mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Mauro Carvalho Chehab <mchehab@redhat.com>
To: unlisted-recipients:; (no To-header on input)
Cc: Mauro Carvalho Chehab <mchehab@redhat.com>,
	Linux Edac Mailing List <linux-edac@vger.kernel.org>,
	Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
	Greg K H <gregkh@linuxfoundation.org>
Subject: [PATCH 14/14] edac: Only expose csrows/channels on legacy API if they're populated
Date: Thu, 29 Mar 2012 14:07:01 -0300	[thread overview]
Message-ID: <1333040821-6253-15-git-send-email-mchehab@redhat.com> (raw)
In-Reply-To: <1333040821-6253-1-git-send-email-mchehab@redhat.com>

This patch actually fixes a bug with the legacy API, where, at the
same csrow, some channels may have different DIMMs. This can happen
on FB-DIMM/RAMBUS and modern Intel controllers.

This is the case, for example, of Nehalem machines:

$ ./edac-ctl --layout
       +-----------------------------------+
       |                mc0                |
       | channel0  | channel1  | channel2  |
-------+-----------------------------------+
slot2: |     0 MB  |     0 MB  |     0 MB  |
slot1: |  1024 MB  |     0 MB  |     0 MB  |
slot0: |  1024 MB  |  1024 MB  |  1024 MB  |
-------+-----------------------------------+

Before this patch, non-filled memories were shown. Now, only what's
filled is there:

grep . /sys/devices/system/edac/mc/mc0/csrow*/ch?*
/sys/devices/system/edac/mc/mc0/csrow0/ch0_ce_count:0
/sys/devices/system/edac/mc/mc0/csrow0/ch0_dimm_label:CPU#0Channel#0_DIMM#0
/sys/devices/system/edac/mc/mc0/csrow0/ch1_ce_count:0
/sys/devices/system/edac/mc/mc0/csrow0/ch1_dimm_label:CPU#0Channel#0_DIMM#1
/sys/devices/system/edac/mc/mc0/csrow1/ch0_ce_count:0
/sys/devices/system/edac/mc/mc0/csrow1/ch0_dimm_label:CPU#0Channel#1_DIMM#0
/sys/devices/system/edac/mc/mc0/csrow2/ch0_ce_count:0
/sys/devices/system/edac/mc/mc0/csrow2/ch0_dimm_label:CPU#0Channel#2_DIMM#0

Thanks-to: Aristeu Rozanski Filho <arozansk@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
---
 drivers/edac/edac_mc_sysfs.c |   24 ++++++++++++++++++++++++
 1 files changed, 24 insertions(+), 0 deletions(-)

diff --git a/drivers/edac/edac_mc_sysfs.c b/drivers/edac/edac_mc_sysfs.c
index 319612c..112fff2 100644
--- a/drivers/edac/edac_mc_sysfs.c
+++ b/drivers/edac/edac_mc_sysfs.c
@@ -347,6 +347,16 @@ static struct device_attribute *dynamic_csrow_ce_count_attr[] = {
 	&dev_attr_legacy_ch5_ce_count.attr
 };
 
+static inline int nr_pages_per_csrow(struct csrow_info *csrow)
+{
+	int chan, nr_pages = 0;
+
+	for (chan = 0; chan < csrow->nr_channels; chan++)
+		nr_pages += csrow->channels[chan].dimm->nr_pages;
+
+	return nr_pages;
+}
+
 /* Create a CSROW object under specifed edac_mc_device */
 static int edac_create_csrow_object(struct mem_ctl_info *mci,
 				    struct csrow_info *csrow, int index)
@@ -371,6 +381,9 @@ static int edac_create_csrow_object(struct mem_ctl_info *mci,
 		return err;
 
 	for (chan = 0; chan < csrow->nr_channels; chan++) {
+		/* Only expose populated DIMMs */
+		if (!csrow->channels[chan].dimm->nr_pages)
+			continue;
 		err = device_create_file(&csrow->dev,
 					 dynamic_csrow_dimm_attr[chan]);
 		if (err < 0)
@@ -405,6 +418,9 @@ static int edac_create_csrow_objects(struct mem_ctl_info *mci)
 	struct csrow_info *csrow;
 
 	for (i = 0; i < mci->num_csrows; i++) {
+		csrow = &mci->csrows[i];
+		if (!nr_pages_per_csrow(csrow))
+			continue;
 		err = edac_create_csrow_object(mci, &mci->csrows[i], i);
 		if (err < 0)
 			goto error;
@@ -414,7 +430,11 @@ static int edac_create_csrow_objects(struct mem_ctl_info *mci)
 error:
 	for (--i; i >= 0; i--) {
 		csrow = &mci->csrows[i];
+		if (!nr_pages_per_csrow(csrow))
+			continue;
 		for (chan = csrow->nr_channels - 1; chan >= 0; chan--) {
+			if (!csrow->channels[chan].dimm->nr_pages)
+				continue;
 			device_remove_file(&csrow->dev,
 						dynamic_csrow_dimm_attr[chan]);
 			device_remove_file(&csrow->dev,
@@ -433,7 +453,11 @@ static void edac_delete_csrow_objects(struct mem_ctl_info *mci)
 
 	for (i = mci->num_csrows - 1; i >= 0; i--) {
 		csrow = &mci->csrows[i];
+		if (!nr_pages_per_csrow(csrow))
+			continue;
 		for (chan = csrow->nr_channels - 1; chan >= 0; chan--) {
+			if (!csrow->channels[chan].dimm->nr_pages)
+				continue;
 			debugf1("Removing csrow %d channel %d sysfs nodes\n",
 				i, chan);
 			device_remove_file(&csrow->dev,
-- 
1.7.8


  parent reply	other threads:[~2012-03-29 17:08 UTC|newest]

Thread overview: 26+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-03-29 17:06 [PATCH 00/14] Fix the EDAC API Mauro Carvalho Chehab
2012-03-29 17:06 ` [PATCH 01/14] edac: rewrite the sysfs code to use struct device Mauro Carvalho Chehab
2012-03-29 22:03   ` Greg K H
2012-03-29 23:19     ` Mauro Carvalho Chehab
2012-03-29 23:40       ` Greg K H
2012-03-30  2:13         ` Mauro Carvalho Chehab
2012-03-30  9:11           ` Borislav Petkov
2012-03-30 10:46             ` Mauro Carvalho Chehab
2012-03-30 15:32               ` Greg K H
2012-03-30 15:31           ` Greg K H
2012-03-30 17:34             ` Mauro Carvalho Chehab
2012-03-30 19:24               ` Greg K H
2012-03-29 17:06 ` [PATCH 02/14] mpc85xx_edac: convert sysfs logic " Mauro Carvalho Chehab
2012-03-29 17:06 ` [PATCH 03/14] amd64_edac: " Mauro Carvalho Chehab
2012-03-29 17:06 ` [PATCH 04/14] i7core_edac: convert it " Mauro Carvalho Chehab
2012-03-29 17:06 ` [PATCH 05/14] edac: Get rid of the old kobj's from the edac mc code Mauro Carvalho Chehab
2012-03-29 17:06 ` [PATCH 06/14] edac: add a new per-dimm API and make the old per-virtual-rank API obsolete Mauro Carvalho Chehab
2012-03-29 17:06 ` [PATCH 07/14] edac: add a sysfs node to report the maximum location for the system Mauro Carvalho Chehab
2012-03-29 17:06 ` [PATCH 08/14] edac: Add debufs nodes to allow doing fake error inject Mauro Carvalho Chehab
2012-03-29 17:06 ` [PATCH 09/14] edac: Create a per-Memory Controller bus Mauro Carvalho Chehab
2012-03-29 17:06 ` [PATCH 10/14] edac: Move grain/dtype/edac_type calculus to be out of channel loop Mauro Carvalho Chehab
2012-03-29 17:06 ` [PATCH 11/14] i82975x_edac: Test nr_pages earlier to save a few CPU cycles Mauro Carvalho Chehab
2012-03-29 17:06 ` [PATCH 12/14] i5100_edac: Fix a warning when compiled with 32 bits Mauro Carvalho Chehab
2012-03-29 17:07 ` [PATCH 13/14] i7300_edac: Get rid of some wrongly-solved rebase conflict Mauro Carvalho Chehab
2012-03-29 17:07 ` Mauro Carvalho Chehab [this message]
2012-03-29 20:46 ` [PATCH 00/14] Fix the EDAC API Aristeu Rozanski Filho

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1333040821-6253-15-git-send-email-mchehab@redhat.com \
    --to=mchehab@redhat.com \
    --cc=gregkh@linuxfoundation.org \
    --cc=linux-edac@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®