vivid-sources

[chickaroo]

i was reading the lkml recently, and saw a patch submitted
[PATCH] fix spurious OOM kills
i guess they are still working on that issue

[fallow]

i was reading the lkml recently, and saw a patch submitted
[PATCH] fix spurious OOM kills
i guess they are still working on that issue

yes ,it`s happens from some time ago...
now I added Con`s vm , not many mm vm* and mm* modyfications.
lufs is fixed and I`m doing final testing
I do some select for page_alloc from menuconfig , I hope that now oom`s will be ok . release is based on bk6 and many of fixes is in it.

greetings

[fallow]

ok, I have some problems with my ftp connection

new release is based on 2610rc2bk6
main mission objectives :
- return to Con`s vm aside of mm`s vm*
- O)1 CPU Scheduler + partial mm fixes
- as usually fbsplash and reiser4+fixes from latest mm
- additional fs`s , config-tools , and fixes like nvidia
- and mm scheduling and other fixes/speed-ups.

"vesa_rrc" for vesa rrc version with definitions for 1024x768@85Hz
"vesa_tng" for vesa tng

and list/info

Code: Select all

2.6.10-rc2-vv_e2 (bk6 based)
CPU Scheduler : O)1 + partial mm fixes
IO Scheduler  : switchable at runtime,default selectable from menuconfig or bootparams
--bk`s------
patch-2.6.10-rc2-bk6.bz2 - vanilla bk snapshot
2.6.10-rc2-libata1-dev1.patch.bz2 
2.6.10-rc2-netdev1.patch.bz2 
bk-acpi.patch 
bk-agpgart.patch 
bk-alsa.patch 
bk-ide-dev.patch 
bk-ntfs.patch 
--base-----
add_vivid_menu.diff - add some features 
 [*]   Preempt The Big Kernel Lock ( Ingo`s stuff ported by Con )  
      Default IO scheduler (CFQ v2 io scheduler + updates)  --->    
*** for select default io scheduler as alternative to bootparams elevator=x that works also. ***  
  [ ] old ll_rw_blk version ( no high i/o load lkml fix ) 
*** for without high i/o load fix (lkml).Some people said that is better without it ***
  [*] [page_alloc] zone->pages_lots=((zone->pages_min*6)/4)*2   
*** in page_alloc page_lots=((pages_min*6)/4)*2 like in bk`s , or as in mwII *6 only.I don`t know what is better.
  [ ] CFLAGS hacking 
*** fixed by Troll some love stuff
  [*] /proc/mm support     
*** from SKAS patch.link for info with SKAS patch line.

2.6.10-rc2-mingoll.diff - Ingos latency fixes ported by Con 
host-skas3-2.6.9-v7.patch - SKAS interesting patches ( http://user-mode-linux.sourceforge.net/skas.html )
skas-update-2.6.10.patch - update for 2610rc
increase_prio0_scanning.diff + mwII.diff - Con`s vm work.                       
2610rc2fbs_temp2.diff - fbsplash 
2.6.10-rc2-nvidia-fix.diff - nvidia fix from ck 
nvidia_compat.diff - yet another nvidia fix from ck
2.6.6-qsort-updated-lkml.patch - yet anohter qsort implementation from pld ( www.pld.org.pl ) 
2.6.6-xfs-qsort-lkml.patch - and pld xfs-qsort patch
config_hz.diff - for setting kernel internal clock freq.
squashfs2.0-patch - http://distributions.linux.com/howtos/SquashFS-HOWTO/gettingready.shtml
supermount-ng208-10rc2ck2.diff - tool for auto(u)mount removable media`s from ck
acerhk.patch - acer keyboards support
cddvd-cmdfilter-drop.patch - fix for non root user`s from ck.
cflags-selection.patch - fixed by Troll love stuff for easy apped own CFLAGS for kernel.
chmp-r4-2.6.9.patch - configurable hid mouse polling ;)
config-nr-tty-devices.patch - as in name :)
daconfig-2.1.1.patch - DaMouse "next generation" of menuconfigname ;)
lirc-2.6.9-20041024-px1.patch - fixed by Pax82 lirc patch.
lufs-0.9.7-vivid.patch.bz2 - lufs fixed for 2610rc2
vesa_rrc / vesa_tng selectable as usually via USE flags - http://dev.gentoo.org/~spock 
  "vesa_rrc" for vesa rrc 
  "vesa_tng" for tng version 
--mm part/fixes/speed-ups/clean-ups----- / info : mm notcie/changelog on www.kernel.org /
Reiser4 from latest mm with all fixes :)
sched-active_load_balance-fixlet.patch
sched-add_requeue_task.patch
sched-adjust_timeslice_granularity.patch
sched-alter_kthread_prio.patch
sched-can_migrate-exception-for-idle-cpus.patch
sched-more-agressive-wake_idle.patch
sched-newidle-fix.patch
sched-remove_interactive_credit.patch
sched-reset-cache_hot_time.patch
sched-use-cached-current-value.patch
CONFIG_SOUND_VIA82CXXX_PROCFS.patch
allow-admin-to-enable-only-some-of-the-magic-sysrq-functions.patch
dont-hide-thread_group_leader-from-grep.patch
enforce-a-gap-between-heap-and-stack.patch
hold-bkl-for-shorter-period-in-generic_shutdown_super.patch
invalidate_inode_pages-mmap-coherency-fix.patch
invalidate_inodes-speedup.patch
make-tree_lock-an-rwlock.patch
oss-ac97-quirk-facility.patch
possible-dcache-bug-debugging-patch.patch
ppc64-reloc_hide.patch
provide-a-filesystem-specific-syncable-page-bit-fix-2.patch
provide-a-filesystem-specific-syncable-page-bit-fix.patch
provide-a-filesystem-specific-syncable-page-bit.patch
reiser3-cleanups.patch

