cvs commit: src/sys/dev/bce if_bce.c src/sys/dev/em if_em.c if_em.h src/sys/dev/mpt mpt.h mpt_pci.c

John Baldwin jhb at freebsd.org
Mon Jan 8 19:20:46 UTC 2007


On Saturday 18 November 2006 10:07, Pawel Worach wrote:
> Pawel Worach wrote:
> >> So em triggered an interrupt and you got an NMI out of it instead.  What
> >> does 'show lapic' show?
> >>
> > 
> > db> show lapic
> > lapic ID = 0
> > version  = 1.4
> > max LVT  = 5
> > SVR      = ff (enabled)
> > TPR      = 00
> > In-service Interrupts:
> > TMR Interrupts:
> > IRR Interrupts:
> > 
> 
> I reverted MSI support from em(4) just to see what would happen to 
> mpt(4), no NMI fires but mpt hardware seems to go off to la-la land.

Can you verify that the following disables MSI for your system:

Index: pci.c
===================================================================
RCS file: /usr/cvs/src/sys/dev/pci/pci.c,v
retrieving revision 1.331
diff -u -r1.331 pci.c
--- pci.c	28 Dec 2006 06:14:42 -0000	1.331
+++ pci.c	8 Jan 2007 19:04:50 -0000
@@ -178,6 +178,13 @@
 	{ 0x02001166, PCI_QUIRK_MAP_REG,	0x90,	 0 },
 
 	/*
+	 * MSI doesn't work with the ServerWorks CNB20-HE Host Bridge
+	 * or the CMIC-SL (AKA ServerWorks GC_LE).
+	 */
+	{ 0x00141106, PCI_QUIRK_DISABLE_MSI,	0,	0 },
+	{ 0x00171106, PCI_QUIRK_DISABLE_MSI,	0,	0 },
+	
+	/*
 	 * MSI doesn't work with the Intel E7501 chipset, at least on
 	 * the Tyan 2721 motherboard.
 	 */

-- 
John Baldwin


More information about the cvs-src mailing list