Summary: | cisco-vpnclient-3des-4.6.03.0190 fails to build against vanilla-sources 2.6.14-rc2 | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Mike Auty (RETIRED) <ikelos> |
Component: | Current packages | Assignee: | Chris Gianelloni (RETIRED) <wolf31o2> |
Status: | VERIFIED FIXED | ||
Severity: | normal | CC: | patrick, tom |
Priority: | High | ||
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Attachments: |
Updated ebuild running the linux-2.6.14 patch.
Patch for 2.6.14 kernel (use ebuild) |
Description
Mike Auty (RETIRED)
2005-09-25 13:20:42 UTC
Well, I'd recommend asking Cisco for a patch. After all, you are paying them for support on it, right? ;] Once someone posts a patch here, I'll add it, but I'm not spending my time writing a patch for a commercial product. I'm not trying to be rude, I just don't think I should be spending my volunteer time working for Cisco without compensation. Feel free to REOPEN this if you get a patch from Cisco (or somewhere else). see http://forums.gentoo.org/viewtopic-t-375714.html for solving the problem. Don't know how to make a patch that's why I don't do it. But I hope, someone is able to... Created attachment 71759 [details]
Updated ebuild running the linux-2.6.14 patch.
First *real* ebuild modification, it works for me.. but may break on older
kernel versions (the patch runs with no condition).
Created attachment 71760 [details, diff]
Patch for 2.6.14 kernel (use ebuild)
As above, first real time fiddling with ebuilds/C.. probably best if
experienced eyes give it a good run over :)
There is now a patch courtesy of DaNIsH (and whilst the patch works, the ebuild will apply it to every system whether or not it's 2.6.14 so more work is needed there), but it's not from Cisco, and it's unclear if any other changes to the skbuffer made in the latest kernel updates will suddenly show up as bugs. However, it'd be nice to see it in a new ebuild just to get people back up and working again, especially as 2.6.14's just gone stable, so I've marked the bug as open again. The patch breaks 2.6.13 (at least) so I have forced it to only apply if the kernel is 2.6.14 for now. I don't know if 2.6.15 will require this patch also. If it does, then I'll need to rework the current conditions in the ebuild to match all kernels >=2.6.14 instead. Added to CVS and taken to stable. Thanks for sorting that. 2.6.15 will require the patch, the changes to the skbuffer seem to have been intentional and will stay. The patch does still work though, and nothing else needs to be added... The new r1 ebuild has a problem with the kernel version checking code. The following line (in src_unpack()) is being used to make sure that the kernel version is greater than 2.6.14. if kernel_is 2 6 && [ ${KV_MINOR} -ge 14 ] The variable KV_PATCH should be used here instead of KV_MINOR. d'oh! Fixed... |