linux/arch
Kim Phillips 7ec11cad3f perf/x86/amd/ibs: Fix sample bias for dispatched micro-ops
[ Upstream commit 0f4cd769c4 ]

When counting dispatched micro-ops with cnt_ctl=1, in order to prevent
sample bias, IBS hardware preloads the least significant 7 bits of
current count (IbsOpCurCnt) with random values, such that, after the
interrupt is handled and counting resumes, the next sample taken
will be slightly perturbed.

The current count bitfield is in the IBS execution control h/w register,
alongside the maximum count field.

Currently, the IBS driver writes that register with the maximum count,
leaving zeroes to fill the current count field, thereby overwriting
the random bits the hardware preloaded for itself.

Fix the driver to actually retain and carry those random bits from the
read of the IBS control register, through to its write, instead of
overwriting the lower current count bits with zeroes.

Tested with:

perf record -c 100001 -e ibs_op/cnt_ctl=1/pp -a -C 0 taskset -c 0 <workload>

'perf annotate' output before:

 15.70  65:   addsd     %xmm0,%xmm1
 17.30        add       $0x1,%rax
 15.88        cmp       %rdx,%rax
              je        82
 17.32  72:   test      $0x1,%al
              jne       7c
  7.52        movapd    %xmm1,%xmm0
  5.90        jmp       65
  8.23  7c:   sqrtsd    %xmm1,%xmm0
 12.15        jmp       65

'perf annotate' output after:

 16.63  65:   addsd     %xmm0,%xmm1
 16.82        add       $0x1,%rax
 16.81        cmp       %rdx,%rax
              je        82
 16.69  72:   test      $0x1,%al
              jne       7c
  8.30        movapd    %xmm1,%xmm0
  8.13        jmp       65
  8.24  7c:   sqrtsd    %xmm1,%xmm0
  8.39        jmp       65

Tested on Family 15h and 17h machines.

Machines prior to family 10h Rev. C don't have the RDWROPCNT capability,
and have the IbsOpCurCnt bitfield reserved, so this patch shouldn't
affect their operation.

It is unknown why commit db98c5faf8 ("perf/x86: Implement 64-bit
counter support for IBS") ignored the lower 4 bits of the IbsOpCurCnt
field; the number of preloaded random bits has always been 7, AFAICT.

Signed-off-by: Kim Phillips <kim.phillips@amd.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: "Arnaldo Carvalho de Melo" <acme@kernel.org>
Cc: <x86@kernel.org>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "Borislav Petkov" <bp@alien8.de>
Cc: Stephane Eranian <eranian@google.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: "Namhyung Kim" <namhyung@kernel.org>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Link: https://lkml.kernel.org/r/20190826195730.30614-1-kim.phillips@amd.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
2019-09-21 07:17:07 +02:00
..
alpha alpha: Fix Eiger NR_IRQS to 128 2019-02-20 10:25:47 +01:00
arc ARC: mm: SIGSEGV userspace trying to access kernel virtual memory 2019-09-16 08:22:14 +02:00
arm ARM: 8901/1: add a criteria for pfn_valid of arm 2019-09-21 07:17:05 +02:00
arm64 arm64: dts: stratix10: add the sysmgr-syscon property from the gmac's 2019-09-16 08:22:09 +02:00
c6x
h8300 h8300: use cc-cross-prefix instead of hardcoding h8300-unknown-linux- 2019-04-05 22:32:55 +02:00
hexagon
ia64 ia64: fix build errors by exporting paddr_to_nid() 2019-06-22 08:15:19 +02:00
m68k m68k: Add -ffreestanding to CFLAGS 2019-03-23 20:10:00 +01:00
microblaze mm: make the __PAGETABLE_PxD_FOLDED defines non-empty 2018-12-29 13:37:57 +01:00
mips mips: fix cacheinfo 2019-08-29 08:28:27 +02:00
nds32 nds32: Fix gcc 8.0 compiler option incompatible. 2019-02-12 19:46:57 +01:00
nios2
openrisc
parisc parisc: Fix build of compressed kernel even with debug enabled 2019-08-06 19:06:55 +02:00
powerpc powerpc/mm/radix: Use the right page size for vmemmap mapping 2019-09-21 07:16:42 +02:00
riscv riscv: remove unused variable in ftrace 2019-09-16 08:21:44 +02:00
s390 s390/bpf: use 32-bit index for tail calls 2019-09-21 07:16:51 +02:00
sh sh: kernel: hw_breakpoint: Fix missing break in switch statement 2019-08-25 10:47:42 +02:00
sparc sparc: perf: fix updated event period in response to PERF_EVENT_IOC_PERIOD 2019-06-25 11:35:57 +08:00
um um: Silence lockdep complaint about mmap_sem 2019-07-31 07:27:04 +02:00
unicore32
x86 perf/x86/amd/ibs: Fix sample bias for dispatched micro-ops 2019-09-21 07:17:07 +02:00
xtensa xtensa: add missing isync to the cpu_reset TLB code 2019-08-25 10:47:47 +02:00
.gitignore
Kconfig jump_label: move 'asm goto' support test to Kconfig 2019-06-04 08:02:34 +02:00