View previous topic :: View next topic |
Author |
Message |
GentooBox Veteran
Joined: 22 Jun 2003 Posts: 1168 Location: Denmark
|
|
Back to top |
|
|
noooob n00b
Joined: 15 Dec 2004 Posts: 5
|
Posted: Thu Dec 16, 2004 7:02 pm Post subject: |
|
|
Anyone have orinoco patched in it?. Or can you please put orinoco patch in it..please.. |
|
Back to top |
|
|
Martin Zoubek n00b
Joined: 08 Dec 2004 Posts: 6
|
Posted: Thu Dec 16, 2004 11:27 pm Post subject: |
|
|
Latency measurement, as I promised.
kernel: 2.6.10-rc3-vv_e1
jackd -R -d alsa -P -o 2 -p 16 -s
i.e. jack latency of 1/3 s, it spits all bigger delays it encounters to the console
tests: playing a video, with sound, of course - ~20% CPU usage, max. latency 2020 us, several times a second there is a latency of about 0.7-1.1 ms
Also some tests from latencytests: 'x11' and 'proc' - no increase in latency, forkbomb killed X
Result: with -p 16 is sound unlistenable, with -p 64 (1.33 ms jack latency) it may be usable, but not good, there will be some droupouts, -p 128 should be OK. But remember that there is only one program producing sound in my test, in real situations you will use several of them, that's what is jack designed for!
Conclusion: This kernel is probably not good for serious audio work.
Looking forward to test your new mm kernel |
|
Back to top |
|
|
vrln Guru
Joined: 11 Sep 2004 Posts: 534 Location: Finland
|
Posted: Fri Dec 17, 2004 12:20 am Post subject: |
|
|
Great patchset! Everything is running perfectly here, using 2.6.10-rc3-vve1. Here is a far fetched request but I'll say it anyway: I respect that you want to keep the patchset unique, so sorry for saying this, but staircase cpu scheduler would be nice as an option :) It has better desktop interactivity for me. (yes I know it's used in so many patchsets but it's not in any patchset that is based on 2.6.10-rc3+).
Entrance (the E17 login manager) kernel panics with 2.6.9 based kernels for me if I run it in opengl mode and press the restart or shutdown button.
also it's like you've made a new release every time I check these forums, keep up the good work :) |
|
Back to top |
|
|
fallow Bodhisattva
Joined: 08 Jan 2004 Posts: 2208 Location: Poland
|
Posted: Fri Dec 17, 2004 11:52 am Post subject: |
|
|
ok :
noooob wrote: | Anyone have orinoco patched in it?. Or can you please put orinoco patch in it..please.. |
done orinoco patch is at sourceforge and also the new available ebuild has it , for use it choose USE="orinoco" flag . also You can download 2610rc3vvmm_e1_orinoco.diff.bz2 from sf.net/projects/vivid-sources and do patching by hand in kernel_sources by : Code: | bzip2 -dc 2610rc3vvmm_e1_orinoco.diff.bz2 | patch -p1 |
release info is changed of course : https://forums.gentoo.org/viewtopic.php?p=1886355#1886355
Gentree wrote: | OK nice patch set, you've got a new convert!
I have been using 2.6.9-r1-nitro for about 4months but the more recent offerings seem less sucessful. I tried going back to love-sources but dev seems pretty much stalled over there and I could not even get the lastest one to boot.
So I started looking for alternatives and came across vivid and so far I like it.
Nothing has broken that was not broken already so signs are good.
Thanks for your efforts. |
Thanks - pretty nice You`re welcome . Some time ago I also was been a love user (Steel300 time etc. )
thanks , ebuild corrected - sorry for that mistake, and ~amd64 keyword added
Martin Zoubek wrote: |
Looking forward to test your new mm kernel
|
Thanks for You tests .I think for testing the _real_ latency it was better than sched2.c , so if You need serious audio work / small latence for that test new vvmm_e1 with Ingo`s Real Time Preempt , IT would be better I think it this case, Superb option is leveling of Preemption deep , I don`t need really realtime preempt for my desktop (I`m going to agree with bollucks that _really_ realtime preempt can slow down in some cases , so I`m choose Preemptible Kernel from menuconfig instead of Full Realtime Preempt ) I`m also doing some audio work but as amateur with Reason and similar I love experimental,drum`n`base,jungle,breakcore,acid jazz,trip hop and etc.etc.
vrln wrote: | Great patchset! Everything is running perfectly here, using 2.6.10-rc3-vve1. Here is a far fetched request but I'll say it anyway: I respect that you want to keep the patchset unique, so sorry for saying this, but staircase cpu scheduler would be nice as an option It has better desktop interactivity for me. (yes I know it's used in so many patchsets but it's not in any patchset that is based on 2.6.10-rc3+).
Entrance (the E17 login manager) kernel panics with 2.6.9 based kernels for me if I run it in opengl mode and press the restart or shutdown button.
also it's like you've made a new release every time I check these forums, keep up the good work |
thanks for Your opinion nice
about staircase, I was been staircase user since 5.x versions also. I think that staircase provided better interactivity than O(1) also.but since 2.6.10-rcX the situation is going to change. Many of mm stuff is going to RC 2610 , also partial latency fixes - in a fact of this is much better latency (small) in Nick Piggin`s sched2.c test than in the past.
I was moved to O(1) with partial / full mm fixes/speed-ups since 2.6.10-rcX because : I don`t feel difference beetwen 2.6.10 based on staircase and based on O(1) as I feel that in the past ( 2.6.9 for example ) . Also I have problems with staircase with java/flash applications and sound . Yes i know that is ISO mode with staircase via schedtool but in my feelings with 2.6.10 O(1) doing better
in the past we have one mm release with menuconfig select beetwen O(1) and staircase . I think when I have some time ( maybe Christmas ) then I can try to reproduce this, but it was not so easy
also If You will to do it , Yo can do patch with staircase for vivid ?
My personally fovorite at now is : O(1) + mm fixes.
cheers _________________ "Time is a companion that goes with us on a journey. It reminds us to cherish each moment, because it will never come again. What we leave behind is not as important as how we have lived" J-L. Picard |
|
Back to top |
|
|
noooob n00b
Joined: 15 Dec 2004 Posts: 5
|
Posted: Fri Dec 17, 2004 6:33 pm Post subject: |
|
|
fallow wrote: | ok :
noooob wrote: | Anyone have orinoco patched in it?. Or can you please put orinoco patch in it..please.. |
done orinoco patch is at sourceforge and also the new available ebuild has it , for use it choose USE="orinoco" flag . also You can download 2610rc3vvmm_e1_orinoco.diff.bz2 from sf.net/projects/vivid-sources and do patching by hand in kernel_sources by : Code: | bzip2 -dc 2610rc3vvmm_e1_orinoco.diff.bz2 | patch -p1 |
release info is changed of course : https://forums.gentoo.org/viewtopic.php?p=1886355#1886355
Gentree wrote: | OK nice patch set, you've got a new convert!
I have been using 2.6.9-r1-nitro for about 4months but the more recent offerings seem less sucessful. I tried going back to love-sources but dev seems pretty much stalled over there and I could not even get the lastest one to boot.
So I started looking for alternatives and came across vivid and so far I like it.
Nothing has broken that was not broken already so signs are good.
Thanks for your efforts. |
Thanks - pretty nice You`re welcome . Some time ago I also was been a love user (Steel300 time etc. )
thanks , ebuild corrected - sorry for that mistake, and ~amd64 keyword added
Martin Zoubek wrote: |
Looking forward to test your new mm kernel
|
Thanks for You tests .I think for testing the _real_ latency it was better than sched2.c , so if You need serious audio work / small latence for that test new vvmm_e1 with Ingo`s Real Time Preempt , IT would be better I think it this case, Superb option is leveling of Preemption deep , I don`t need really realtime preempt for my desktop (I`m going to agree with bollucks that _really_ realtime preempt can slow down in some cases , so I`m choose Preemptible Kernel from menuconfig instead of Full Realtime Preempt ) I`m also doing some audio work but as amateur with Reason and similar I love experimental,drum`n`base,jungle,breakcore,acid jazz,trip hop and etc.etc.
vrln wrote: | Great patchset! Everything is running perfectly here, using 2.6.10-rc3-vve1. Here is a far fetched request but I'll say it anyway: I respect that you want to keep the patchset unique, so sorry for saying this, but staircase cpu scheduler would be nice as an option It has better desktop interactivity for me. (yes I know it's used in so many patchsets but it's not in any patchset that is based on 2.6.10-rc3+).
Entrance (the E17 login manager) kernel panics with 2.6.9 based kernels for me if I run it in opengl mode and press the restart or shutdown button.
also it's like you've made a new release every time I check these forums, keep up the good work |
thanks for Your opinion nice
about staircase, I was been staircase user since 5.x versions also. I think that staircase provided better interactivity than O(1) also.but since 2.6.10-rcX the situation is going to change. Many of mm stuff is going to RC 2610 , also partial latency fixes - in a fact of this is much better latency (small) in Nick Piggin`s sched2.c test than in the past.
I was moved to O(1) with partial / full mm fixes/speed-ups since 2.6.10-rcX because : I don`t feel difference beetwen 2.6.10 based on staircase and based on O(1) as I feel that in the past ( 2.6.9 for example ) . Also I have problems with staircase with java/flash applications and sound . Yes i know that is ISO mode with staircase via schedtool but in my feelings with 2.6.10 O(1) doing better
in the past we have one mm release with menuconfig select beetwen O(1) and staircase . I think when I have some time ( maybe Christmas ) then I can try to reproduce this, but it was not so easy
also If You will to do it , Yo can do patch with staircase for vivid ?
My personally fovorite at now is : O(1) + mm fixes.
cheers |
Wohooo..thanks alot, for orinoco patch, on that kernel..running very well here. (now I am playing around with kismet,airsnort etc)...no problem whatsoever.
Tumbs up..keep up good work, from now on..I am using this kernel.
Cheers for that |
|
Back to top |
|
|
gringo Advocate
Joined: 27 Apr 2003 Posts: 3793
|
Posted: Fri Dec 17, 2004 7:33 pm Post subject: |
|
|
Quote: | staircase cpu scheduler would be nice as an option |
maybe you can have a look to plugsched, Con´s approach to this.
Didnt tried your latest patch (yet), but will let you know
cheers _________________ Error: Failing not supported by current locale |
|
Back to top |
|
|
NiLuJe Guru
Joined: 06 Nov 2004 Posts: 303 Location: Paris, France
|
Posted: Sat Dec 18, 2004 12:25 am Post subject: |
|
|
Works fine here Except that ati-drivers won't build with the vvm_e1 release .. (It worked with vv_e1)
Code: | CC [M] /var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.o
/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.c:311:22: drm_proc.h : Aucun fichier ou répertoire de ce type
/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.c: In function `firegl_stub_putminor':
/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.c:519: attention : « inter_module_put » est obsolète (déclaré à include/linux/module.h:577)
/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.c:521: attention : « inter_module_unregister » est obsolète (déclaré à include/linux/module.h:573)
/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.c: In function `firegl_stub_register':
/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.c:541: attention : « inter_module_register » est obsolète (déclaré à include/linux/module.h:572)
/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.c:572: attention : « inter_module_put » est obsolète (déclaré à include/linux/module.h:577)
/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.c: In function `firegl_init_module':
/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.c:613: erreur: « DRM_PROC_ENTRIES » non déclaré (première utilisation dans cette fonction)
/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.c:613: erreur: (Chaque identificateur non déclaré est rapporté une seule fois
/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.c:613: erreur: pour chaque fonction dans laquelle il apparaît.)