ebuild -> http://www.munet.eu.org/~fallow/vv_e2/v ... -r2.ebuild
all dir -> http://www.munet.eu.org/~fallow/vv_e2/

of course feedback , ideas , cooperation will be nice as usually
greetings

[chickaroo]

beautiful! everything is working perfectly (including lufs)
your kernels rock!

[fallow]

thnx
nice to hear that
You`re welcome

greetings

[travisau]

First time I've been able to get ivtv, reiser4, splash, AND lirc working in the same kernel. Excellent job with these sources! Time to migrate the rest of my boxes over.

[fallow]

great & nice

now vivid-sources also on sourceforge.net
2 lines - so vivid-stable and vv_e

http://sourceforge.net/projects/vivid-sources

now available is latest vv_e with ebuild for sourceforge server and stable release also in near future.

of course future info also will be placed here
greetings

[vrln]

Sorry for asking this, but what is vesa rrc? I've never heard of the term. Is it the same thing as the vanilla vesafb? (afaik it is not possible to set refresh rates with vesafb). I would use vesafb-tng, but it locks up on my system (ht p4).

edit: I'll take that vesafb-tng comment back. Just compiled a kernel with vesafb-tng and it works this time... Last time I tried it locked up during boot.

[fallow]

Sorry for asking this, but what is vesa rrc? I've never heard of the term. Is it the same thing as the vanilla vesafb? (afaik it is not possible to set refresh rates with vesafb). I would use vesafb-tng, but it locks up on my system (ht p4).

edit: I'll take that vesafb-tng comment back. Just compiled a kernel with vesafb-tng and it works this time... Last time I tried it locked up during boot.

no problem , hm , so - good , with a fact of working
If the problem will apear again , then maybe You wrote here this error?:)

vesa rrc is a old version of vesa extension patch.
all definitions is in vesafb_modes.h ( also many of reviews on this forum ).
I`m still using vesa_rrc

greetings

[fallow]

One of Users in Polish forum reports bug in radeonfb.

this is a fix for it :

to patch - http://osdn.dl.sourceforge.net/sourcefo ... 2.diff.bz2
--or--
http://prdownloads.sourceforge.net/vivi ... 2?download

of course in sources dir :

Code: Select all

 # bzip2 -dc radfb_fix2.diff.bz2 | patch -p1 

greetings

[travisau]

Does anyone else have trouble emerging php? I've tried the CFQ and anticipatory scheduler so it doesn't appear to be related to that. I do have a reiser4 filesystem but don't know if its part of the problem or not.

>>> Install php-4.3.9 into /var/tmp/portage/php-4.3.9/image/ category dev-php
Installing PHP CLI binary: /var/tmp/portage/php-4.3.9/image//usr/bin/
Installing PHP CLI man page: /var/tmp/portage/php-4.3.9/image//usr/share/man/man1/
Installing shared extensions: /var/tmp/portage/php-4.3.9/image//usr/lib/php/extensions/no-debug-non-zts-20020429/
Installing PEAR environment: /var/tmp/portage/php-4.3.9/image//usr/lib/php/
[PEAR] Archive_Tar - installed: 1.1
make[1]: *** [install-pear-installer] Segmentation fault
make: *** [install-pear] Error 2

!!! ERROR: dev-php/php-4.3.9 failed.
!!! Function php-sapi_src_install, Line 525, Exitcode 2
!!! (no error message)
!!! If you need support, post the topmost build error, NOT this status messa

[gringo]

[PEAR] Archive_Tar - installed: 1.1
make[1]: *** [install-pear-installer] Segmentation fault
make: *** [install-pear] Error 2

!!! ERROR: dev-php/php-4.3.9 failed.
!!! Function php-sapi_src_install, Line 525, Exitcode 2
!!! (no error message)
!!! If you need support, post the topmost build error, NOT this status messa

this has nothing to do with reiser4, its gcc-3.4.3 related i think, just try with another compiler version.
BTW, great job with vivid fallow !!! Whats your opinion about patches included in -mjb, some of them look quite cool and are easy to include

cheers

[travisau]

Thanks for the quick reply! As gringo said it doesn't appear to be a vivid issue but has something to do with reiser4. See

http://gentoo-wiki.com/TIP_reiser4emergetip

[gringo]

http://gentoo-wiki.com/TIP_reiser4emergetip

oops, didnt know about this, thanks !
BUT it compiled here with 3.3.4 & reiser4

cheers

[fallow]

hey

BTW, great job with vivid fallow !!! Whats your opinion about patches included in -mjb, some of them look quite cool and are easy to include
cheers

thanks :)
yes You`re right,I forgot about mjb
latest is for 2.6.9 , 2.6.10-rc has many differencies to 269.
I`m going to add every one whos isn`t fixed yet in 2610rc in next vivid release
thanks for this tip

with php I have the same error with gcc 3.4.3 and 3.3.3
Like this from travisau
I have full reiser4 system
greetings

[travisau]

Hey guys..just an update on the php issue thing. I had a spare swap partition laying around on another drive so I mounted it up as /var/tmp/portage (formatted as an ext3). Absolutely nothing else changed on my system and now php merges fine. I think we've nailed this one down to a reiser4 issue.

Now I don't have any more complaints with vivid. I guess I'll have to upgrade my kernel now so I'll have something to work on!

[gringo]

I`m going to add every one whos isn`t fixed yet in 2610rc in next vivid release
thanks for this tip Smile

thanx to you for your hard work ! I try to always rollup my own patchset and actually im working a lot with mm as there is lot of interesting stuff in there.
Maybe you can add f.ex. kexec & dumpcrash in new releases, only a few easy rejects and seems to be quite stable now

with php I have the same error with gcc 3.4.3 and 3.3.3

that´s weird, as said it compiles fine here with 3.3.4 and i have reiser4 sata raids ... maybe im just in luck

cheers

[chickaroo]

star@Selphie ~ $ uptime
04:40:52 up 6 days, 13:20, 6 users, load average: 0.24, 0.84, 1.13

[fallow]

star@Selphie ~ $ uptime
04:40:52 up 6 days, 13:20, 6 users, load average: 0.24, 0.84, 1.13

nice

Code: Select all

20:38:17 up  4:37,  3 users,  load average: 2.15, 1.68, 1.35

hihi

I am working on vv_e3 now , I`m going to add mjb patches , kexec and I`m going to look on dumpcrash also @gringo

greetings

[gringo]

I`m going to add mjb patches , kexec and I`m going to look on dumpcrash also @gringo

hehe, you are fast

cheers

[fallow]

hm, I have some problems with kexec .

I`ve grab it from 2610rc2mm4 but it doesn`t work for me.
I have freeze , also with # halt when power down is calling.
I have this error always with every mm , since 268mm4 i think
but I was thinking all the time that this bug with mm is with ACPI hehe so , it`s kexec .
But it`s working for many people . I don`t know what is wrong with my cfg , so is kexec working for You Gringo ?

I think that adding it via USE flag will be good idea
for some my friends it`s working , for me no

greetings

[gringo]

I don`t know what is wrong with my cfg , so is kexec working for You Gringo ?

yes, it mostly works on my pentiumIV ( ht enabled, so smp) and works great on my other amd box.

Im using these ebuild + scripts:
http://bugs.gentoo.org/show_bug.cgi?id=50313

Hope that helps

cheers

[fallow]

ok, after some technical and non-technical problems hehe - new release based on 2.6.10-rc3 .

as usually ...main mission objectives :
* 2.6.10-rc3 based with O(1) CPU Sched with mm fixes
* O(1) CPU Sched control interface via /proc/sys/sched (mjb)
* fbsplash,vesa(tng/rrc) support
* partial mm patches,partial mjb patches,partial ck fixes
* reiser4 from latest mm with fixes
* bk drivers update and some (menu)config improvements
* kexec and crushdump from mm via USE flag="kex_cdump"

"vesa_rrc" - vesa rrc with default definitions for 1024x768@85Hz
"vesa_tng" - new version of vesa modes patch
"kex_cdump" - kernel exec and crushdump from mm

and nfo ( full version - http://osdn.dl.sourceforge.net/sourcefo ... nd_nfo.txt )

Code: Select all

--current release info--
CPU Scheduler : O(1) 2.6.10-rc3 based with partial mm patches/fixes
IO Scheduler(s) : switchable at runtime,default selectable from menuconfig or bootparams
--patches--
vivid_menu_2610rc3vve1.diff
[*] Preemptible Kernel (moved here for better visibility)                 
[*]   Preempt The Big Kernel Lock ( Ingo`s stuff ported by Con ) (NEW)    
    Default IO scheduler (CFQ v2 io scheduler + updates)  --->    
//for select default io scheduler as alternative to bootparams elevator=x that works also         
[ ] old ll_rw_blk version ( no high i/o load lkml fix ) (NEW)          
//for without high i/o load fix (lkml).Some people said that is better without it    
[*] [page_alloc] zone->pages_lots  = (zone->pages_min * 7) / 4; (NEW)    
//page_alloc:page_lots=((pages_min*7)/4) like in bk`s , or as in mwII *6 only.I don`t know what is better. 
[ ] CFLAGS hacking (NEW)   
//fixed by Troll some love stuff                                               
[*] /proc/mm support (NEW)     
//from SKAS patch.link for info with SKAS patch line.                
[ ] 1Gb Low Memory Support (CK) (NEW)                                     
(63) Maximum tty device number (NEW)                                      
    Kernel internal timer frequency (1000)  --->                          
[ ] Configurable USB HID Mouse Polling (CHMP) (NEW)                       
                                                            
--bk`s--
netdev1.patch.bz2
libata1-dev1.patch.bz2
bk-agpgart.patch
bk-alsa.patch
bk-ide-dev.patch
bk-ntfs.patch
--base--
2.6.10-rc2-mingoll.diff - Ingo`s low latency fixes ported by Con
2.6.10-rc2-nvidia-fix.diff - nvidia fix from ck
2.6.6-qsort-updated-lkml.patch - yet anohter qsort implementation from pld
2.6.6-xfs-qsort-lkml.patch - and pld xfs-qsort patch
2610rc3-fbsplash_temp.diff - fbsplash
cflags-selection.patch - fixed by Troll love stuff for easy apped own CFLAGS for kernel.
chmp-r4-2.6.9.patch - configurable hid mouse polling ;)
config-nr-tty-devices.patch - as in name :)
config_hz.diff - for setting kernel internal clock freq.
daconfig-2.1.1.patch - DaMouse "next generation" of menuconfigname ;)
cddvd-cmdfilter-drop.patch -  fix for non root user`s from ck.
acerhk.patch - acer keyboard support
gds_smbfs-dos-fix.patch - smbf dos fix from g-d-s
host-skas3-2.6.9-v7.patch - SKAS interesting patches ( http://user-mode-linux.sourceforge.net/skas.html )
skas-update-2.6.10.patch - update for 2610rc
lirc-2.6.9-20041024-px1.patch - fixed by Pax82 lirc patch.
lufs-2510rc_vivid.patch.bz2 - lufs fixed for 2610rc2
increase_prio0_scanning.diff + mwII.diff - Con`s vm work. 
nvidia_compat.diff -yet another nvidia fix from ck
squashfs2.0-patch - http://squashfs.sourceforge.net/
supermount-ng208-10rc2ck2.diff - tool for auto(u)mount removable media`s from ck
1g_lowmem1_i386.diff - as in name , from ck 
--mjb partial fixes-- info http://kerneltrap.org/book/view/2242
111-ppc64_reloc_hide.txt
220-numameminfo.txt
305-aiofix2.txt
310-percpu_real_loadavg.txt
320-irqbal_fast.txt
340-numa_mem_equals.txt
380-protocol254.txt
390-slabtune.txt
410-fasync_lock_rcu.txt
480-physnode_map.txt
500-sched_tunables.txt
--mm part--
reiser4_with_fixes.diff

sched-active_load_balance-fixlet.patch
sched-add-cond_resched_softirq.patch
sched-add_requeue_task.patch
sched-adjust_timeslice_granularity.patch
sched-alter_kthread_prio.patch
sched-can_migrate-exception-for-idle-cpus.patch
sched-fix-scheduling-latencies-in-vgaconc.patch
sched-more-agressive-wake_idle.patch
sched-newidle-fix.patch
sched-no-need-to-recalculate-rq.patch
sched-remove_interactive_credit.patch
sched-reset-cache_hot_time.patch
sched-use-cached-current-value.patch
CONFIG_SOUND_VIA82CXXX_PROCFS.patch
acpi-flush-tlb-when-pagetable-changed.patch
allow-admin-to-enable-only-some-of-the-magic-sysrq-functions.patch
assign-pkmap_base-dynamically.patch
assign_irq_vector-section-fix.patch
cancel_rearming_delayed_work.patch
cdromc-make-several-functions-static.patch
dont-hide-thread_group_leader-from-grep.patch
enforce-a-gap-between-heap-and-stack.patch
enhanced-i-o-accounting-data-patch.patch
ext3-cleanup-handling-of-aborted-transactions.patch
ext3-handle-attempted-delete-of-bitmap-blocks.patch
ext3-handle-attempted-double-delete-of-metadata.patch
ext3-use-generic_open_file-to-fix-possible-preemption-bugs.patch
file-sync-no-i_sem.patch
firestream-warnings.patch
fix-reiserfs-quota-debug-messages.patch
fork-total_forks-not-counted-under-tasklist_lock.patch
hold-bkl-for-shorter-period-in-generic_shutdown_super.patch
htree-telldir-fix.patch
ide_arch_obsolete_init-fix.patch
invalidate_inode_pages-mmap-coherency-fix.patch
invalidate_inodes-speedup.patch
iosched_def_sel.diff
kmap_atomic-fallout.patch
kmap_atomic-takes-char-fix.patch
kmap_atomic-takes-char.patch
make-sure-ioremap-only-tests-valid-addresses.patch
make-sysrq-f-call-oom_kill.patch
make-tree_lock-an-rwlock.patch
mempolicy-optimization.patch
minimal-ide-disk-updates.patch
oss-ac97-quirk-facility-fix.patch
oss-ac97-quirk-facility.patch
possible-dcache-bug-debugging-patch.patch
provide-a-filesystem-specific-syncable-page-bit-fix-2.patch
provide-a-filesystem-specific-syncable-page-bit-fix.patch
provide-a-filesystem-specific-syncable-page-bit.patch
rcu-eliminate-rcu_ctrlblklock.patch
rcu-eliminate-rcu_datalast_qsctr.patch
register_disk-hack-warnings.patch
register_disk-hack.patch
reiser3-cleanups.patch
reiserfs-cleaning-up-const-checks.patch
serialize-access-to-ide-devices.patch

ebuild , patches ,nfo & mini how-to/readme at
http://sourceforge.net/projects/vivid-sources

of course feedback,ideas,propositions,cooperations etc. etc.
greetings

[gringo]

new release based on 2.6.10-rc3

hehe, as said, youre fast !

Works nice here so far, but theres one thing i wanted to ask, maybe you know more about this, since 2.6.9 with staircase +cfq2 i get tons of this in my logs:

Code: Select all

 cfq: depth 4 reached, tagging now on

I read in ck ml that Jens said this was for debugging purposes or something but also that this should not appear anymore now.
It seems to be harmless anyway, but do you know something about this ?

cheers

[Genkaku]

Code: Select all

 CC      kernel/resource.o
  CC      kernel/sysctl.o
kernel/sysctl.c:455: error: parse error before '{' token
kernel/sysctl.c:91: warning: 'ngroups_max' defined but not used
make[1]: *** [kernel/sysctl.o] Error 1

Hmmm...