The patch titled
kdump: mpt fusion driver initialization failure fix
has been removed from the -mm tree. Its filename is
mpt-fusion-driver-initialization-failure-fix.patch
This patch was dropped because it was merged into mainline or a subsystem tree
------------------------------------------------------
Subject: kdump: mpt fusion driver initialization failure fix
From: Vivek Goyal <vgoyal@xxxxxxxxxx>
MPT fusion driver initialization fails while second kernel is booting,
after a system crash (if kdump kernel is configured). Oops message is
pasted below. I have attached a patch to fix the issue. Your suggestions
are welcome.
*****************************************************************************
Fusion MPT base driver 3.03.08
Copyright (c) 1999-2005 LSI Logic Corporation
Fusion MPT SAS Host driver 3.03.08 ACPI: PCI Interrupt 0000:01:00.0[A] -> Link [LNKA] -> GSI 5 (level, low) -> IRQ 5
mptbase: Initiating ioc0 bringup
BUG: unable to handle kernel paging request at virtual address 00002608
printing eip:
c11782fd
*pde = 00000000
Oops: 0000 [#1]
Modules linked in:
CPU: 0
EIP: 0060:[<c11782fd>] Not tainted VLI
EFLAGS: 00010046 (2.6.17-rc1-16M #2)
EIP is at mptscsih_io_done+0x27/0x3a3
eax: c4fed000 ebx: c4fed000 ecx: 00002600 edx: 00000298
esi: c11782d6 edi: 00002600 ebp: 00000000 esp: c1332f74
ds: 007b es: 007b ss: 0068
Process swapper (pid: 0, threadinfo=c1332000 task=c128f9c0) Stack: <0>0000006c 00000020 00000298 00002600 c4fed000 c4fed000 c11782d6 0000260 0
00000000 c1172c49 c4fed000 c1305b40 00000005 00000000 c1172d75 c48877e0
c1029687 00000000 c1307fb8 00000000 c1305a00 00000001 00000000 c1307fb8
Call Trace:
<c11782d6> mptscsih_io_done+0x0/0x3a3 <c1172c49> mpt_turbo_reply+0xbb/0xd3
<c1172d75> mpt_interrupt+0x22/0x2b <c1029687> misrouted_irq+0x63/0xcb
<c10297b3> note_interrupt+0x43/0x98 <c10292f9> __do_IRQ+0x68/0x8f
<c1003fac> do_IRQ+0x36/0x4e
=======================
<c1002aa6> common_interrupt+0x1a/0x20 <c1001150> mwait_idle+0x1a/0x2a
<c10010bf> cpu_idle+0x40/0x5c <c1308610> start_kernel+0x17a/0x17c Code: 5e 5f 5d c3 55 89 cd 57 56 53 83 ec 14 89 54 24 0c 89 44 24 10 8b 90 cc 00 00 00 8b 4c 24 0c 81 c2 98 02 00 00 85 ed 89 54 24 08 <0f> b7 79 08 89 fe 74 04 0f b7 75 08 66 39 f7 75 0d 8b 44 24 0c
*******************************************************************************
o Kdump capture kernel boot fails during initialization of MPT fusion driver.
(LSI Logic / Symbios Logic SAS1064E PCI-Express Fusion-MPT SAS (rev 01))
o Problem is easily reproducible, if system crashed while some disk activity
like cp operation was going on.
o After a system crash, devices are not shutdown and capture kenrel starts
booting while skipping BIOS. Hence underlying device is left in operational
state. In this case scsi contoller was left with interrupt line asserted
reply FIFO was not empty. When driver starts initializing in the second
kernel, it receives the interrupt the moment request_irq() is called.
Interrupt handler, reads the message from reply FIFO and tries to access
the associated message frame and panics, as in the new kernel's context
that message frame is not valid at all.
o In this scenario, probably we should delay the request_irq() call. First
bring up the IOC, reset it if needed and then should register for irq.
o I have tested the patch with SAS1064E and 53c1030 controllers.
Signed-off-by: Vivek Goyal <vgoyal@xxxxxxxxxx>
Acked-by: "Moore, Eric Dean" <Eric.Moore@xxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxx>
---
drivers/message/fusion/mptbase.c | 88 ++++++++++++++++-------------
1 file changed, 50 insertions(+), 38 deletions(-)
diff -puN drivers/message/fusion/mptbase.c~mpt-fusion-driver-initialization-failure-fix drivers/message/fusion/mptbase.c
--- devel/drivers/message/fusion/mptbase.c~mpt-fusion-driver-initialization-failure-fix 2006-06-01 20:41:51.000000000 -0700
+++ devel-akpm/drivers/message/fusion/mptbase.c 2006-06-01 20:41:51.000000000 -0700
@@ -1386,39 +1386,6 @@ mpt_attach(struct pci_dev *pdev, const s
/* Set lookup ptr. */
list_add_tail(&ioc->list, &ioc_list);
- ioc->pci_irq = -1;
- if (pdev->irq) {
- if (mpt_msi_enable && !pci_enable_msi(pdev))
- printk(MYIOC_s_INFO_FMT "PCI-MSI enabled\n", ioc->name);
-
- r = request_irq(pdev->irq, mpt_interrupt, SA_SHIRQ, ioc->name, ioc);
-
- if (r < 0) {
-#ifndef __sparc__
- printk(MYIOC_s_ERR_FMT "Unable to allocate interrupt %d!\n",
- ioc->name, pdev->irq);
-#else
- printk(MYIOC_s_ERR_FMT "Unable to allocate interrupt %s!\n",
- ioc->name, __irq_itoa(pdev->irq));
-#endif
- list_del(&ioc->list);
- iounmap(mem);
- kfree(ioc);
- return -EBUSY;
- }
-
- ioc->pci_irq = pdev->irq;
-
- pci_set_master(pdev); /* ?? */
- pci_set_drvdata(pdev, ioc);
-
-#ifndef __sparc__
- dprintk((KERN_INFO MYNAM ": %s installed at interrupt %d\n", ioc->name, pdev->irq));
-#else
- dprintk((KERN_INFO MYNAM ": %s installed at interrupt %s\n", ioc->name, __irq_itoa(pdev->irq)));
-#endif
- }
-
/* Check for "bound ports" (929, 929X, 1030, 1035) to reduce redundant resets.
*/
mpt_detect_bound_ports(ioc, pdev);
@@ -1428,11 +1395,7 @@ mpt_attach(struct pci_dev *pdev, const s
printk(KERN_WARNING MYNAM
": WARNING - %s did not initialize properly! (%d)\n",
ioc->name, r);
-
list_del(&ioc->list);
- free_irq(ioc->pci_irq, ioc);
- if (mpt_msi_enable)
- pci_disable_msi(pdev);
if (ioc->alt_ioc)
ioc->alt_ioc->alt_ioc = NULL;
iounmap(mem);
@@ -1651,6 +1614,7 @@ mpt_do_ioc_recovery(MPT_ADAPTER *ioc, u3
int handlers;
int ret = 0;
int reset_alt_ioc_active = 0;
+ int irq_allocated = 0;
printk(KERN_INFO MYNAM ": Initiating %s %s\n",
ioc->name, reason==MPT_HOSTEVENT_IOC_BRINGUP ? "bringup" : "recovery");
@@ -1734,6 +1698,48 @@ mpt_do_ioc_recovery(MPT_ADAPTER *ioc, u3
}
}
+ /*
+ * Device is reset now. It must have de-asserted the interrupt line
+ * (if it was asserted) and it should be safe to register for the
+ * interrupt now.
+ */
+ if ((ret == 0) && (reason == MPT_HOSTEVENT_IOC_BRINGUP)) {
+ ioc->pci_irq = -1;
+ if (ioc->pcidev->irq) {
+ if (mpt_msi_enable && !pci_enable_msi(ioc->pcidev))
+ printk(MYIOC_s_INFO_FMT "PCI-MSI enabled\n",
+ ioc->name);
+ rc = request_irq(ioc->pcidev->irq, mpt_interrupt,
+ SA_SHIRQ, ioc->name, ioc);
+ if (rc < 0) {
+#ifndef __sparc__
+ printk(MYIOC_s_ERR_FMT "Unable to allocate "
+ "interrupt %d!\n", ioc->name,
+ ioc->pcidev->irq);
+#else
+ printk(MYIOC_s_ERR_FMT "Unable to allocate "
+ "interrupt %s!\n", ioc->name,
+ __irq_itoa(ioc->pcidev->irq));
+#endif
+ if (mpt_msi_enable)
+ pci_disable_msi(ioc->pcidev);
+ return -EBUSY;
+ }
+ irq_allocated = 1;
+ ioc->pci_irq = ioc->pcidev->irq;
+ pci_set_master(ioc->pcidev); /* ?? */
+ pci_set_drvdata(ioc->pcidev, ioc);
+#ifndef __sparc__
+ dprintk((KERN_INFO MYNAM ": %s installed at interrupt "
+ "%d\n", ioc->name, ioc->pcidev->irq));
+#else
+ dprintk((KERN_INFO MYNAM ": %s installed at interrupt "
+ "%s\n", ioc->name,
+ __irq_itoa(ioc->pcidev->irq)));
+#endif
+ }
+ }
+
/* Prime reply & request queues!
* (mucho alloc's) Must be done prior to
* init as upper addresses are needed for init.
@@ -1833,7 +1839,7 @@ mpt_do_ioc_recovery(MPT_ADAPTER *ioc, u3
ret = mptbase_sas_persist_operation(ioc,
MPI_SAS_OP_CLEAR_NOT_PRESENT);
if(ret != 0)
- return -1;
+ goto out;
}
/* Find IM volumes
@@ -1914,6 +1920,12 @@ mpt_do_ioc_recovery(MPT_ADAPTER *ioc, u3
/* FIXME? Examine results here? */
}
+out:
+ if ((ret != 0) && irq_allocated) {
+ free_irq(ioc->pci_irq, ioc);
+ if (mpt_msi_enable)
+ pci_disable_msi(ioc->pcidev);
+ }
return ret;
}
_
Patches currently in -mm which might be from vgoyal@xxxxxxxxxx are
git-infiniband.patch
git-mtd.patch
register-hot-added-memory-to-iomem-resource.patch
updated-kdump-documentation.patch
-
To unsubscribe from this list: send the line "unsubscribe mm-commits" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