/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.c:619: erreur: « FGLDRM_proc_list » non déclaré (première utilisation dans cette fonction)
/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.c: In function `__ke_block_all_signals':
/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.c:815: attention : passage de l'argument n°1 de « block_all_signals » d'un type pointeur incompatible
/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.c: In function `__ke_get_vm_phys_addr':
/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.c:1597: attention : déclaration implicite de la fonction « pgd_offset_is_obsolete »
/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.c:1597: attention : affectation transforme un entier en pointeur sans transtypage
/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.c: In function `__ke_request_irq':
/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.c:1908: attention : affectation d'un type pointeur incompatible
/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.c: In function `do_vm_shm_nopage':
/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.c:2060: attention : affectation transforme un entier en pointeur sans transtypage
/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.c: In function `do_vm_pcie_nopage':
/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.c:2189: erreur: « drm_device_t » non déclaré (première utilisation dans cette fonction)
/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.c:2189: erreur: « dev » non déclaré (première utilisation dans cette fonction)
/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.c:2189: erreur: erreur d'analyse syntaxique avant un élément lexical « ) »
/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.c: In function `__ke_vm_phys_addr_str':
/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.c:2439: attention : affectation transforme un entier en pointeur sans transtypage
/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.c: At top level:
/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.c:2528: attention : initialisation d'un type pointeur incompatible
/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.c: In function `__ke_vm_map':
/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.c:2589: attention : déclaration implicite de la fonction « remap_page_range »
/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.c: In function `__ke_agp_uninit':
/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.c:3014: attention : « inter_module_put » est obsolète (déclaré à include/linux/module.h:577)
/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.c: At top level:
/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.c:301: attention : « FGLDRM_mem_info » défini mais n'a pas été utilisé
make[1]: *** [/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/firegl_public.o] Erreur 1
make: *** [_module_/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod] Erreur 2
make: Leaving directory `/usr/src/linux-2.6.10-rc3-vvmm_e1'
* DRM module not built |
(Oops, gcc warnings in french .. ) |
|
Back to top |
|
|
fallow Bodhisattva
Joined: 08 Jan 2004 Posts: 2208 Location: Poland
|
Posted: Sat Dec 18, 2004 9:14 am Post subject: |
|
|
noooob wrote: | Wohooo..thanks alot, for orinoco patch, on that kernel..running very well here. (now I am playing around with kismet,airsnort etc)...no problem whatsoever.
Tumbs up..keep up good work, from now on..I am using this kernel.
Cheers for that |
OK , nice - so orinoco will be in patchlists of future releases
and thx for good words
gringo wrote: | Quote: | staircase cpu scheduler would be nice as an option |
maybe you can have a look to plugsched, Con´s approach to this.
Didnt tried your latest patch (yet), but will let you know
cheers |
hihi Gringo - thanks for You efforts/ideas , I fortgot also about Con`s plugsched also like about mjb hehe.
Right - in a fact that we have 2 lines now, so mm release will have plugsched without much work as I think thx
NiLuJe wrote: | Works fine here Except that ati-drivers won't build with the vvm_e1 release .. (It worked with vv_e1)
|
hm, I really don`t have ati-card 2.6.10-rc3-vv_e1 works fine with nvidia 6111 and 6629 , and latest mm 2.6.10-rc3-vvmm_e1 works with 6629 from portage here. ( I think it also works with 6111 but maybe pgd_patches will be required or 4levels).
I will take a look in the near future days but I even never merged it ( ati-drv`s )
btw.maybe alternative ati drivers - something like ati-gatos if I remember the name correct will work ...
cheers _________________ "Time is a companion that goes with us on a journey. It reminds us to cherish each moment, because it will never come again. What we leave behind is not as important as how we have lived" J-L. Picard |
|
Back to top |
|
|
Martin Zoubek n00b
Joined: 08 Dec 2004 Posts: 6
|
Posted: Sat Dec 18, 2004 5:09 pm Post subject: |
|
|
NiLuJe wrote: | Works fine here Except that ati-drivers won't build with the vvm_e1 release .. (It worked with vv_e1)
|
Same problem, fglrx won't build. Nvidia is OK.
I've encountered another problem which was not in vv_e1 - loading cmd64x module locks system. That's pretty important to me, I've root partition on disk connected to card with this chipset Any solution? |
|
Back to top |
|
|
fallow Bodhisattva
Joined: 08 Jan 2004 Posts: 2208 Location: Poland
|
Posted: Sat Dec 18, 2004 6:24 pm Post subject: |
|
|
Martin Zoubek wrote: | NiLuJe wrote: | Works fine here Except that ati-drivers won't build with the vvm_e1 release .. (It worked with vv_e1)
|
Same problem, fglrx won't build. Nvidia is OK.
I've encountered another problem which was not in vv_e1 - loading cmd64x module locks system. That's pretty important to me, I've root partition on disk connected to card with this chipset Any solution? |
hm , vvmm_e1 is full mm based - mm is experimental so is high risk to broke something, vv_e1 with partial mm was done exectly for that , to eleminate risc of broking . that is the reason of 2 vivid lines partial and full mm based.
I chek lkml to find something about broking cmd64* with mm1 , but with no results...
I compiled it as module and loaded via modprobe - ok- everything is fine so i think is direct mm problem, but I don`t have this controller.
Also I checked mm1 diff for changes in cmd6* file - no results .
So mm1 doesn`t change directly the driver , It must be some indirect related to cmd6* but I don`t know what.
There is no bk-pci but maybe You try to reverse for example patches in relation with cmd6* bk-ide-dev for example
about ati with vivid mm based : I don`t have any experience because i never had ati card , I will take a look today .
cheers _________________ "Time is a companion that goes with us on a journey. It reminds us to cherish each moment, because it will never come again. What we leave behind is not as important as how we have lived" J-L. Picard |
|
Back to top |
|
|
NiLuJe Guru
Joined: 06 Nov 2004 Posts: 303 Location: Paris, France
|
Posted: Sat Dec 18, 2004 7:31 pm Post subject: |
|
|
Thanks |
|
Back to top |
|
|
fallow Bodhisattva
Joined: 08 Jan 2004 Posts: 2208 Location: Poland
|
Posted: Sat Dec 18, 2004 9:06 pm Post subject: |
|
|
about ati drivers and full mm vivid - 2.6.10-rc3-vvmm_e1
as in partial mm vivid everything is working hehe , in full mm based not "normal full mm problems"
after reversing mm`s bk-drm.patch ati-drivers was compiled.
If somebody doesn`t know how to do that ( better to say:) )
to reversing it download from http://sourceforge.net/projects/vivid-sources -> 2.6.10-rc3-vvmm_e1 -> bk-drm.patch.bz2
and in kernel_source_dir do :
Code: | # bzip2 -dc bk-drm.patch.bz2 | patch -p1 -R |
but also 2 warnings :
Code: |
Building modules, stage 2.
MODPOST
*** Warning: "remap_page_range" [/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/fglrx.ko] undefined!
*** Warning: "pgd_offset_is_obsolete" [/var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/fglrx.ko] undefined!
CC /var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/fglrx.mod.o
LD [M] /var/tmp/portage/ati-drivers-3.14.6/work/lib/modules/fglrx/build_mod/fglrx.ko
make: Leaving directory `/usr/src/linux-2.6.10-rc3-vvmm_e1'
|
so seems that ati-drivers requires fixes similar to the nvidia drivers with remap_page_range from 6111 and pgd_offset from 6629 ...
