Ron Johnson wrote:
On 07/30/08 14:51, Hugo Vanwoerkom wrote:Ron Johnson wrote:On 07/30/08 10:02, David Baron wrote:1. All those section mismatches! This is all over google for the release candidates but still around as of .8. Seem not to have any effect, the thing boots OK, but ....2. Hal on startup gets an invalid op/code busy from the IDE, resets IDE canceling DMA and other enabled functions. These can be restored using hdparm but eventually, unreadable sectors occur and the file system is damaged. There may, in fact, be problems with the drive involved, but these problems rarely if ever occur in 2.6.23.It's either the hardware or the chipset, because the nvidia MCP55 chipset with sata_nv and jmicron drivers havebeen working perfectly for me so far.I think 2.6.25 is too dangerous to use at this point.Add this to your sources.list and build a 2.6.26 kernel: deb http://kernel-archive.buildserver.net/debian-kernel trunk mainDid you? I built a 2.6.26 from kernel.org. Two items: 1. the NVidia driver 96.43-07 applies without having to patch anythingThat's pretty old. Do you have a 52xx or 6xxx series card?
It's coming. As of yet it's an mx4000
173.14.05 compiles perfectly... <pause>I just realized that I haven't rebooted yet, and so am still running a kernel from linux-source-2.6.25 v2.6.25-7.gcc version 4.3.1 (Debian 4.3.1-6) builds nvidia 173.14.05 and runs it well.2. the VMware vmserver 1.0.6 needs update117c and then starting vmware gets an out-of-memory error.Can't comment on that since I don't run VMware.Beyond that I am of the opinion that 2.6.26 (and 2.6.25) is stickier than 2.6.24, the last kernel that the -ck patch could be applied to.It must be hardware incompatibility issues. Life was so much easier in 1984...
No. It's that Con Kolivas quit and nobody knows how to patch the newer kernels for performance the way he did. I started looking at his patch and it is c o m p l i c a t e d!
Hugo