Summary: | madwifi-driver-0.1_pre20040726 does not compile with linux-2.6.8.1 | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Stephen Tallowitz <dev> |
Component: | Current packages | Assignee: | Mobile Herd (OBSOLETE) <mobile+disabled> |
Status: | VERIFIED TEST-REQUEST | ||
Severity: | normal | CC: | jan.callewaert, per, reillyp, solar |
Priority: | High | ||
Version: | unspecified | ||
Hardware: | x86 | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Attachments: |
proc_dointvec.patch
madwifi-driver-0.1_pre20040824.ebuild.patch |
Description
Stephen Tallowitz
2004-08-15 10:11:20 UTC
I have the exact same problems as you. madwifi-driver-0.1_pre20040726.ebuild compiled fine in 2.6.7 (official vanilla sources), but it fails with the error you mentioned above with 2.6.8.1 (official vanilla sources). I'm also having the identical problem with gentoo-dev-sources-2.6.8-r1 same problem with gentoo-dev-sources-2.6.8 here. Same problem here. I have found a patch on https://sourceforge.net/tracker/index.php?func=detail&aid=1008185&group_id=82936&atid=567755, but I don't know how to patch the ebuild. Created attachment 38225 [details, diff]
proc_dointvec.patch
Cleaned up the patch originating at the SourceForge site (linked from a
previous comment) by removing unneeded/mistakenly included stuff. It should now
apply cleanly to madwifi-driver-0.1_pre20040824 (without any line relocations).
*** Bug 60626 has been marked as a duplicate of this bug. *** Created attachment 38228 [details, diff]
madwifi-driver-0.1_pre20040824.ebuild.patch
Fixed the madwifi ebuild to apply the previous patch.
Don't know if anybody needs this, but here are the instructions on how to apply the previous two patches (simple but dirty way): 1. cp proc_dointvec.patch /usr/portage/net-wireless/madwifi-driver/files 2. cd /usr/portage/net-wireless/madwifi-driver 3. patch -p0 < madwifi-driver-0.1_pre20040824.ebuild.patch 4. ebuild madwifi-driver-0.1_pre20040824.ebuild digest 5. ACCEPT_KEYWORDS=~x86 emerge madwifi-driver-0.1_pre20040824 Hope it works for you too! Also occurs on 2.6.8-r3. Patch works to resolve. Things work now with gentoo-dev-sources 2.6.8-r3 and madwifi-driver 0.1_pre20040824-r1. I don't know if that patch was merged or what, but thanks for whomever fixed it... Compiles, but in use tells me that the card cannot go into scan mode. Mobile Herd: Seems solar has fixed this but not marked this bug as resolved: 29 Aug 2004; <solar@gentoo.org> +files/madwifi-driver-2.6.8-20040814-proc_dointvec.patch, +files/madwifi-multi-ssid-support.patch, +madwifi-driver-0.1_pre20040824-r1.ebuild: update for 2.6.8 kernels reillyp: I suspect that the problem you mention in comment #11 is not related to this bug or the resolution. The patch applied in the new ebuild seems only to adjust for a Linux driver API change. Maybe you should open another bug describing the exact problems you have? Also, you might want to check the SourceForge bug tracker (although I couldn't find anything similar from a quick look): http://sourceforge.net/tracker/?atid=567753&group_id=82936&func=browse I was not on the CC: of this bug. Are we happy with madwifi-driver now? Any reason to not mark 0.1_pre20040824-r1 stable? Who has tested on 2.4.x kernels? changing resolution to TEST-REQUEST Happy? Yes. Stable? Not really 100%, but (it seems) it is the only option for this hardware. I have experienced total hangs when combining with iproute2 (used by wondershaper) from time to time. Irritating but not frequent enough to merit further investigation (not sure it is even this driver). 2.4 kernels? Works on Debian with CVS from sometime in March, but I haven't tested on Gentoo with a 2.4 kernel. it works for me in gentoo-dev-sources-2.6.8-r3. No hitches whatsoever. Closing as per comment #16. |