Bug#305330: starfire NIC driver on alpha
Steve,
Yes, the bug is also present in 2.6.18. Using the current kernel in
Debian testing, I get the following error:
This driver has dma_addr_t issues, please send email to maintainer
The fix is to apply the following one line change to starfire.c:
--- starfire.orig.c 2007-03-11 22:54:36.000000000 -0700
+++ starfire.c 2007-03-11 22:54:04.000000000 -0700
@@ -151,7 +151,7 @@
* This SUCKS.
* We need a much better method to determine if dma_addr_t is 64-bit.
*/
-#if (defined(__i386__) && defined(CONFIG_HIGHMEM64G)) || defined
(__x86_64__) || defined (__ia64__) || defined(__mips64__) || (defined
(__mips__) && defined(CONFIG_HIGHMEM) && defined
(CONFIG_64BIT_PHYS_ADDR))
+#if (defined(__i386__) && defined(CONFIG_HIGHMEM64G)) || defined
(__x86_64__) || defined (__ia64__) || defined(__mips64__) || defined
(__alpha__) || (defined(__mips__) && defined(CONFIG_HIGHMEM) &&
defined(CONFIG_64BIT_PHYS_ADDR))
/* 64-bit dma_addr_t */
#define ADDR_64BITS /* This chip uses 64 bit addresses. */
#define netdrv_addr_t u64
I have been running the starfire module with this fix on one of my
alphas using 2.6.18 for a few days now. Seems to be working just
fine. I have been running the same fix in the Debian stable 2.6.8
kernel for a long time now.
Thanks,
Bob
On Mar 10, 2007, at 1:33 AM, Steve Langasek wrote:
Hi Bob,
On Tue, Feb 27, 2007 at 04:47:39PM -0800, Bob Lindell wrote:
I use Adaptec quad NICs in my 3 alphas. I submitted a bug about 2
years ago with a simple patch to the starfire driver to make it work
on the alpha. Basically, a "|| defined (__alpha__)" needs to be
added to an expression in the driver code. The submiited bug is
#305330. Could this get into etch? Without it, I don't have a
network when I do an install and I have to recompile the starfire
driver and update it before I can get the machine back on the network
and complete any network based package installation.
Is it certain that this bug affects 2.6.18? Have you tested your
patch
against the current etch kernel, to be sure that it's still
applicable?
Certainly, the kernel message you quoted is no longer present in
the source.
Norbert said he was going to contact the upstream maintainer of the
driver,
but there's no information about what came of that. Committing the
patch to
the Debian tree is easy enough, but I'm hesitant to do that without
more
information.
--
Steve Langasek Give me a lever long enough and a
Free OS
Debian Developer to set it on, and I can move the
world.
vorlon@debian.org http://
www.debian.org/
Reply to: