Sophie

Sophie

distrib > Scientific%20Linux > 5x > x86_64 > by-pkgid > fc11cd6e1c513a17304da94a5390f3cd > files > 2940

kernel-2.6.18-194.11.1.el5.src.rpm

From: AMEET M. PARANJAPE <aparanja@redhat.com>
Date: Wed, 22 Oct 2008 13:29:24 -0500
Subject: [ppc64] EEH PCI-E: recovery fails E1000; support MSI
Message-id: 48FF7104.8040109@REDHAT.COM
O-Subject: Re: [PATCH 1/2 RHEL5.3 bz445299] EEH recovery fails E1000 PCI-E; support for EEH on PCI-E using MSI
Bugzilla: 445299
RH-Acked-by: David Howells <dhowells@redhat.com>
RH-Acked-by: David Howells <dhowells@redhat.com>
RH-Acked-by: Pete Zaitcev <zaitcev@redhat.com>
RH-Acked-by: David Howells <dhowells@redhat.com>

RHBZ#:
======
https://bugzilla.redhat.com/show_bug.cgi?id=445299

Description:
===========
EEH recovery fails after injects EEH error to 4 ports E1000 PCI-E card(coleman)
on Power 6 machine

The following is the shown message after EEH injection.

[root@devl4e-shoal-lp1 ~]# errinjct ioa-bus-error -f 17 -s net/eth3 -m 0
Injecting an ioa-bus-error...
Call to RTAS errinjct succeeded!

The device then fails to recover, as can be seen here:
# Call Trace:
[C000000002A6FAC0] [C00000000000FFDC] .show_stack+0x68/0x1b0 (unreliable)
[C000000002A6FB60] [C00000000004EFA4] .eeh_dn_check_failure+0x25c/0x29c
[C000000002A6FC10] [C00000000004F1F0] .eeh_check_failure+0xdc/0x104
[C000000002A6FC90] [D00000000037ABF0] .e1000_watchdog_task+0xb8/0xb00 [e1000]
[C000000002A6FD50] [C00000000007E66C] .run_workqueue+0xdc/0x168
[C000000002A6FDF0] [C00000000007F3D8] .worker_thread+0x128/0x198
[C000000002A6FEE0] [C000000000083EB8] .kthread+0x128/0x178
[C000000002A6FF90] [C000000000026A14] .kernel_thread+0x4c/0x68
EEH: This PCI device has failed 1 times since last reboot:
location=U789D.001.DQDMLZZ-P1-C3 driver=e1000 pci addr=0002:03:00.1
e1000: eth3: e1000_reset: Hardware Error
Badness in disable_msi_mode at arch/powerpc/kernel/msi.c:749

This problem prevents e1000e devices using MSI on powerpc from automatically
recovering after a PCI error is detected.  Automatic I/O error recovery is a
vital feature in customer environments where availability is critical.

RHEL Version Found:
================
RHEL 5.1

kABI Status:
============
Will test once brew is back up

Brew:
=====
Will test once brew is back up

Upstream Status:
================
The commit to mainline appears here:
http://git.kernel.org/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=b56a5a23bfecd9cac9187164a9d5f22d287c48b9.

 I kept these in the core eeh code to avoid touching the pci subsystem directly.
The commit to export pci_restore_msi_state appears here:
http://git.kernel.org/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=94688cf2454986309fbcd495233ba2423786a14a
The commit to increase EEH_MAX_FAILS appears here:
http://git.kernel.org/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=2fd30be8dae25386fc5167c34c6d73201334a8d4
A commit similar to what's posted here for e1000e is here:
http://git.kernel.org/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=aad32739641d3a75818fbe653d4b0d530e965f2f

