Has anyone actually tried to find which patch breaks the keyboard? I'd
like to get this keyboard thing fixed in g-d-s, and since I'm not
affected, I can't test it...
Also, what are the specific problems preventing us from going up to
2.6.8 (and even 2.6.9) over 2.6.7? The main issues that I know of which
are not resolved in 2.6.7-gentoo-r17 are:
* SMP "broken" - for all 2.6.x
* Hypersparc DMA broken (patch on sparclinux) - for all 2.6.x
* hme doesn't work on Hypersparc - 2.6.9
* SMP with high load causes lockup - for all 2.6.x
* high load / disk access causes (rare) lockup on UP and SMP (reported
for a US II on 2.6.0-test10 and 2.6.9) - AFAIK, for all 2.6.x
* stop-a doesn't work over serial console (patch on sparclinux, but I
don't use serial console, so I haven't tested it) - for all 2.6.x
* keyboard sometimes doesn't work - 2.6.7-gentoo-rx
No issues that I see were added with 2.6.8, and additionally, I've been
running g-d-s-2.6.8-r10 (with the sparc patch set forward ported from
2.6.7-r17) without a hickup for the past day, and it's been constantly
running 'dbench 32' for the past 90 minutes... I have a UP UltraSparc
IIi... so would someone like to fill me in? Is it a sparc32 specific
reason that I missed?
On Fri, 2004-11-26 at 16:11 -0800, Jeremy Huddleston wrote:
> On Fri, 2004-11-26 at 23:45 +0000, Ciaran McCreesh wrote:
> > d-s-2.6.7 works. g-d-s-2.6.7 doesn't, hence why I moan about the gentoo
> > patchset so much :) Unfortunately I don't have access to the box that I
> > could trigger that on any more...
> Ok, that's actually good news... So if someone can trigger the keyboard
> bug in the g-d-s-2.6.7-r15 kernel, then try adding this to the
> gentoo-dev-sources-2.6.7-r15 ebuild and re-emerging/compiling the
> kernel... This will exclude all the non-sparc patches...
> UNIPATCH_EXCLUDE="1100_ip_tables.patch 1105_CAN-2004-0497.patch
> 1110_proc.patch 1115_net_eql.patch 1120-CAN-2004-0415.patch
> 1125-proc_pid_cmdline-race-fix.patch 1130_nfs-fix.patch
> 1135_netfilter-dos-fix.patch 1140_CAN-2004-0814.patch
> 1145_smbfs-dos.patch 1310_k8_cardbus_io.patch
> 1315_alpha-sysctl-uac.patch 1320_x86_64-2.6.7-2.patch
> 1321_x86_64-noiommu.patch 1325_iptables-headers.patch
> 1900_ulogd_timestamp.patch 2115_fa311-mac-address-fix.patch
> 2700_ppc-pegasos-2.6.6.patch 2905_mouse_click_fix.patch
> 2910_wacom_fix.patch 4505_bootsplash-3.1.4-sp-0.1-2.6.7.patch
> 4705_squashfs-2.0.patch 4715_supermount-2.0.4-2.6.7.patch
> Test that out, slowly remove them until you find the culprit...
> I am suspicious of 2905_mouse_click_fix.patch as it's the only one that
> touches the input code, so try re-adding that one first to see if it
> rebreaks the keyboard.
Jeremy Huddleston <email@example.com>