On Sat, 1 Nov 1997, James Lewis Nance wrote:
> Hello All,
> I have found several problems with 2.1.60 and 2.1.61 which I think
> I should report. I have not done much testing of the 2.1.5X kernels,
> so I can not say for sure when these problems appeared.
> The machine I am using is built around a Cyrix 6x86 P166 processor.
> It has IDE disks. I can be more specific if anyone wants more details.
> The machine runs well with all the 2.0.X kernels I have tried, and I
> am currently running it with 2.0.31. Here are the problems I am
> having with 2.1.6X kernels:
>
> 1) The command "/sbin/hdparm -u 1 -d 1 -c 1 /dev/hdc" fails to put the
> drive into dma mode for 2.1.6X kernels, but it works find with 2.0.X
> kernels.
Suspect: Motherboard variety? Not sure.
> 2) When I apply Mike Jagdis'es 2.1.39 Cyrix patch, the cpu type as
> reported by uname is iE86 instead of i586.
Patch: (the 2.1.39 cyrix patch causes this bug....)
==========================================================================
--- linux/include/asm-i386/bugs.h~ Sun Oct 26 01:14:10 1997
+++ linux/include/asm-i386/bugs.h Sun Oct 26 01:35:08 1997
@@ -171,7 +171,6 @@
check_tlb();
check_fpu();
check_hlt();
- system_utsname.machine[1] = '0' + (x86 & 0x0f);
check_amd_k6();
- system_utsname.machine[1] = '0' + x86;
+ system_utsname.machine[1] = '0' + (x86 & 0x0f);
}
==========================================================================
> 3) If I don't apply the Cyrix patch mentioned above, the kernel fails
> to boot. It complains that my processor does not honor the WP bit
> and that I should recompile for a 386. I don't have any problems
> booting 2.0.X kernels w/o the Cyrix patches, and /proc/cpuinfo shows
> that the wp bit works.
May I toss another vote in favour of this patch being adopted?
<grin>
Especially if there's computers that won't run without it....
(I haven't tested since 2.1.39....)
G'day, eh? :)
- Teunis