[PATCHv2] x86/mce: Avoid reading every machine check bank register twice.

[PATCHv2] x86/mce: Avoid reading every machine check bank register twice.

[Luck, Tony]

Reading machine check bank registers is slow. There is a trend of
increasing the number of banks, and the number of cores. The main section
of do_machine_check() is a serialized section where each cpu in turn
checks every bank. Even on a little two socket SandyBridge-EP system
that multiplies out as:

	2 sockets * 8 cores * 2 hyperthreads * 20 banks = 640 MSRs

We already scan the banks in parallel in mce_no_way_out() to see if there
is a fatal error anywhere in the system. If we build a cache of VALID
bits during this scan, we can avoid uselessly re-reading banks that have
no data. Note that this cache is only a hint. If the valid bit is set in a
shared bank, all cpus that share that bank will see it during the parallel
scan, but the first to find it in the sequential scan will (usually) clear
the bank.

Signed-off-by: Tony Luck <tony.luck@intel.com>
---

Version 2:
	+ mce_no_way_out() now scans all banks to build the full bitmap (instead of
	  breaking out early if it saw a fatal issue).
	+ changed name of bitmap from "hint" to "valid_banks"

 arch/x86/kernel/cpu/mcheck/mce.c |   16 +++++++++++-----
 1 files changed, 11 insertions(+), 5 deletions(-)

diff --git a/arch/x86/kernel/cpu/mcheck/mce.c b/arch/x86/kernel/cpu/mcheck/mce.c
index d086a09..66e1c51 100644
--- a/arch/x86/kernel/cpu/mcheck/mce.c
+++ b/arch/x86/kernel/cpu/mcheck/mce.c
@@ -641,16 +641,18 @@ EXPORT_SYMBOL_GPL(machine_check_poll);
  * Do a quick check if any of the events requires a panic.
  * This decides if we keep the events around or clear them.
  */
-static int mce_no_way_out(struct mce *m, char **msg)
+static int mce_no_way_out(struct mce *m, char **msg, unsigned long *validp)
 {
-	int i;
+	int i, ret = 0;
 
 	for (i = 0; i < banks; i++) {
 		m->status = mce_rdmsrl(MSR_IA32_MCx_STATUS(i));
+		if (m->status & MCI_STATUS_VAL)
+			__set_bit(i, validp);
 		if (mce_severity(m, tolerant, msg) >= MCE_PANIC_SEVERITY)
-			return 1;
+			ret = 1;
 	}
-	return 0;
+	return ret;
 }
 
 /*
@@ -1011,6 +1013,7 @@ void do_machine_check(struct pt_regs *regs, long error_code)
 	 */
 	int kill_it = 0;
 	DECLARE_BITMAP(toclear, MAX_NR_BANKS);
+	DECLARE_BITMAP(valid_banks, MAX_NR_BANKS);
 	char *msg = "Unknown";
 
 	atomic_inc(&mce_entry);
@@ -1025,7 +1028,8 @@ void do_machine_check(struct pt_regs *regs, long error_code)
 	final = &__get_cpu_var(mces_seen);
 	*final = m;
 
-	no_way_out = mce_no_way_out(&m, &msg);
+	memset(valid_banks, 0, sizeof(valid_banks));
+	no_way_out = mce_no_way_out(&m, &msg, valid_banks);
 
 	barrier();
 
@@ -1045,6 +1049,8 @@ void do_machine_check(struct pt_regs *regs, long error_code)
 	order = mce_start(&no_way_out);
 	for (i = 0; i < banks; i++) {
 		__clear_bit(i, toclear);
+		if (!test_bit(i, valid_banks))
+			continue;
 		if (!mce_banks[i].ctl)
 			continue;
 
-- 
1.7.9.rc2.1.g69204

Re: [PATCHv2] x86/mce: Avoid reading every machine check bank register twice.

[Borislav Petkov]

On Wed, Apr 18, 2012 at 03:19:40PM -0700, Luck, Tony wrote:
> Reading machine check bank registers is slow. There is a trend of
> increasing the number of banks, and the number of cores. The main section
> of do_machine_check() is a serialized section where each cpu in turn
> checks every bank. Even on a little two socket SandyBridge-EP system
> that multiplies out as:
> 
> 	2 sockets * 8 cores * 2 hyperthreads * 20 banks = 640 MSRs
> 
> We already scan the banks in parallel in mce_no_way_out() to see if there
> is a fatal error anywhere in the system. If we build a cache of VALID
> bits during this scan, we can avoid uselessly re-reading banks that have
> no data. Note that this cache is only a hint. If the valid bit is set in a
> shared bank, all cpus that share that bank will see it during the parallel
> scan, but the first to find it in the sequential scan will (usually) clear
> the bank.
> 
> Signed-off-by: Tony Luck <tony.luck@intel.com>
> ---
> 
> Version 2:
> 	+ mce_no_way_out() now scans all banks to build the full bitmap (instead of
> 	  breaking out early if it saw a fatal issue).
> 	+ changed name of bitmap from "hint" to "valid_banks"

Looks good.

Acked-by: Borislav Petkov <borislav.petkov@amd.com>

Thanks.

-- 
Regards/Gruss,
Boris.

Advanced Micro Devices GmbH
Einsteinring 24, 85609 Dornach
GM: Alberto Bozzo
Reg: Dornach, Landkreis Muenchen
HRB Nr. 43632 WEEE Registernr: 129 19551