EDAC, sb_edac: Remove double buffering of error records
In the bad old days the functions from x86_mce_decoder_chain could be
called in machine check context. So we used to carefully copy them and
defer processing until later. But in
f29a7aff4bd60 ("x86/mce: Avoid potential deadlock due to printk() in MCE context")
we switched the logging code to save the record in a genpool, and call
the functions that registered to be notified later from a work queue.
So drop all the double buffering and do all the work we want to do as
soon as sbridge_mce_check_error() is called.
Signed-off-by: Tony Luck <tony.luck@intel.com>
Cc: Aristeu Rozanski <arozansk@redhat.com>
Cc: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Cc: linux-edac <linux-edac@vger.kernel.org>
Cc: patrickg@supermicro.com
Link: http://lkml.kernel.org/r/100025611cd780d9bca72792b2b2146760da53e0.1460756761.git.tony.luck@intel.com
Signed-off-by: Borislav Petkov <bp@suse.de>
diff --git a/drivers/edac/sb_edac.c b/drivers/edac/sb_edac.c
index 93f0d41..3421674 100644
--- a/drivers/edac/sb_edac.c
+++ b/drivers/edac/sb_edac.c
@@ -363,16 +363,6 @@
/* Memory type detection */
bool is_mirrored, is_lockstep, is_close_pg;
- /* Fifo double buffers */
- struct mce mce_entry[MCE_LOG_LEN];
- struct mce mce_outentry[MCE_LOG_LEN];
-
- /* Fifo in/out counters */
- unsigned mce_in, mce_out;
-
- /* Count indicator to show errors not got */
- unsigned mce_overrun;
-
/* Memory description */
u64 tolm, tohm;
struct knl_pvt knl;
@@ -3075,63 +3065,8 @@
}
/*
- * sbridge_check_error Retrieve and process errors reported by the
- * hardware. Called by the Core module.
- */
-static void sbridge_check_error(struct mem_ctl_info *mci)
-{
- struct sbridge_pvt *pvt = mci->pvt_info;
- int i;
- unsigned count = 0;
- struct mce *m;
-
- /*
- * MCE first step: Copy all mce errors into a temporary buffer
- * We use a double buffering here, to reduce the risk of
- * loosing an error.
- */
- smp_rmb();
- count = (pvt->mce_out + MCE_LOG_LEN - pvt->mce_in)
- % MCE_LOG_LEN;
- if (!count)
- return;
-
- m = pvt->mce_outentry;
- if (pvt->mce_in + count > MCE_LOG_LEN) {
- unsigned l = MCE_LOG_LEN - pvt->mce_in;
-
- memcpy(m, &pvt->mce_entry[pvt->mce_in], sizeof(*m) * l);
- smp_wmb();
- pvt->mce_in = 0;
- count -= l;
- m += l;
- }
- memcpy(m, &pvt->mce_entry[pvt->mce_in], sizeof(*m) * count);
- smp_wmb();
- pvt->mce_in += count;
-
- smp_rmb();
- if (pvt->mce_overrun) {
- sbridge_printk(KERN_ERR, "Lost %d memory errors\n",
- pvt->mce_overrun);
- smp_wmb();
- pvt->mce_overrun = 0;
- }
-
- /*
- * MCE second step: parse errors and display
- */
- for (i = 0; i < count; i++)
- sbridge_mce_output_error(mci, &pvt->mce_outentry[i]);
-}
-
-/*
- * sbridge_mce_check_error Replicates mcelog routine to get errors
- * This routine simply queues mcelog errors, and
- * return. The error itself should be handled later
- * by sbridge_check_error.
- * WARNING: As this routine should be called at NMI time, extra care should
- * be taken to avoid deadlocks, and to be as fast as possible.
+ * Check that logging is enabled and that this is the right type
+ * of error for us to handle.
*/
static int sbridge_mce_check_error(struct notifier_block *nb, unsigned long val,
void *data)
@@ -3176,21 +3111,7 @@
"%u APIC %x\n", mce->cpuvendor, mce->cpuid,
mce->time, mce->socketid, mce->apicid);
- smp_rmb();
- if ((pvt->mce_out + 1) % MCE_LOG_LEN == pvt->mce_in) {
- smp_wmb();
- pvt->mce_overrun++;
- return NOTIFY_DONE;
- }
-
- /* Copy memory error at the ringbuffer */
- memcpy(&pvt->mce_entry[pvt->mce_out], mce, sizeof(*mce));
- smp_wmb();
- pvt->mce_out = (pvt->mce_out + 1) % MCE_LOG_LEN;
-
- /* Handle fatal errors immediately */
- if (mce->mcgstatus & 1)
- sbridge_check_error(mci);
+ sbridge_mce_output_error(mci, mce);
/* Advice mcelog that the error were handled */
return NOTIFY_STOP;
@@ -3276,9 +3197,6 @@
mci->dev_name = pci_name(pdev);
mci->ctl_page_to_phys = NULL;
- /* Set the function pointer to an actual operation function */
- mci->edac_check = sbridge_check_error;
-
pvt->info.type = type;
switch (type) {
case IVY_BRIDGE: