| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-Id: <1286276048-20556-7-git-send-email-bp@amd64.org>
Date: Tue, 5 Oct 2010 12:54:06 +0200
From: Borislav Petkov <bp@...64.org>
To: <norsk5@...oo.com>
Cc: <linux-edac@...r.kernel.org>, <x86@...nel.org>,
<linux-kernel@...r.kernel.org>,
Borislav Petkov <borislav.petkov@....com>
Subject: [PATCH 6/8] EDAC, MCE: Complete NB MCE decoders
From: Borislav Petkov <borislav.petkov@....com>
Add support for decoding F14h BU MCEs and improve decoding of the
remaining families.
Signed-off-by: Borislav Petkov <borislav.petkov@....com>
---
drivers/edac/amd64_edac.h | 1 -
drivers/edac/mce_amd.c | 210 +++++++++++++++++++++++++++++++++-----------
drivers/edac/mce_amd.h | 3 +-
3 files changed, 158 insertions(+), 56 deletions(-)
diff --git a/drivers/edac/amd64_edac.h b/drivers/edac/amd64_edac.h
index 13e1d6f..044aee4 100644
--- a/drivers/edac/amd64_edac.h
+++ b/drivers/edac/amd64_edac.h
@@ -482,7 +482,6 @@ extern const char *rrrr_msgs[16];
extern const char *to_msgs[2];
extern const char *pp_msgs[4];
extern const char *ii_msgs[4];
-extern const char *ext_msgs[32];
extern const char *htlink_msgs[8];
#ifdef CONFIG_EDAC_DEBUG
diff --git a/drivers/edac/mce_amd.c b/drivers/edac/mce_amd.c
index adb5d25..cf6d44a 100644
--- a/drivers/edac/mce_amd.c
+++ b/drivers/edac/mce_amd.c
@@ -5,6 +5,8 @@
static struct amd_decoder_ops *fam_ops;
+static u8 nb_err_cpumask = 0xf;
+
static bool report_gart_errors;
static void (*nb_bus_decoder)(int node_id, struct mce *m, u32 nbcfg);
@@ -61,45 +63,16 @@ EXPORT_SYMBOL_GPL(to_msgs);
const char *ii_msgs[] = { "MEM", "RESV", "IO", "GEN" };
EXPORT_SYMBOL_GPL(ii_msgs);
-/*
- * Map the 4 or 5 (family-specific) bits of Extended Error code to the
- * string table.
- */
-const char *ext_msgs[] = {
- "K8 ECC error", /* 0_0000b */
- "CRC error on link", /* 0_0001b */
- "Sync error packets on link", /* 0_0010b */
- "Master Abort during link operation", /* 0_0011b */
- "Target Abort during link operation", /* 0_0100b */
- "Invalid GART PTE entry during table walk", /* 0_0101b */
- "Unsupported atomic RMW command received", /* 0_0110b */
- "WDT error: NB transaction timeout", /* 0_0111b */
- "ECC/ChipKill ECC error", /* 0_1000b */
- "SVM DEV Error", /* 0_1001b */
- "Link Data error", /* 0_1010b */
- "Link/L3/Probe Filter Protocol error", /* 0_1011b */
- "NB Internal Arrays Parity error", /* 0_1100b */
- "DRAM Address/Control Parity error", /* 0_1101b */
- "Link Transmission error", /* 0_1110b */
- "GART/DEV Table Walk Data error" /* 0_1111b */
- "Res 0x100 error", /* 1_0000b */
- "Res 0x101 error", /* 1_0001b */
- "Res 0x102 error", /* 1_0010b */
- "Res 0x103 error", /* 1_0011b */
- "Res 0x104 error", /* 1_0100b */
- "Res 0x105 error", /* 1_0101b */
- "Res 0x106 error", /* 1_0110b */
- "Res 0x107 error", /* 1_0111b */
- "Res 0x108 error", /* 1_1000b */
- "Res 0x109 error", /* 1_1001b */
- "Res 0x10A error", /* 1_1010b */
- "Res 0x10B error", /* 1_1011b */
- "ECC error in L3 Cache Data", /* 1_1100b */
- "L3 Cache Tag error", /* 1_1101b */
- "L3 Cache LRU Parity error", /* 1_1110b */
- "Probe Filter error" /* 1_1111b */
+static const char *f10h_nb_mce_desc[] = {
+ "HT link data error",
+ "Protocol error (link, L3, probe filter, etc.)",
+ "Parity error in NB-internal arrays",
+ "Link Retry due to IO link transmission error",
+ "L3 ECC data cache error",
+ "ECC error in L3 cache tag",
+ "L3 LRU parity bits error",
+ "ECC Error in the Probe Filter directory"
};
-EXPORT_SYMBOL_GPL(ext_msgs);
static bool f10h_dc_mce(u16 ec)
{
@@ -370,19 +343,97 @@ wrong_ls_mce:
pr_emerg(HW_ERR "Corrupted LS MCE info?\n");
}
+static bool k8_nb_mce(u16 ec, u8 xec)
+{
+ bool ret = true;
+
+ switch (xec) {
+ case 0x1:
+ pr_cont("CRC error detected on HT link.\n");
+ break;
+
+ case 0x5:
+ pr_cont("Invalid GART PTE entry during GART table walk.\n");
+ break;
+
+ case 0x6:
+ pr_cont("Unsupported atomic RMW received from an IO link.\n");
+ break;
+
+ case 0x0:
+ case 0x8:
+ pr_cont("DRAM ECC error detected on the NB.\n");
+ break;
+
+ case 0xd:
+ pr_cont("Parity error on the DRAM addr/ctl signals.\n");
+ break;
+
+ default:
+ ret = false;
+ break;
+ }
+
+ return ret;
+}
+
+static bool f10h_nb_mce(u16 ec, u8 xec)
+{
+ bool ret = true;
+ u8 offset = 0;
+
+ if (k8_nb_mce(ec, xec))
+ return true;
+
+ switch(xec) {
+ case 0xa ... 0xc:
+ offset = 10;
+ break;
+
+ case 0xe:
+ offset = 11;
+ break;
+
+ case 0xf:
+ if (TLB_ERROR(ec))
+ pr_cont("GART Table Walk data error.\n");
+ else if (BUS_ERROR(ec))
+ pr_cont("DMA Exclusion Vector Table Walk error.\n");
+ else
+ ret = false;
+
+ goto out;
+ break;
+
+ case 0x1c ... 0x1f:
+ offset = 24;
+ break;
+
+ default:
+ ret = false;
+
+ goto out;
+ break;
+ }
+
+ pr_cont("%s.\n", f10h_nb_mce_desc[xec - offset]);
+
+out:
+ return ret;
+}
+
+static bool f14h_nb_mce(u16 ec, u8 xec)
+{
+ return false;
+}
+
void amd_decode_nb_mce(int node_id, struct mce *m, u32 nbcfg)
{
- u32 ec = m->status & 0xffff;
+ u8 xec = (m->status >> 16) & 0x1f;
+ u16 ec = m->status & 0xffff;
u32 nbsh = (u32)(m->status >> 32);
- u32 nbsl = (u32)m->status;
-
- /*
- * GART TLB error reporting is disabled by default. Bail out early.
