2.6.8.1-nitro5 "it cleans your breath while it cleans y

Message

teutzz · Post by **teutzz** » Fri Aug 27, 2004 12:31 pm

i reversed it, staircase 8.0, (it did however had test1 on top of it) (this is what i originally did) and it failed
as for sc_mw can i apply it on top (reversing if necessary) of watermark1 + watermark_diff + watermark2 (applied in that order)?

neonik · Post by **neonik** » Fri Aug 27, 2004 12:44 pm

teutzz wrote:i reversed it, staircase 8.0, (it did however had test1 on top of it) (this is what i originally did) and it failed
as for sc_mw can i apply it on top (reversing if necessary) of watermark1 + watermark_diff + watermark2 (applied in that order)?

You should reverse staircase-8.0 and s8.0_test1 to apply the new staircase patch.

sc_mw can go on top of those, either way run patch with --dry-run first.

Here a list of perfectly working patches:

Jens Bergmann wrote:schedrange.diff
schedbatch2.4.diff
schediso2.5.diff
sched-adjust-p4gain
mapped_watermark2.diff
defaultcfq.diff
config_hz.diff
1g_lowmem_i386.diff
akpm-latency-fix.patch
9000-SuSE-117-writeback-lat.patch
cddvd-cmdfilter-drop.patch
cool-spinlocks-i386.diff
bio_uncopy_user-mem-leak.patch
bio_uncopy_user2.diff
ioport-latency-fix-2.6.8.1.patch
make-tree_lock-an-rwlock.patch
invalidate_inodes-speedup.patch
2.6.8.1-ck4-version.diff

And read this message: http://bhhdoa.org.au/pipermail/ck/2004- ... 00601.html

sdaffis · Post by **sdaffis** » Fri Aug 27, 2004 1:22 pm

Hey, it doesn't clean my teeth.. False advertising!

teutzz · Post by **teutzz** » Fri Aug 27, 2004 2:31 pm

@neonik: i think the sc_mw patch did it, i think it was the final piece of the puzzle where as this kernel is concearned... really
as for staircase8.1 i didn't have the guts or patiance to apply it... maybe later if i don't have anything else to do

neonik · Post by **neonik** » Fri Aug 27, 2004 2:37 pm

teutzz wrote:@neonik: i think the sc_mw patch did it, i think it was the final piece of the puzzle where as this kernel is concearned... really
as for staircase8.1 i didn't have the guts or patiance to apply it... maybe later if i don't have anything else to do

Wonderful, then I guess it's more than worthwile to advise everyone having any slowdown issues to give it a try.

teutzz · Post by **teutzz** » Fri Aug 27, 2004 3:07 pm

neonik wrote:
teutzz wrote:@neonik: i think the sc_mw patch did it, i think it was the final piece of the puzzle where as this kernel is concearned... really
as for staircase8.1 i didn't have the guts or patiance to apply it... maybe later if i don't have anything else to do
Wonderful, then I guess it's more than worthwile to advise everyone having any slowdown issues to give it a try.

i would also advice the ones who experience swaping more than they are used to

seppe · Post by **seppe** » Fri Aug 27, 2004 3:10 pm

FYI, I'm creating a new nitro .. this time with voluntary preempt, it's not going to be an easy job. When I applied staircase8.1 against it, I got a lot of failed hunks. I fixed them, and it compiles cleanly .. but I still have to add like 40 other patches or so

Robin79 · Post by **Robin79** » Fri Aug 27, 2004 3:47 pm

seppe: i will w8 with pleasure for the next nitro!!!! Me loves nitro-sources i aint even using Love anymore cause the nitro is so much more stable than the love (and f**king mm-sources) and the ck patches just speeding up my system so much!

miseiler · Post by **miseiler** » Fri Aug 27, 2004 4:00 pm

Vide wrote:
Strange, my radeon card worked fine. It might sound silly, but you HAVE emerged ati-drivers then opengl-update ati since you installed the new kernel, right? And are you using the 3.9.0 drivers or the 3.11.1 drivers, which work for me?

What does fglrxinfo show?
Obviously I reemerged and updated everything, I'm not in mesa
The nly thing is that I'm using 3.9.0 ati driver, and I'm running a NForce2 mobo. What's your chipset?

Anyway I'll give a try to older ati driver version

3.11.1 is the newest version. ACCEPT_KEYWORDS=~x86 emerge ati-drivers ati-drivers-extra to get it.

Or you can do it properly (since I'll get yelled at for suggesting the above), and do the following:

Code: Select all

echo "media-video/ati-drivers ~x86" >> /etc/portage/package.keywords
echo "media-video/ati-drivers-extra ~x86" >> /etc/portage/package.keywords
emerge ati-drivers ati-drivers-extra

Hope that helps.

neonik · Post by **neonik** » Fri Aug 27, 2004 4:11 pm

Con released the new patchset -ck5 with all the fixes for -ck4.
http://ck.kolivas.org/patches/2.6/2.6.8.1/2.6.8.1-ck5/

Con Kolivas wrote:Added since ck4:
+s8.0_s8.1
Sync with latest staircase patch. Tiny fix affecting very short running tasks that use a lot of cpu (like xmame).

+mapped_watermark_fix.diff
+sc_mw.diff
Mapped watermark was released a touch too early and was too aggressive leading to an oom easily. These fix that, make it more efficient and make it possible to properly inactivate it by setting mapped to 0.

+1g_change_config.diff
Change the 1Gb lowmem so that the default is off (thanks for suggestion Michael Buesch)

http://bhhdoa.org.au/pipermail/ck/2004- ... d.html#end

miseiler · Post by **miseiler** » Fri Aug 27, 2004 4:56 pm

Couple of things.

First, the s8.0_test1 and s8.0_s8.1 patches (from ck4dev and ck5, respectively) are both meant to patch against the already-applied staircase 8.0. Also, these are the SAME PATCHES...so you don't need to go reversing staircase, then reapplying the "new" patches or whatnot

... If you've already applied the ckdev s8.0_test1 patch, you are already at the newest version of staircase.

Second, the sc_mw patch is meant to be applied to a kernel already using mapped_watermark.diff AND mapped_watermark_fix.diff. I think neonik already mentioned this, but it's worth repeating, I think...if nothing else to clear up any confusion.

So. To assure that you have the latest patchset from ck5 in your nitro5 kernel, you need to apply these patches in the following order:

s8.0_s8.1 OR s8.0_test1
mapped_watermark_fix.diff
sc_mw.diff

miseiler · Post by **miseiler** » Fri Aug 27, 2004 5:19 pm

Oh, and I have to say the above fixes all problems entirely. I've been abusing my filesystem (compressing, decompressing, tarring, untarring, moving around 500MB files, etc) and according to gkrellm my free memory has never gone under 1/3rd of my total RAM, let alone swapped out. Which, I should add, is exactly what Con said the mapped_watermark business was supposed to do from the beginning

I am very, very impressed.

Rainmaker · Post by **Rainmaker** » Fri Aug 27, 2004 11:45 pm

Well, I took out the mapped and replaced it with mapped_watermark3 (released today

)

It seems to help for my swap issues. I haven't tested it fully, I'll do that tomorrow sometime. At first glance, all looks well, though I must say I still have a lot of harddisk I/O throughput... A LOT less then I had with the default watermark patch, but still a lot compared to 2.6.7-r11

I tried upgrading the staircase scheduler, but couldn't reverse the patch because of failed hunks. I think I could figure them out, but I'm too tired now

. I think I'll keep running this kernel until the next release comes out. Great work!

st_lim · Post by **st_lim** » Sat Aug 28, 2004 3:52 am

Hi,
I initially had the swapping problem, and was thinking that nitro5 was the absolute worst piece of crap.
I finally got all this working. but compiling without 4k stacks. I did patch in the s8.0_8.1, mapped_watermark_fix and sc_mw.diff. and set the low mem kernel patch.

Hope this helps. I think we have sufficient stuff to get nitro 6 up and running??

Ciao
ST Lim

kohno · Post by **kohno** » Sat Aug 28, 2004 4:18 am

Why is that I don't find reiser4 option in this kernel? Some one told me to deselect 4K stack, then it'd appear, but that didn't work for me.

Code: Select all

 Linux Kernel v2.6.8-nitro5 Configuration
 ─────────────────────────────────────────────────────────────────────────────────────────
  ┌─────────────────────────────────── File systems ───────────────────────────────────┐
  │  Arrow keys navigate the menu.  <Enter> selects submenus --->.  Highlighted        │
  │  letters are hotkeys.  Pressing <Y> includes, <N> excludes, <M> modularizes        │
  │  features.  Press <Esc><Esc> to exit, <?> for Help.  Legend: [*] built-in  [ ]     │
  │  excluded  <M> module  < > module capable                                          │
  │ ┌────────────────────────────────────────────────────────────────────────────────┐ │
  │ │         <*> Second extended fs support                                         │ │
  │ │         [ ]   Ext2 extended attributes                                         │ │
  │ │         <*> Ext3 journalling file system support                               │ │
  │ │         [*]   Ext3 extended attributes                                         │ │
  │ │         [ ]     Ext3 POSIX Access Control Lists                                │ │
  │ │         [ ]     Ext3 Security Labels                                           │ │
  │ │         [ ] JBD (ext3) debugging support                                       │ │
  │ │         <*> Reiserfs support                                                   │ │
  │ │         [ ]   Enable reiserfs debug mode                                       │ │
  │ │         [*]   Stats in /proc/fs/reiserfs                                       │ │
  │ │         [ ]   ReiserFS extended attributes                                     │ │
  │ │         <*> JFS filesystem support                                             │ │
  │ │         [ ]   JFS POSIX Access Control Lists                                   │ │
  │ │         [ ]   JFS debugging                                                    │ │
  │ │         [ ]   JFS statistics

TierMann · Post by **TierMann** » Sat Aug 28, 2004 5:13 am

The option you may be missing is under Code Maturity Level Options --> "Prompt for development and/or incomplete code/drivers". That should be on to view all the "EXPERIMENTAL" options.

Reiser4 will show up regardless of the 4k stack setting because it's only dependant on the "EXPERIMENTAL" (above) setting. It just (with recent versions) won't let you compile with 4k stacks and Reiser4 enabled at the same time.

Hope that helps.

kohno · Post by **kohno** » Sat Aug 28, 2004 5:20 am

TierMann, thank you for your prompt reply, the "Prompt for development and/or incomplete code/drivers" is already on, but still I don't see Reiser4.

Bot24 · Post by **Bot24** » Sat Aug 28, 2004 7:42 am

Will somebody please releave the confusion us Nitro n00b's are having by giving instructions as to how I should get this ck5 patch set into Nitro and then that other stuff too, or updating the ebuild so that we can be lazy.

jewps · Post by **jewps** » Sat Aug 28, 2004 7:53 am

Nitro6 should be out soon, should be composed of ck5 and vol pre-empt. Seppe asked me to fix his work-in-progress patch of nitro6 but it was borked so I couldn't work on it. I think we should hold out till 2.6.9 + ck5, whenever Con decides to release some patches for .9.

Sorry I couldn't fix the problems seppe, IMO, we should hold till 2.6.9.

Vide · Post by **Vide** » Sat Aug 28, 2004 11:12 am

miseiler wrote: 3.11.1 is the newest version. ACCEPT_KEYWORDS=~x86 emerge ati-drivers ati-drivers-extra to get it.

Hope that helps.

Ok, emerged the new ati-drivers...now I can't get my old 2.6.5-r1 to recognize ATI OpenGL anymore (I'm always in Mesa..but everything is as before in my confs!!), while it's recognized in nitro5
The only problem is only that if I launch fgl_glxgears my machine hang up with blank screen (well, no signal on the screen at all, it's like no cable attached or PC powered off)

Sigh

Please, tell me what mobo chipset you have, I think it's important..

Bot24 · Post by **Bot24** » Sat Aug 28, 2004 6:08 pm

To get rid of the ATI stuff were the monitor stops getting a signal, you need to disable DGA. You might need to update to the latest drivers on all your kernels because there is a module component. I have some kernel module option about AGPLock in my xorg.conf because this fixes crashes with Blender, the latest Crystal Space Libs, and resizing windows that contain an overlay a lot.

Did you get ATI drivers working with acceleration on VESAtng? Mine were stuck in low resolution.

Robin79 · Post by **Robin79** » Sat Aug 28, 2004 6:14 pm

Me want the new nitro eheheh

Crisp · Post by **Crisp** » Sat Aug 28, 2004 7:41 pm

kohno wrote:TierMann, thank you for your prompt reply, the "Prompt for development and/or incomplete code/drivers" is already on, but still I don't see Reiser4.

I've had the same problem; appearantly, once you have run make menuconfig and a kernel compile, all special options disappear when you re-run "make menuconfig" (I only saw the regular kernel options). One solution is to re-emerge the source. I am not well enough into the kernel source to pinpoint the exact problem, but perhaps it helps for Seppe to pindown the problem. I no longer have this problem with the 2.6.8.1-r6 nitro source.

[edit: part of post moved to nitro 2.6.8-r6 thread]

2.6.8.1-nitro5 "it cleans your breath while it cleans y

Finally, working