I try to do something with it , anyone to try fix ati-drivers also ?
cheers _________________ "Time is a companion that goes with us on a journey. It reminds us to cherish each moment, because it will never come again. What we leave behind is not as important as how we have lived" J-L. Picard |
|
Back to top |
|
|
Macumba Tux's lil' helper
Joined: 13 Sep 2004 Posts: 112
|
Posted: Sat Dec 18, 2004 10:28 pm Post subject: |
|
|
i got this error when i make make menuconfig, or s.th like this (make config, gconfig, xconfig...)
Code: |
scripts/kconfig/conf arch/x86_64/Kconfig
file lib/Kconfig.RT already scanned?
make[1]: *** [config] Error 1
make: *** [config] Error 2
|
edit
THANKS!! _________________
Code: | dd if=/dev/urandom of=/dev/speech |
ogg - mp3
Last edited by Macumba on Sat Dec 18, 2004 11:24 pm; edited 1 time in total |
|
Back to top |
|
|
fallow Bodhisattva
Joined: 08 Jan 2004 Posts: 2208 Location: Poland
|
Posted: Sat Dec 18, 2004 11:05 pm Post subject: |
|
|
Macumba wrote: | i got this error when i make make menuconfig, or s.th like this (make config, gconfig, xconfig...)
Code: |
scripts/kconfig/conf arch/x86_64/Kconfig
file lib/Kconfig.RT already scanned?
make[1]: *** [config] Error 1
make: *** [config] Error 2
|
|
sorry my mistake
for x86_64 please use this patch : x86_64_config_fix1.diff from http://sourceforge.net/projects/vivid-sources -> 2.6.10-rc3-vvmm_e1 or to do this by hand. It`s one line error .
Code: |
diff -Naur linux-2.6.10-rc3-vvmm_e1/arch/x86_64/Kconfig t2/arch/x86_64/Kconfig
--- linux-2.6.10-rc3-vvmm_e1/arch/x86_64/Kconfig 2004-12-17 12:19:57.000000000 +0100
+++ t2/arch/x86_64/Kconfig 2004-12-18 23:56:53.164007664 +0100
@@ -236,7 +236,7 @@
cost of slightly increased overhead in some places. If unsure say
N here.
-source "lib/Kconfig.RT"
+
config RWSEM_GENERIC_SPINLOCK
bool
|
You must delete thins line -> source "lib/Kconfig.RT" from arch/x86_64/Kconfig
sorry and cheers.
[EDIT1]
ebuild corrected -> new ebuild -> 2610rc3vvmm_e1-ebuild_latest3.tar.bz2 on sf.net
[/EDIT1] _________________ "Time is a companion that goes with us on a journey. It reminds us to cherish each moment, because it will never come again. What we leave behind is not as important as how we have lived" J-L. Picard |
|
Back to top |
|
|
Gentree Watchman
Joined: 01 Jul 2003 Posts: 5350 Location: France, Old Europe
|
Posted: Sun Dec 19, 2004 1:53 pm Post subject: |
|
|
I've no time to get upto speed on patching and reverse patching but I have two machines here I;ve just put 2.6.10-rc3-vv_e1 on .
One is GeForce ti 4200 (nvidia-kernel-1.0.6111-r3 ) the other is ATI radeon VE running the kernel ati drivers.
Both work fine . (the ati model is not supported by flgxcccccc\xz)
However, I just tried a quick test playing an audio CD and recording a 2s sample with audacity. The sound was , frankly shit.
You could tell what it was but you would not want to listen to it !
Now I just rebooted to 2.6.9-nitro1 and at a first hearing the sound was clean.
I'm sure with all the scheduler options and stuff on the vv kernel I can get someting set up way better if someone can offer some pointers.
I'm not looking to make a 32 track recording studio but if I can get a decent stereo snip of a CD and a clean live recording I will be chuffed.
TIA. _________________ Linux, because I'd rather own a free OS than steal one that's not worth paying for.
