Re: [PATCH] perf, amd, ibs: Update IBS MSRs and feature definitions

From: Aravind Gopalakrishnan
Date: Thu Nov 06 2014 - 11:57:55 EST


On 11/6/2014 10:34 AM, Borislav Petkov wrote:
On Thu, Nov 06, 2014 at 10:26:22AM -0600, Aravind Gopalakrishnan wrote:
diff --git a/arch/x86/include/uapi/asm/msr-index.h b/arch/x86/include/uapi/asm/msr-index.h
index e21331c..ba7b609 100644
--- a/arch/x86/include/uapi/asm/msr-index.h
+++ b/arch/x86/include/uapi/asm/msr-index.h
@@ -206,6 +206,8 @@
#define MSR_AMD64_IBSOP_REG_MASK ((1UL<<MSR_AMD64_IBSOP_REG_COUNT)-1)
#define MSR_AMD64_IBSCTL 0xc001103a
#define MSR_AMD64_IBSBRTARGET 0xc001103b
+#define MSR_AMD64_IBS_FETCH_EXTD_CTL 0xc001103c
+#define MSR_AMD64_IBSOPDATA4 0xc001103d
#define MSR_AMD64_IBS_REG_COUNT_MAX 8 /* includes MSR_AMD64_IBSBRTARGET */
/* Fam 16h MSRs */
diff --git a/arch/x86/kernel/cpu/perf_event_amd_ibs.c b/arch/x86/kernel/cpu/perf_event_amd_ibs.c
index cbb1be3e..a61f5c6 100644
--- a/arch/x86/kernel/cpu/perf_event_amd_ibs.c
+++ b/arch/x86/kernel/cpu/perf_event_amd_ibs.c
@@ -565,6 +565,21 @@ static int perf_ibs_handle_irq(struct perf_ibs *perf_ibs, struct pt_regs *iregs)
perf_ibs->offset_max,
offset + 1);
} while (offset < offset_max);
+ if (event->attr.sample_type & PERF_SAMPLE_RAW) {
+ /*
+ * Read IbsBrTarget and IbsOpData4 separately
+ * depending on their availability.
+ * Can't add to offset_max as they are staggered
+ */
+ if (ibs_caps & IBS_CAPS_BRNTRGT) {
+ rdmsrl(MSR_AMD64_IBSBRTARGET, *buf++);
Are those MSRs present on everything that supports IBS and if not, you
probably should do rdmsr_safe() here instead and handle the error case.
Or do something with f/m/s checking or so...

But Why?
IBS_CAPS_BRNTRGT and IBS_CAPS_OPDATA4 indicate support for the respective MSRs and
I am only loading the MSR contents upon checking for their availability. So it's not like an exception is
generated for a rdmsr command on an unimplemented/reserved MSR.

And the nice thing about the feature identifiers is that I don't have to do f/m/s checks right?
I mean, if some other future processor decides to implement it, then we don't have to revisit the code
to make a change to the f/m/s condition.
And if they don't want to use those MSRs then it's still OK as the feature bits are not going to be set..


Thanks,
-Aravind.

+ size++;
+ }
+ if (ibs_caps & IBS_CAPS_OPDATA4) {
+ rdmsrl(MSR_AMD64_IBSOPDATA4, *buf++);
+ size++;
+ }
+ }
ibs_data.size = sizeof(u64) * size;
regs = *iregs;
--
1.9.1


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/