2.6.11.11-dark1: First blood

Message

predatorfreak · Post by **predatorfreak** » Sun May 29, 2005 5:18 pm

New dark time, upgrade to 2.6.11.11 and rediff execshield-nx to fix the rejects against 2.6.11.11.

/* dark-sources
* Release: 2.6.11.11-dark1
* Name: Predator
* Apply over: 2.6.11
* Maintainer: predatorfreak
* Highlights: Upgrade to 2.6.11 and rediff execshield against 2.6.11.11
* Comments: Removed the rose_route fix was merged mainline, so I ripped it from -ac. Execshield-nx reject's solved.
*/

/* Dark core (base) */
patch-2.6.11.11.bz2 /* Linux 2.6.11.11 */
patch-ac7-r1.patch /* ac7, minus the config_hz, 2.6.11.7 and rose_route patches */
2.6.11-staircase-10.7.diff /* Staircase 10.7 rediffed for dark */
s10.7_s11-r2.diff /* Update staircase 10.7 to 11, r2 diffed for dark */
s11_s11.1.diff /* Update staircase 11 to 11.1, SMP improvements */
s11.1_s11.2.diff /* Update staircase 11.1 to 11.2, prevents 32-bit overflows */
schedrange.diff /* Infrastructure for more policies */
schediso2.12.diff /* SCHED_ISO for soft realtime scheduling */
schedbatch2.7.diff /* SCHED_BATCH for idle scheduling */

/* Fixes merged from external trees */
2.6.11-ck-scsifix.diff /* SCSI hangfix merged from -ck */
blockdev-mnt-racefix.diff /* blockdev mount race fixes merged from -mm */
buffer-overflow-fixs.diff /* A few minor buffer overflow fixes merged from -mm */
mm-fixes.diff /* Random fixes for files in the mm/ directory merged from -mm */
vmlinux-fix-physical-addrs-all.diff /* Fixes for vmlinux merged from -mm */
properly-share-process-and-session-keyrings-with-clone_thread.patch /* Merged from -mm, rediffed for reject's */
run-softirqs-on-proper-processor-on-offline.patch /* Merged from -mm */
tty-output-lossage-fix.patch /* tty output lossage fix merged from -mm */
cant-unmount-bad-inode.patch /* Unmount fix merged from -mm */
oops-fix.diff /* oops fix merged from -mm */
copy_pte_range-latency-fix.patch /* Minor latency fix merged from -mm */
sys_setpriority-euid-semantics-fix.patch /* Merged from -mm */
mtrr.patch /* mtrr support fixes, merged from -tiny */
2900_alps-tapping.patch /* Merge ALPS tapping fix from gentoo-sources */

/* cfq v3 (cfq-ts) */
cfq-ts-21-rollup.diff /* cfq-ts-21 and fixes from -mm */
isobatch_ionice2.diff /* SCHED_ISO/BATCH ionice values */
rt_ionice.diff /* SCHED_RR ionice value */

/* Additional features */
realtime-lsm.diff /* realtime-lsm, which allows non-root users access to SCHED_RR */
kmalloc-account.patch /* print memory allocation to /proc/kmalloc */
badram-2.6.11.diff /* BadRAM, which works around bad ram spots */
execshield-nx-r2.diff /* execsheild-nx for buffer overflow protection, rediff for rejects */
mwIII.diff /* Mapped watermark III, merged from -ck */

/* Additional hardware support */
libata-dev-2.6_03-03-2005.patch /* libata-dev snapshot, features new SATA driver's and fixes hdparm support */
sk98lin-8.16.2.3.patch /* Latest version of sk98lin straight from SysKonnect */
skge-netdev-2.6.11-mm4.patch /* SKGE network driver to replace sk98lin, this is considered to be experimental */

/* Other */
irqbal-fast.diff /* irq balance fast merged from -mjb */
slabtune.diff /* slab tuning merged from -mjb */

Diff: http://prdownloads.sourceforge.net/dark ... 2?download

HardenCoonor · Post by **HardenCoonor** » Tue May 31, 2005 11:02 am

Very nice kernel patchset(s). I use this since i got away from Reiser4, and on my server since i noticed there was some BADRAM (just 4K), which probably nuked my reiser4 install there.

Yesterday i did 2 emerges (one kernel, the others were progs), and listened to amarok playing music. Very smooth, no lags, no stuttering. Under KDE 3.4 with arts.

Still, maybe the ALSA driver can be upgraded to 1.0.9, but i did not have major issues with sound.

vipernicus · Post by **vipernicus** » Tue May 31, 2005 1:50 pm

Would you mind making a broken out? I'm having problems with execshield and 12-rc5, maybe I can get tips from your diff.

predatorfreak · Post by **predatorfreak** » Tue May 31, 2005 2:41 pm

HardenCoonor wrote:Very nice kernel patchset(s). I use this since i got away from Reiser4, and on my server since i noticed there was some BADRAM (just 4K), which probably nuked my reiser4 install there.

Yesterday i did 2 emerges (one kernel, the others were progs), and listened to amarok playing music. Very smooth, no lags, no stuttering. Under KDE 3.4 with arts.

Still, maybe the ALSA driver can be upgraded to 1.0.9, but i did not have major issues with sound.

I prefer to stick with what is mainline for alsa, as it generally works out of the box. Also, always good to have postive feedback.

vipernicus wrote:Would you mind making a broken out? I'm having problems with execshield and 12-rc5, maybe I can get tips from your diff.

My diff is strictly against 2.6.11.11, I'd recommend trying the 2.6.11-bk10 diff from Ingo, which is updated to work with the merging of address space randomization code to mainline (this was merged in 2.6.11-bk, so it is in all of the 2.6.12-rc series). Mine simply fixes rejects caused by the sanity check improvements merged in 2.6.11.11.

vipernicus · Post by **vipernicus** » Tue May 31, 2005 4:08 pm

predatorfreak wrote:My diff is strictly against 2.6.11.11, I'd recommend trying the 2.6.11-bk10 diff from Ingo, which is updated to work with the merging of address space randomization code to mainline (this was merged in 2.6.11-bk, so it is in all of the 2.6.12-rc series). Mine simply fixes rejects caused by the sanity check improvements merged in 2.6.11.11.

Awesome, thanks.

HardenCoonor · Post by **HardenCoonor** » Wed Jun 01, 2005 10:59 am

I prefer to stick with what is mainline for alsa, as it generally works out of the box.

That's true.

Also, always good to have postive feedback.

That is why i wrote this. I used the last 4-5 versions of your patchset on my server (it is just a K6-2 450), and had no issues with the kernel, which unfortunately cannot repair my buggy harddisks and ram.

Thank you once more, and keep up the good work.

predatorfreak · Post by **predatorfreak** » Wed Jun 01, 2005 6:32 pm

HardenCoonor wrote:
I prefer to stick with what is mainline for alsa, as it generally works out of the box.
That's true.

Also, always good to have postive feedback.
That is why i wrote this. I used the last 4-5 versions of your patchset on my server (it is just a K6-2 450), and had no issues with the kernel, which unfortunately cannot repair my buggy harddisks and ram.

Thank you once more, and keep up the good work.

Will do.

vipernicus · Post by **vipernicus** » Wed Jun 01, 2005 10:49 pm

predatorfreak wrote:My diff is strictly against 2.6.11.11, I'd recommend trying the 2.6.11-bk10 diff from Ingo, which is updated to work with the merging of address space randomization code to mainline (this was merged in 2.6.11-bk, so it is in all of the 2.6.12-rc series). Mine simply fixes rejects caused by the sanity check improvements merged in 2.6.11.11.

Yeah, that one patched alot better, but my main problem still is mm/mmap.c, it is so much different, I can't figure out how to get it working.

predatorfreak · Post by **predatorfreak** » Thu Jun 02, 2005 2:11 am

vipernicus wrote:
predatorfreak wrote:My diff is strictly against 2.6.11.11, I'd recommend trying the 2.6.11-bk10 diff from Ingo, which is updated to work with the merging of address space randomization code to mainline (this was merged in 2.6.11-bk, so it is in all of the 2.6.12-rc series). Mine simply fixes rejects caused by the sanity check improvements merged in 2.6.11.11.
Yeah, that one patched alot better, but my main problem still is mm/mmap.c, it is so much different, I can't figure out how to get it working.

My 2.6.11.11 diff fixes this, they rearranged the sanity check. Although, putting in the code at the end of the sanity check and removing the EXPORT_SYMBOL line worked fine for me, compiles and runs fine aswell.

Edit: Typo's.

vipernicus · Post by **vipernicus** » Thu Jun 02, 2005 4:34 pm

So, are you saying to make the changes and take out the EXPORT GET_UNMAPPED_AREA?

predatorfreak · Post by **predatorfreak** » Thu Jun 02, 2005 5:08 pm

vipernicus wrote:So, are you saying to make the changes and take out the EXPORT GET_UNMAPPED_AREA?

no, the EXPORT_SYMBOL which also fails when patching.

vipernicus · Post by **vipernicus** » Thu Jun 02, 2005 5:34 pm

predatorfreak wrote:
vipernicus wrote:So, are you saying to make the changes and take out the EXPORT GET_UNMAPPED_AREA?
no, the EXPORT_SYMBOL which also fails when patching.

Oops, misunderstood what you were saying earlier, sorry.