Fix i82975x_edac OOPS (BZ#848149)

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
This commit is contained in:
Mauro Carvalho Chehab 2012-10-15 21:55:32 -03:00
parent e2d59620f5
commit 1d13150d5e
2 changed files with 241 additions and 0 deletions

234
i82975x-edac-fix.patch Normal file
View File

@ -0,0 +1,234 @@
commit 9370a8d717720f6b17221490fea8d798396d9f2f
Author: Mauro Carvalho Chehab <mchehab@redhat.com>
Date: Mon Oct 15 21:49:35 2012 -0300
i82975x_edac: Use the edac standard debug macro
Instead of declaring its own debug macro, that requires
to uncomment part of the code, use the edac standard macro
to add the debug code, and the edac debug level to print it,
just like any other EDAC driver.
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
diff --git a/drivers/edac/i82975x_edac.c b/drivers/edac/i82975x_edac.c
index a980204..f998d2c 100644
--- a/drivers/edac/i82975x_edac.c
+++ b/drivers/edac/i82975x_edac.c
@@ -435,11 +435,9 @@ static void i82975x_init_csrows(struct mem_ctl_info *mci,
}
}
-/* #define i82975x_DEBUG_IOMEM */
-
-#ifdef i82975x_DEBUG_IOMEM
-static void i82975x_print_dram_timings(void __iomem *mch_window)
+static void i82975x_print_dram_config(void __iomem *mch_window, u32 mchbar, u32 *drc)
{
+#ifdef CONFIG_EDAC_DEBUG
/*
* The register meanings are from Intel specs;
* (shows 13-5-5-5 for 800-DDR2)
@@ -448,26 +446,63 @@ static void i82975x_print_dram_timings(void __iomem *mch_window)
*/
static const int caslats[4] = { 5, 4, 3, 6 };
u32 dtreg[2];
+ u8 c0drb[4];
+ u8 c1drb[4];
+
+ if (!edac_debug_level)
+ return;
+
+ i82975x_printk(KERN_INFO, "MCHBAR real = %0x, remapped = %p\n",
+ mchbar, mch_window);
+
+ c0drb[0] = readb(mch_window + I82975X_DRB_CH0R0);
+ c0drb[1] = readb(mch_window + I82975X_DRB_CH0R1);
+ c0drb[2] = readb(mch_window + I82975X_DRB_CH0R2);
+ c0drb[3] = readb(mch_window + I82975X_DRB_CH0R3);
+ c1drb[0] = readb(mch_window + I82975X_DRB_CH1R0);
+ c1drb[1] = readb(mch_window + I82975X_DRB_CH1R1);
+ c1drb[2] = readb(mch_window + I82975X_DRB_CH1R2);
+ c1drb[3] = readb(mch_window + I82975X_DRB_CH1R3);
+ i82975x_printk(KERN_INFO, "DRBCH0R0 = 0x%02x\n", c0drb[0]);
+ i82975x_printk(KERN_INFO, "DRBCH0R1 = 0x%02x\n", c0drb[1]);
+ i82975x_printk(KERN_INFO, "DRBCH0R2 = 0x%02x\n", c0drb[2]);
+ i82975x_printk(KERN_INFO, "DRBCH0R3 = 0x%02x\n", c0drb[3]);
+ i82975x_printk(KERN_INFO, "DRBCH1R0 = 0x%02x\n", c1drb[0]);
+ i82975x_printk(KERN_INFO, "DRBCH1R1 = 0x%02x\n", c1drb[1]);
+ i82975x_printk(KERN_INFO, "DRBCH1R2 = 0x%02x\n", c1drb[2]);
+ i82975x_printk(KERN_INFO, "DRBCH1R3 = 0x%02x\n", c1drb[3]);
+
+ i82975x_printk(KERN_INFO, "DRC_CH0 = %0x, %s\n", drc[0],
+ ((drc[0] >> 21) & 3) == 1 ?
+ "ECC enabled" : "ECC disabled");
+ i82975x_printk(KERN_INFO, "DRC_CH1 = %0x, %s\n", drc[1],
+ ((drc[1] >> 21) & 3) == 1 ?
+ "ECC enabled" : "ECC disabled");
+
+ i82975x_printk(KERN_INFO, "C0 BNKARC = %0x\n",
+ readw(mch_window + I82975X_C0BNKARC));
+ i82975x_printk(KERN_INFO, "C1 BNKARC = %0x\n",
+ readw(mch_window + I82975X_C1BNKARC));
dtreg[0] = readl(mch_window + 0x114);
dtreg[1] = readl(mch_window + 0x194);
- i82975x_printk(KERN_INFO, "DRAM Timings : Ch0 Ch1\n"
+ i82975x_printk(KERN_INFO,
+ "DRAM Timings : Ch0 Ch1\n"
" RAS Active Min = %d %d\n"
" CAS latency = %d %d\n"
" RAS to CAS = %d %d\n"
" RAS precharge = %d %d\n",
(dtreg[0] >> 19 ) & 0x0f,
- (dtreg[1] >> 19) & 0x0f,
+ (dtreg[1] >> 19) & 0x0f,
caslats[(dtreg[0] >> 8) & 0x03],
- caslats[(dtreg[1] >> 8) & 0x03],
+ caslats[(dtreg[1] >> 8) & 0x03],
((dtreg[0] >> 4) & 0x07) + 2,
- ((dtreg[1] >> 4) & 0x07) + 2,
+ ((dtreg[1] >> 4) & 0x07) + 2,
(dtreg[0] & 0x07) + 2,
- (dtreg[1] & 0x07) + 2
+ (dtreg[1] & 0x07) + 2
);
-
-}
#endif
+}
static int i82975x_probe1(struct pci_dev *pdev, int dev_idx)
{
@@ -480,10 +515,6 @@ static int i82975x_probe1(struct pci_dev *pdev, int dev_idx)
u32 drc[2];
struct i82975x_error_info discard;
int chans;
-#ifdef i82975x_DEBUG_IOMEM
- u8 c0drb[4];
- u8 c1drb[4];
-#endif
edac_dbg(0, "\n");
@@ -495,45 +526,11 @@ static int i82975x_probe1(struct pci_dev *pdev, int dev_idx)
mchbar &= 0xffffc000; /* bits 31:14 used for 16K window */
mch_window = ioremap_nocache(mchbar, 0x1000);
-#ifdef i82975x_DEBUG_IOMEM
- i82975x_printk(KERN_INFO, "MCHBAR real = %0x, remapped = %p\n",
- mchbar, mch_window);
-
- c0drb[0] = readb(mch_window + I82975X_DRB_CH0R0);
- c0drb[1] = readb(mch_window + I82975X_DRB_CH0R1);
- c0drb[2] = readb(mch_window + I82975X_DRB_CH0R2);
- c0drb[3] = readb(mch_window + I82975X_DRB_CH0R3);
- c1drb[0] = readb(mch_window + I82975X_DRB_CH1R0);
- c1drb[1] = readb(mch_window + I82975X_DRB_CH1R1);
- c1drb[2] = readb(mch_window + I82975X_DRB_CH1R2);
- c1drb[3] = readb(mch_window + I82975X_DRB_CH1R3);
- i82975x_printk(KERN_INFO, "DRBCH0R0 = 0x%02x\n", c0drb[0]);
- i82975x_printk(KERN_INFO, "DRBCH0R1 = 0x%02x\n", c0drb[1]);
- i82975x_printk(KERN_INFO, "DRBCH0R2 = 0x%02x\n", c0drb[2]);
- i82975x_printk(KERN_INFO, "DRBCH0R3 = 0x%02x\n", c0drb[3]);
- i82975x_printk(KERN_INFO, "DRBCH1R0 = 0x%02x\n", c1drb[0]);
- i82975x_printk(KERN_INFO, "DRBCH1R1 = 0x%02x\n", c1drb[1]);
- i82975x_printk(KERN_INFO, "DRBCH1R2 = 0x%02x\n", c1drb[2]);
- i82975x_printk(KERN_INFO, "DRBCH1R3 = 0x%02x\n", c1drb[3]);
-#endif
-
drc[0] = readl(mch_window + I82975X_DRC_CH0M0);
drc[1] = readl(mch_window + I82975X_DRC_CH1M0);
-#ifdef i82975x_DEBUG_IOMEM
- i82975x_printk(KERN_INFO, "DRC_CH0 = %0x, %s\n", drc[0],
- ((drc[0] >> 21) & 3) == 1 ?
- "ECC enabled" : "ECC disabled");
- i82975x_printk(KERN_INFO, "DRC_CH1 = %0x, %s\n", drc[1],
- ((drc[1] >> 21) & 3) == 1 ?
- "ECC enabled" : "ECC disabled");
- i82975x_printk(KERN_INFO, "C0 BNKARC = %0x\n",
- readw(mch_window + I82975X_C0BNKARC));
- i82975x_printk(KERN_INFO, "C1 BNKARC = %0x\n",
- readw(mch_window + I82975X_C1BNKARC));
- i82975x_print_dram_timings(mch_window);
- goto fail1;
-#endif
+ i82975x_print_dram_config(mch_window, mchbar, drc);
+
if (!(((drc[0] >> 21) & 3) == 1 || ((drc[1] >> 21) & 3) == 1)) {
i82975x_printk(KERN_INFO, "ECC disabled on both channels.\n");
goto fail1;
commit 8992ed2f4295eab137e1713fa16be5546a759373
Author: Mauro Carvalho Chehab <mchehab@redhat.com>
Date: Mon Oct 15 21:48:48 2012 -0300
i82975x_edac: Fix dimm label initialization
The driver has only 4 hardcoded labels, but allows much more memory.
Fix it by removing the hardcoded logic, using snprintf() instead.
[ 19.833972] general protection fault: 0000 [#1] SMP
[ 19.837733] Modules linked in: i82975x_edac(+) edac_core firewire_ohci firewire_core crc_itu_t nouveau mxm_wmi wmi video i2c_algo_bit drm_kms_helper ttm drm i2c_core
[ 19.837733] CPU 0
[ 19.837733] Pid: 390, comm: udevd Not tainted 3.6.1-1.fc17.x86_64.debug #1 Dell Inc. Precision WorkStation 390 /0MY510
[ 19.837733] RIP: 0010:[<ffffffff813463a8>] [<ffffffff813463a8>] strncpy+0x18/0x30
[ 19.837733] RSP: 0018:ffff880078535b68 EFLAGS: 00010202
[ 19.837733] RAX: ffff880069fa9708 RBX: ffff880078588000 RCX: ffff880069fa9708
[ 19.837733] RDX: 000000000000001f RSI: 5f706f5f63616465 RDI: ffff880069fa9708
[ 19.837733] RBP: ffff880078535b68 R08: ffff880069fa9727 R09: 000000000000fffe
[ 19.837733] R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000000003
[ 19.837733] R13: 0000000000000000 R14: ffff880069fa9290 R15: ffff880079624a80
[ 19.837733] FS: 00007f3de01ee840(0000) GS:ffff88007c400000(0000) knlGS:0000000000000000
[ 19.837733] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 19.837733] CR2: 00007f3de00b9000 CR3: 0000000078dbc000 CR4: 00000000000007f0
[ 19.837733] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[ 19.837733] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
[ 19.837733] Process udevd (pid: 390, threadinfo ffff880078534000, task ffff880079642450)
[ 19.837733] Stack:
[ 19.837733] ffff880078535c18 ffffffffa017c6b8 00040000816d627f ffff880079624a88
[ 19.837733] ffffc90004cd6000 ffff880079624520 ffff88007ac21148 0000000000000000
[ 19.837733] 0000000000000000 0004000000000000 feda000078535bc8 ffffffff810d696d
[ 19.837733] Call Trace:
[ 19.837733] [<ffffffffa017c6b8>] i82975x_init_one+0x2e6/0x3e6 [i82975x_edac]
...
Fix bug reported at:
https://bugzilla.redhat.com/show_bug.cgi?id=848149
And, very likely:
https://bbs.archlinux.org/viewtopic.php?id=148033
https://bugzilla.kernel.org/show_bug.cgi?id=47171
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
diff --git a/drivers/edac/i82975x_edac.c b/drivers/edac/i82975x_edac.c
index 069e26c..a980204 100644
--- a/drivers/edac/i82975x_edac.c
+++ b/drivers/edac/i82975x_edac.c
@@ -370,10 +370,6 @@ static enum dev_type i82975x_dram_type(void __iomem *mch_window, int rank)
static void i82975x_init_csrows(struct mem_ctl_info *mci,
struct pci_dev *pdev, void __iomem *mch_window)
{
- static const char *labels[4] = {
- "DIMM A1", "DIMM A2",
- "DIMM B1", "DIMM B2"
- };
struct csrow_info *csrow;
unsigned long last_cumul_size;
u8 value;
@@ -423,9 +419,10 @@ static void i82975x_init_csrows(struct mem_ctl_info *mci,
dimm = mci->csrows[index]->channels[chan]->dimm;
dimm->nr_pages = nr_pages / csrow->nr_channels;
- strncpy(csrow->channels[chan]->dimm->label,
- labels[(index >> 1) + (chan * 2)],
- EDAC_MC_LABEL_LEN);
+
+ snprintf(csrow->channels[chan]->dimm->label, EDAC_MC_LABEL_LEN, "DIMM %c%d",
+ (chan == 0) ? 'A' : 'B',
+ index);
dimm->grain = 1 << 7; /* 128Byte cache-line resolution */
dimm->dtype = i82975x_dram_type(mch_window, index);
dimm->mtype = MEM_DDR2; /* I82975x supports only DDR2 */

View File

@ -708,6 +708,8 @@ Patch14010: lis3-improve-handling-of-null-rate.patch
Patch19000: ips-noirq.patch
Patch19001: i82975x-edac-fix.patch
# ARM
# Flattened devicetree support
Patch21000: arm-omap-dt-compat.patch
@ -1424,6 +1426,8 @@ ApplyPatch lis3-improve-handling-of-null-rate.patch
ApplyPatch ips-noirq.patch
ApplyPatch i82975x-edac-fix.patch
ApplyPatch power-x86-destdir.patch
#rhbz 754518
@ -2305,6 +2309,9 @@ fi
# '-' | |
# '-'
%changelog
* Mon Oct 15 2012 Mauro Carvalho Chehab <mchehab@redhat.com> - 3.6.2-3
- Fix i82975x_edac OOPS
* Sat Oct 13 2012 Josh Boyer <jwboyer@redhat.com> - 3.6.2-2
- Linux v3.6.2
- Fix whitespace corruption in linux-3.6-arm-build-fixup.patch