- */
- if (TLB_ERROR(ec) && !report_gart_errors)
- return;
- pr_emerg(HW_ERR "Northbridge Error, node %d", node_id);
+ pr_emerg(HW_ERR "Northbridge Error, node %d: ", node_id);
/*
* F10h, revD can disable ErrCpu[3:0] so check that first and also the
@@ -391,20 +442,50 @@ void amd_decode_nb_mce(int node_id, struct mce *m, u32 nbcfg)
if ((boot_cpu_data.x86 == 0x10) &&
(boot_cpu_data.x86_model > 7)) {
if (nbsh & K8_NBSH_ERR_CPU_VAL)
- pr_cont(", core: %u\n", (u8)(nbsh & 0xf));
+ pr_cont(", core: %u", (u8)(nbsh & nb_err_cpumask));
} else {
- u8 assoc_cpus = nbsh & 0xf;
+ u8 assoc_cpus = nbsh & nb_err_cpumask;
if (assoc_cpus > 0)
pr_cont(", core: %d", fls(assoc_cpus) - 1);
+ }
- pr_cont("\n");
+ switch (xec) {
+ case 0x2:
+ pr_cont("Sync error (sync packets on HT link detected).\n");
+ return;
+
+ case 0x3:
+ pr_cont("HT Master abort.\n");
+ return;
+
+ case 0x4:
+ pr_cont("HT Target abort.\n");
+ return;
+
+ case 0x7:
+ pr_cont("NB Watchdog timeout.\n");
+ return;
+
+ case 0x9:
+ pr_cont("SVM DMA Exclusion Vector error.\n");
+ return;
+
+ default:
+ break;
}
- pr_emerg(HW_ERR "%s.\n", EXT_ERR_MSG(nbsl));
+ if (!fam_ops->nb_mce(ec, xec))
+ goto wrong_nb_mce;
+
+ if (boot_cpu_data.x86 == 0xf || boot_cpu_data.x86 == 0x10)
+ if ((xec == 0x8 || xec == 0x0) && nb_bus_decoder)
+ nb_bus_decoder(node_id, m, nbcfg);
- if (BUS_ERROR(ec) && nb_bus_decoder)
- nb_bus_decoder(node_id, m, nbcfg);
+ return;
+
+wrong_nb_mce:
+ pr_emerg(HW_ERR "Corrupted NB MCE info?\n");
}
EXPORT_SYMBOL_GPL(amd_decode_nb_mce);
@@ -434,11 +515,30 @@ static inline void amd_decode_err_code(u16 ec)
pr_emerg(HW_ERR "Huh? Unknown MCE error 0x%x\n", ec);
}
+/*
+ * Filter out unwanted MCE signatures here.
+ */
+static bool amd_filter_mce(struct mce *m)
+{
+ u8 xec = (m->status >> 16) & 0x1f;
+
+ /*
+ * NB GART TLB error reporting is disabled by default.
+ */
+ if (m->bank == 4 && xec == 0x5 && !report_gart_errors)
+ return true;
+
+ return false;
+}
+
int amd_decode_mce(struct notifier_block *nb, unsigned long val, void *data)
{
struct mce *m = (struct mce *)data;
int node, ecc;
+ if (amd_filter_mce(m))
+ return NOTIFY_STOP;
+
pr_emerg(HW_ERR "MC%d_STATUS: ", m->bank);
pr_cont("%sorrected error, other errors lost: %s, "
@@ -513,16 +613,20 @@ static int __init mce_amd_init(void)
case 0xf:
fam_ops->dc_mce = k8_dc_mce;
fam_ops->ic_mce = k8_ic_mce;
+ fam_ops->nb_mce = k8_nb_mce;
break;
case 0x10:
fam_ops->dc_mce = f10h_dc_mce;
fam_ops->ic_mce = k8_ic_mce;
+ fam_ops->nb_mce = f10h_nb_mce;
break;
case 0x14:
+ nb_err_cpumask = 0x3;
fam_ops->dc_mce = f14h_dc_mce;
fam_ops->ic_mce = f14h_ic_mce;
+ fam_ops->nb_mce = f14h_nb_mce;
break;
default:
diff --git a/drivers/edac/mce_amd.h b/drivers/edac/mce_amd.h
index dc81dba..0d0637d 100644
--- a/drivers/edac/mce_amd.h
+++ b/drivers/edac/mce_amd.h
@@ -7,7 +7,6 @@
#define ERROR_CODE(x) ((x) & 0xffff)
#define EXT_ERROR_CODE(x) (((x) >> 16) & 0x1f)
-#define EXT_ERR_MSG(x) ext_msgs[EXT_ERROR_CODE(x)]
#define LOW_SYNDROME(x) (((x) >> 15) & 0xff)
#define HIGH_SYNDROME(x) (((x) >> 24) & 0xff)
@@ -83,7 +82,6 @@ extern const char *rrrr_msgs[];
extern const char *pp_msgs[];
extern const char *to_msgs[];
extern const char *ii_msgs[];
-extern const char *ext_msgs[];
/*
* relevant NB regs
@@ -102,6 +100,7 @@ struct err_regs {
struct amd_decoder_ops {
bool (*dc_mce)(u16);
bool (*ic_mce)(u16);
+ bool (*nb_mce)(u16, u8);
};
void amd_report_gart_errors(bool);
--
1.7.3.1.50.g1e633
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists