* [Linux-ia64] Bug in sba_iommu.c (2.5.64)
@ 2003-04-01 4:39 Peter Chubb
0 siblings, 0 replies; only message in thread
From: Peter Chubb @ 2003-04-01 4:39 UTC (permalink / raw)
To: linux-ia64
OK, thanks to some inspired debugging work by Matthew Chapman, we now
have the error. sba_map_sg() was passing the uninitialised dma
address from the sg-list to sba_map_single() instead of the virtual
address. This meant that the dma when it happened scribbled over
random memory causing an MCA later, or possibly caused an MCA immediately.
Patch against 2.5.64 appended.
=== linux-userdrivers/arch/ia64/hp/common/sba_iommu.c 1.10 vs edited ==--- 1.10/arch/ia64/hp/common/sba_iommu.c Wed Feb 5 12:06:01 2003
+++ edited/arch/ia64/hp/common/sba_iommu.c Tue Apr 1 14:31:50 2003
@@ -1119,7 +1119,7 @@
/* Fast path single entry scatterlists. */
if (nents = 1) {
sba_sg_iova(sglist) = sba_map_single(dev,
- (void *) sba_sg_iova(sglist),
+ (void *)sba_sg_address(sglist),
sba_sg_len(sglist), direction);
sba_sg_iova_len(sglist) = sba_sg_len(sglist);
#ifdef CONFIG_PROC_FS
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2003-04-01 4:39 UTC | newest]
Thread overview: (only message) (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2003-04-01 4:39 [Linux-ia64] Bug in sba_iommu.c (2.5.64) Peter Chubb
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox