doesn't give you the normal scheduling spread one would expect for performance. Other things like EDAC use cpu_llc_id so they will be b0rked too. So, the APIC ID is preset in APICx020 for bits 3 and above: they contain the core complex, node and socket IDs. The LLC is at the core complex level so we can find a unique cpu_llc_id by right shifting the APICID by 3 because then the least significant bit will be the Core Complex ID. Tested-by: Borislav Petkov <[email protected]> Signed-off-by: Yazen Ghannam <[email protected]> [ Cleaned up and extended the commit message. ] Signed-off-by: Borislav Petkov <[email protected]> Acked-by: Thomas Gleixner <[email protected]> Cc: Aravind Gopalakrishnan <[email protected]> Cc: Linus Torvalds <[email protected]> Cc: Peter Zijlstra <[email protected]> Fixes: 3849e91 ("x86/AMD: Fix last level cache topology for AMD Fam17h systems") Link: http://lkml.kernel.org/r/[email protected] Signed-off-by: Ingo Molnar <[email protected]> Signed-off-by: Greg Kroah-Hartman <[email protected]>
1 parent cf5ae29 commit f8c74cf

File tree

1 file changed

+1
-5
lines changed

1 file changed

+1
-5
lines changed

arch/x86/kernel/cpu/amd.c

Lines changed: 1 addition & 5 deletions
Original file line numberDiff line numberDiff line change
@@ -347,7 +347,6 @@ static void amd_detect_cmp(struct cpuinfo_x86 *c)
347347
#ifdef CONFIG_SMP
348348
unsigned bits;
349349
int cpu = smp_processor_id();
350-
unsigned int socket_id, core_complex_id;
351350

352351
bits = c->x86_coreid_bits;
353352
/* Low order bits define the core id (index of core in socket) */
@@ -365,10 +364,7 @@ static void amd_detect_cmp(struct cpuinfo_x86 *c)
365364
if (c->x86 != 0x17 || !cpuid_edx(0x80000006))
366365
return;
367366

368-
socket_id = (c->apicid >> bits) - 1;
369-
core_complex_id = (c->apicid & ((1 << bits) - 1)) >> 3;
370-
371-
per_cpu(cpu_llc_id, cpu) = (socket_id << 3) | core_complex_id;
367+
per_cpu(cpu_llc_id, cpu) = c->apicid >> 3;
372368
#endif
373369
}
374370

0 commit comments

Comments
 (0)