Gentoo because I'm a masochist
AthlonXP-M on A7N8X. Portage ~x86 |
|
Back to top |
|
|
rommel Veteran
Joined: 19 Apr 2002 Posts: 1145 Location: Williamsburg Virginia
|
Posted: Sun Dec 19, 2004 3:03 pm Post subject: |
|
|
i am still unable to get sound to work correctly. strange that the module isnt seen but can be modprobed... the problem is the same with strait mm or with either vv kernel on source forge. modprobing the driver gives me a card in proc but no device in gnome. this and the bonnie++ scores for vv as compared to nitro4 were aweful. and Gentree i am not blaming vv for anything but was asking in my previous post only if someone else had had a similar issue. i havent tried with with a strait vanilla 2.6.10 kernel since i am using reiser4 on both installs. |
|
Back to top |
|
|
fallow Bodhisattva
Joined: 08 Jan 2004 Posts: 2208 Location: Poland
|
Posted: Sun Dec 19, 2004 4:44 pm Post subject: |
|
|
Gentree : I`m using 2.6.10-rc3-vv_e1 . I try to do that via audacity and everythink is ok here. Also I ask my friend Troll(vivid-stable maintainer) to do that, He said that also sound was ok.
We both have emu10k1 based msx cards.
nitro1 is 2.6.9 based , vv_e1 is 2.6.10-rc3 based , maybe in vanilla-rc is a some bug.Propably You don`t have proper time but testing pure vanilla 2.6.10-rc3 will be a good solution to localize error.
Rommel : I really don`t know - sorry . with emu10k1 it works correctly also Gentree said in some past post that is ok with loading modules ...hm , what sound card do You have via82* ?
Code: | CONFIG_SOUND_VIA82CXXX_PROCFS.patch | this is the one patch in vv_e1 for sound only, You can Try reverse it maybe is it.
this patch is in mm broken-out -> http://www.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.10-rc3/2.6.10-rc3-mm1/2.6.10-rc3-mm1-broken-out.tar.bz2
Code: | patch -p1 -R < patch_name | in kernel_src for reversing of course.
Rommel , Can You post here bonnie++ result of Your comparison ?
it`s may be interesting in casue of 2.6.9 and 2.6.10-rc based kernels. Did You use the same IO schedulers in both ?
however , if You like nitro better and everything You want You have in it then imho is better to stay with nitro
I don`t want to enforce abybody to vivid hihi
and , in addition vv_e1 is rc based so it`s more propability that something will be broken - this is release candidate and nitro is based on stable 2.6.9 , but don`t have some 2.6.10-rcX stuff .
Everytning has + and - hehe:)
I really like partial mm stuff like vv_e1 , full mm-based is too much unstable and error-risky for me , but from time to time...
cheers _________________ "Time is a companion that goes with us on a journey. It reminds us to cherish each moment, because it will never come again. What we leave behind is not as important as how we have lived" J-L. Picard |
|
Back to top |
|
|
Gentree Watchman
Joined: 01 Jul 2003 Posts: 5350 Location: France, Old Europe
|
Posted: Sun Dec 19, 2004 6:12 pm Post subject: |
|
|
rommel wrote: | strange that the module isnt seen but can be modprobed... the problem is the same with strait mm or with either vv kernel on source forge. modprobing the driver gives me a card in proc but no device in gnome. this and the bonnie++ scores for vv as compared to nitro4 were aweful. and Gentree i am not blaming vv for anything but was asking in my previous post only if someone else had had a similar issue. i havent tried with with a strait vanilla 2.6.10 kernel since i am using reiser4 on both installs. |
I'm on R4 as well. I dont think your problem is kernel related. The module is there since you can modprobe it.
It sounds to me like there is a fault in your alsa config. What do you have configured that you expect will cause the alsasound init script to load that specific module?
HTH _________________ Linux, because I'd rather own a free OS than steal one that's not worth paying for.
