| Summary: | bcm570x-7.0.0.ebuild (New package) | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | Thomas Eckert <eckert.thomas> |
| Component: | New packages | Assignee: | x86-kernel (DEPRECATED) <x86-kernel> |
| Status: | VERIFIED FIXED | ||
| Severity: | enhancement | CC: | avenj, steel300, x86-kernel |
| Priority: | High | ||
| Version: | unspecified | ||
| Hardware: | All | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
| Attachments: |
bcm570x-7.0.0.ebuild
bcm570x-7.1.22.ebuild |
||
|
Description
Thomas Eckert
2003-11-08 11:32:07 UTC
Created attachment 20449 [details]
bcm570x-7.0.0.ebuild
I don't want to push too hard, but I have to negative tests with dual-opteron boards and the in-kernel tg3-driver; if modular the machine keeps running at i386 speed for a few seconds and then freezes, if it is compiled in the the machine seems to freeze instantly (but the debugging in this case is very limited). I'd really be happy if this bcm makes it into portage. grrr s/to negative/two (2) negative/ err, going to close this. The bcm5700 driver is in module form in the gentoo-dev-sources-2.6.0_beta10-r1 look for it under the 1000MB adapters .... -Brad so what (no offense, really)? if i don't use the gentoo-patched kernel i'm lost. beside that there are even ebuild for the intel "e100" and "e1000" drivers (and there two are included in the vanilla kernel). then patch using this: http://dev.gentoo.org/~brad_mssw/kernel_patches/2.6.0/genpatches-0.6/202_bcm5700_broadcom_gigabit_drvr.patch you should use gentoo-dev-sources anyhow, it's not going to become as patched as gentoo-sources. Also, the tg3 driver is improving, and that's where everything needs to be going. The bcm5700 driver is not pretty and should not be officially supported. -Brad i get your point, but I don't use 2.6.x kernels. i _never ever_ use prepatched kernels from anyone (i want to keep as close to vanilla as possible). your dev-sources may be really good and only patched with selected stuff -- but nevertheless i want to decide myself (and that's what Gentoo is about and one reason why i use it). i personally do not care if it goes to portage (i have this ebuild) -- but i had serious problems with tg3 that where fixed with bcm (it may be "not pretty" code but it does a far better job that tg3). so i wanted to save other people the time to reinvent the wheel. if it does not make it in, ok (i'm not going to "complain" any more); at least it's documented in bugzilla. Thomas sorry but need to reopen Created attachment 26933 [details]
bcm570x-7.1.22.ebuild
I'd like to bounce this back to bug-wranglers as it's not an isolated amd64
related thing -- but (once again) bugzilla strickes back and won't let me do
this
:(
ahh, a bug needs to be open to re-assign (my apologies to bugzilla ;) Why was this bounced back to AMD64? The only Gentoo-supported kernel on AMD64 already has this, so this is an issue for x86 people (maybe the kernel team?) this bug is frustrating guys: the interresting thing is, that gentoo-docs (http://www.gentoo.org/proj/en/base/amd64/index.xml?part=2&chap=2) mention the bcm570x-driver, I read somewhere that it is included on the boot-cd for amd64 but for the average user it's needlessly hard to get her system up and running. I understand your argument concerning the e100/e1000 drivers. I have added the ebuild to portage under net-misc. Closed, added to portage. |