Test Status:
============
Was thoroughly tested on the RH 2.6.18-88 kernel in 5.2, but awaiting brew to
confirm on 5.3

diff --git a/arch/powerpc/kernel/msi-rtas.c b/arch/powerpc/kernel/msi-rtas.c
index e2d952b..6229b0e 100644
--- a/arch/powerpc/kernel/msi-rtas.c
+++ b/arch/powerpc/kernel/msi-rtas.c
@@ -165,6 +165,7 @@ static int rtas_setup_msi_irqs(struct pci_dev *pdev, int nvec, int type)
 	struct pci_dn *pdn;
 	int hwirq, virq, i, rc;
 	struct msi_desc *entry;
+	struct msi_msg msg;
 
 	pdn = get_pdn(pdev);
 	if (!pdn)
@@ -214,6 +215,11 @@ static int rtas_setup_msi_irqs(struct pci_dev *pdev, int nvec, int type)
 
 		dev_dbg(&pdev->dev, "rtas_msi: allocated virq %d\n", virq);
 		set_irq_msi(virq, entry);
+
+		/* Read config space back so we can restore after reset*/
+		read_msi_msg(virq, &msg);
+		entry->msg = msg;
+
 		unmask_msi_irq(virq);
 	}
 
diff --git a/arch/powerpc/kernel/msi.c b/arch/powerpc/kernel/msi.c
index 4e5a3e9..e543d3b 100644
--- a/arch/powerpc/kernel/msi.c
+++ b/arch/powerpc/kernel/msi.c
@@ -262,7 +262,6 @@ static struct msi_desc* alloc_msi_entry(void)
 	return entry;
 }
 
-#ifdef CONFIG_PM
 int pci_save_msi_state(struct pci_dev *dev)
 {
 	return 0;
@@ -325,7 +324,7 @@ void pci_restore_msix_state(struct pci_dev *dev)
 	control |= PCI_MSIX_FLAGS_ENABLE;
 	pci_write_config_word(dev, pos + PCI_MSIX_FLAGS, control);
 }
-#endif	/* CONFIG_PM */
+EXPORT_SYMBOL_GPL(pci_restore_msi_state);
 
 /**
  * msi_capability_init - configure device's MSI capability structure
diff --git a/arch/powerpc/platforms/pseries/eeh.c b/arch/powerpc/platforms/pseries/eeh.c
index 34b87a6..778d2b5 100644
--- a/arch/powerpc/platforms/pseries/eeh.c
+++ b/arch/powerpc/platforms/pseries/eeh.c
@@ -128,6 +128,59 @@ static unsigned long slot_resets;
 /* --------------------------------------------------------------- */
 /* Below lies the EEH event infrastructure */
 
+int save_pcie_reg(struct pci_dev *dev)
+{
+	int pos, i = 0;
+	struct pci_cap_saved_state *save_state;
+	u16 *cap;
+
+	pos = pci_find_capability(dev, PCI_CAP_ID_EXP);
+	if (pos <= 0)
+		return 0;
+
+	save_state = pci_find_saved_cap(dev, PCI_CAP_ID_EXP);
+	if (save_state)
+		return 0;
+
+	save_state = kzalloc(sizeof(*save_state) + sizeof(u16) * 4, GFP_KERNEL);
+	if (!save_state) {
+		dev_err(&dev->dev, "Out of memory in save_pcie_reg\n");
+		return -ENOMEM;
+	}
+
+	cap = (u16 *)&save_state->data[0];
+
+	pci_read_config_word(dev, pos + PCI_EXP_DEVCTL, &cap[i++]);
+	pci_read_config_word(dev, pos + PCI_EXP_LNKCTL, &cap[i++]);
+	pci_read_config_word(dev, pos + PCI_EXP_SLTCTL, &cap[i++]);
+	pci_read_config_word(dev, pos + PCI_EXP_RTCTL, &cap[i++]);
+	save_state->cap_nr = PCI_CAP_ID_EXP;
+	pci_add_saved_cap(dev, save_state);
+
+	return 0;
+}
+
+void restore_pcie_reg(struct pci_dev *dev)
+{
+	int i = 0, pos;
+	struct pci_cap_saved_state *save_state;
+	u16 *cap;
+
+	save_state = pci_find_saved_cap(dev, PCI_CAP_ID_EXP);
+	pos = pci_find_capability(dev, PCI_CAP_ID_EXP);
+	if (!save_state || pos <= 0)
+		return;
+	cap = (u16 *)&save_state->data[0];
+
+	pci_write_config_word(dev, pos + PCI_EXP_DEVCTL, cap[i++]);
+	pci_write_config_word(dev, pos + PCI_EXP_LNKCTL, cap[i++]);
+	pci_write_config_word(dev, pos + PCI_EXP_SLTCTL, cap[i++]);
+	pci_write_config_word(dev, pos + PCI_EXP_RTCTL, cap[i++]);
+}
+
+EXPORT_SYMBOL_GPL(save_pcie_reg);
+EXPORT_SYMBOL_GPL(restore_pcie_reg);
+
 static void rtas_slot_error_detail(struct pci_dn *pdn, int severity,
                                    char *driver_log, size_t loglen)
 {
diff --git a/drivers/pci/msi.c b/drivers/pci/msi.c
index 2a63a74..96bec55 100644
--- a/drivers/pci/msi.c
+++ b/drivers/pci/msi.c
@@ -574,6 +574,7 @@ void pci_restore_msi_state(struct pci_dev *dev)
 	pci_remove_saved_cap(save_state);
 	kfree(save_state);
 }
+EXPORT_SYMBOL_GPL(pci_restore_msi_state);
 
 int pci_save_msix_state(struct pci_dev *dev)
 {
diff --git a/drivers/pci/pci.h b/drivers/pci/pci.h
index 9550591..61cb569 100644
--- a/drivers/pci/pci.h
+++ b/drivers/pci/pci.h
@@ -41,26 +41,22 @@ extern void pci_remove_legacy_files(struct pci_bus *bus);
 
 /* Lock for read/write access to pci device and bus lists */
 extern struct rw_semaphore pci_bus_sem;
-
 extern unsigned int pci_pm_d3_delay;
+
 #ifdef CONFIG_PCI_MSI
 void disable_msi_mode(struct pci_dev *dev, int pos, int type);
 void pci_no_msi(void);
-#else
-static inline void disable_msi_mode(struct pci_dev *dev, int pos, int type) { }
-static inline void pci_no_msi(void) { }
-#endif
-#if defined(CONFIG_PCI_MSI) && defined(CONFIG_PM)
 int pci_save_msi_state(struct pci_dev *dev);
 int pci_save_msix_state(struct pci_dev *dev);
-void pci_restore_msi_state(struct pci_dev *dev);
 void pci_restore_msix_state(struct pci_dev *dev);
 #else
+static inline void disable_msi_mode(struct pci_dev *dev, int pos, int type) { }
+static inline void pci_no_msi(void) { }
 static inline int pci_save_msi_state(struct pci_dev *dev) { return 0; }
 static inline int pci_save_msix_state(struct pci_dev *dev) { return 0; }
-static inline void pci_restore_msi_state(struct pci_dev *dev) {}
 static inline void pci_restore_msix_state(struct pci_dev *dev) {}
 #endif
+
 static inline int pci_no_d1d2(struct pci_dev *dev)
 {
 	unsigned int parent_dstates = 0;
diff --git a/include/linux/pci.h b/include/linux/pci.h
index 68ac37b..42d636a 100644
--- a/include/linux/pci.h
+++ b/include/linux/pci.h
@@ -657,6 +657,7 @@ static inline int pci_enable_msix(struct pci_dev* dev,
 	struct msix_entry *entries, int nvec) {return -1;}
 static inline void pci_disable_msix(struct pci_dev *dev) {}
 static inline void msi_remove_pci_irq_vectors(struct pci_dev *dev) {}
+static inline void pci_restore_msi_state(struct pci_dev *dev) {}
 #else
 extern void pci_scan_msi_device(struct pci_dev *dev);
 extern int pci_enable_msi(struct pci_dev *dev);
@@ -665,6 +666,15 @@ extern int pci_enable_msix(struct pci_dev* dev,
 	struct msix_entry *entries, int nvec);
 extern void pci_disable_msix(struct pci_dev *dev);
 extern void msi_remove_pci_irq_vectors(struct pci_dev *dev);
+extern void pci_restore_msi_state(struct pci_dev *dev);
+#endif
+
+#ifndef CONFIG_EEH
+static inline int save_pcie_reg(struct pci_dev *dev) {return -1;}
+static inline void restore_pcie_reg(struct pci_dev *dev) {}
+#else
+extern int save_pcie_reg(struct pci_dev *dev);
+extern void restore_pcie_reg(struct pci_dev *dev);
 #endif
 
 extern void pci_block_user_cfg_access(struct pci_dev *dev);

Perl :: Catalyst :: PostgreSQL :: RPM Valid HTML 4.01 Transitional