LFS-6.6, Stage2, glibc, nscd.c:442

linux fan linuxscratch at gmail.com
Thu May 27 21:21:15 PDT 2010


On 5/27/10, Paul Rogers <paulgrogers at fastmail.fm> wrote:

[Ken Moffat]
>> getting from a .config that worked in 2.6.18 to one that works in
>> 2.6.32.  Particularly, libata [ most ide drives now use libata and
>> therefore /dev/sdXn instead of hdXn ].
[Paul Rogers]
> Yes, I've seen that in some of the other distros I've gotten from LXF.
> Using SCSI emulation for IDE drives just seems WRONG somehow.  I'm a
> follower of Einstein's dictum, "As simple as possible, but no simpler."
> The systems I'm building on/for (P3's) all use PATA drives, and I was
> hoping to deal with SATA later in a POD-3.2 version.
>

I maintain some pure IDE drives as hdXn by using the config parameters:
CONFIG_HAVE_IDE=y
CONFIG_IDE=y
in addition to:
CONFIG_ATA=y
CONFIG_SATA_SIL24=y
CONFIG_SATA_PROMISE=y
for the SATAS as sdXn, providing the peaceful coexistence of both worlds.

This old heap of harware just _won't_  stop working!
I got from 2.6.18 to 2.6.32, over time, a few steps at a time and the
technique was to:
copy last good config to kernel souces .config
make menuconfig
change one thing and change it back (unless really changing something)
"make menuconfig" pulls the relevant new vs. deprecated pieces into line.
go through the whole menus with an eye on reality checking.

If I want to boot with my IDE /dev/hdXY's plugged in (always at the
fixed cable positions), I can, and I can access the /dev/sdXY's too,
or visa versa.

I still think 2.6.18 is a good kernel, but there was a clean break in
compatibility with newer software somewhere around 2.6.22.



More information about the lfs-support mailing list