| Summary: | Gentoo-test-sources-r1 doesn't have Athlon-MP processor config option | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | Jesse Adelman <jesse> |
| Component: | [OLD] Core system | Assignee: | Brian Jackson (RETIRED) <iggy> |
| Status: | RESOLVED FIXED | ||
| Severity: | normal | CC: | fghellar, heiko.baums |
| Priority: | High | ||
| Version: | unspecified | ||
| Hardware: | x86 | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
|
Description
Jesse Adelman
2003-11-23 15:06:48 UTC
Sorry, it's "-r1", not "-r8". Sorry. Along the same line, Athlon XP is not there either. show me that it makes any difference and I'll add it If it doesn't make any differnece, then why is it in gentoo-sources' 2.4.20 kernel config? Isn't this the test kernel that will become gentoo-sources, and, as such, should have all the options of the current gentoo-sources, not less (unless there's a reason for it, like the module is no longer in the Kernel tree)? Perhaps it should be removed from there if it doesn't make any differnece? However, I would assume that it at least adds -march=athlon-mp and/or -mcpu=athlon-mp to gcc's options, rather than just -march=athlon. Now, that may make no difference, either, but it may in future versions of GCC, no? Otherwise the GCC devs wouldn't have added it as an option, right? And, as Felipe added, there are other processor types missing from the choice list in this kernel config in addition to just the Athlon-MP. I appreciate everyone's work on the default Gentoo kernel - evidently the kernel package managers get no love most of the time. But, your work is appreciated! The thing is, different people did the work for the 2.4.20 based gentoo-sources and the 2.4.22 based gentoo(-test)-sources. I was trying to trim the number of patches that got applied, and hence the number of patches we had to maintain. I have nothing against the patch, I just want people to justify what gets added to gentoo-sources so we don't end up with nightmarish patch sets. I am not however going to go back and remove stuff from an already working kernel. Hope this explains things more clearly. >> " show me that it makes any difference and I'll add it " That's the idea of source code... Anyway, it does make difference. Your kernel gets compiled as plain, vanilla code which runs as plain, vanilla code or it gets compiled for your architecture using the proper flags, and thus forth. Equals faster and / or better kernel. >> " the number of patches we had to maintain. " In the end, I don't mind maintaining big patchsets and I'm sure users don't mind using big patchsets. What matters is that patches are neither slacked upon, but are not added into the kernel without consideration. If properly done, your kernels would work... I don't see an upper limit for patches or such, so I see no reason why we can intergrate more patches - our Patch-Like-O-Matic system allows for this and allows patchset configuration at merge time, so I see no reason why more patches to benefit the user can't go in. All that's needed for this is that things are properly tested, cared, and released. >> "nightmarish patch sets" As long as they are tamed and cared for, that won't // shouldn't happen... Like I said, having 20 or 200 patches is not bad. Or good. What's important is that these patches do their job, make a difference and have an impact which the user wants and requires. Some patches can only benefit, while others can cause headaches. The decisions are up to you... *** Bug 35170 has been marked as a duplicate of this bug. *** First I'm sorry to make a duplicate bug report but I haven't found this report when searching for an existing report for gentoo-sources 2.4.22-r1. But now I have a question. Is there a difference between kernel optimization for Athlon and for Athlon XP or not? If there is no difference then Athlon XP should be added at least in the help text of the "Processor family" option that has to be selected for Athlon XP optimization so that it's clear what option to select for this processor. If there is a difference then of course there should be an additional option for Athlon XP because one of the main benefits of Gentoo Linux is the possibility to optimize everything perfectly for the particular hardware. And then it doesn't matter how many patches have to be applied to the kernel. Sorry, but I have to admit that I'm just a bit confused at present. fixed in gentoo-sources-2.4.22-r4, though I'm still skeptical of it's benefits |