Gentoo because I'm a masochist
AthlonXP-M on A7N8X. Portage ~x86 |
|
Back to top |
|
|
Gentree Watchman
Joined: 01 Jul 2003 Posts: 5350 Location: France, Old Europe
|
Posted: Sun Dec 19, 2004 6:15 pm Post subject: |
|
|
Quote: | however , if You like nitro better and everything You want You have with it then imho is better to stay with nitro |
well I was not going to ditch vv before giving it a fair go!
I felt that there was probably something I could set up better to reduce the latency, but maybe not.
I get the impression that nitro is stalling a bit and I had some cd burn issues remaining with the 2.9.1 on which it was based. That's why I thought I would test out vv, after all it is mm (partial) based.
If someone invovled in developing this patch set tells me I'm not likely to get anything better I'll stop digging, but I'd like to give it a fair test.
_________________ Linux, because I'd rather own a free OS than steal one that's not worth paying for.
Gentoo because I'm a masochist
AthlonXP-M on A7N8X. Portage ~x86 |
|
Back to top |
|
|
rommel Veteran
Joined: 19 Apr 2002 Posts: 1145 Location: Williamsburg Virginia
|
Posted: Sun Dec 19, 2004 6:51 pm Post subject: |
|
|
hey Fallow... well i am pretty sure that the patch for via82xx is for oss... i use alsa... here are the bonnie++ for the latest nitro and vv.
Code: |
2.6.9-nitro4
# bonnie++ -d /foo/ -s 4096 -r 2048 -u root
Using uid:0, gid:0.
Writing a byte at a time...done
Writing intelligently...done
Rewriting...done
Reading a byte at a time...done
Reading intelligently...done
start 'em...done...done...done...done...done...
Create files in sequential order...done.
Stat files in sequential order...done.
Delete files in sequential order...done.
Create files in random order...done.
Stat files in random order...done.
Delete files in random order...done.
Version 1.93c ------Sequential Output------ --Sequential Input- --Random-
Concurrency 1 -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP
JaiBaba 4G 235 88 15381 21 17763 5 535 99 50945 6 196.6 10
Latency 63115us 85420ms 46072ms 52295us 99348us 3202ms
Version 1.93c ------Sequential Create------ --------Random Create--------
JaiBaba -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
files /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
16 30272 52 +++++ +++ 20549 99 27118 94 +++++ +++ 21314 98
Latency 11333us 2947us 2948us 103us 93us 138us
2.6.10-rc3-mmvv_e1
# bonnie++ -d /foo/ -s 4096 -r 2048 -u root
Using uid:0, gid:0.
Writing a byte at a time...done
Writing intelligently...done
Rewriting...done
Reading a byte at a time...done
Reading intelligently...done
start 'em...done...done...done...done...done...
Create files in sequential order...done.
Stat files in sequential order...done.
Delete files in sequential order...done.
Create files in random order...done.
Stat files in random order...done.
Delete files in random order...done.
Version 1.93c ------Sequential Output------ --Sequential Input- --Random-
Concurrency 1 -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP
JaiBaba 4G 201 93 13914 2 16539 6 306 97 30974 8 167.3 7
Latency 73864us 128s 11185ms 55976us 25152us 4506ms
Version 1.93c ------Sequential Create------ --------Random Create--------
JaiBaba -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
files /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
16 +++++ +++ +++++ +++ 20879 98 28600 91 +++++ +++ 21636 98
Latency 163us 2712us 2743us 92us 29us 136us
|
Last edited by rommel on Sun Dec 19, 2004 8:54 pm; edited 1 time in total |
|
Back to top |
|
|
Gentree Watchman
Joined: 01 Jul 2003 Posts: 5350 Location: France, Old Europe
|
Posted: Sun Dec 19, 2004 7:36 pm Post subject: |
|
|
could we split the last line of those code blocks , they're make the thread a bit unreadable. _________________ Linux, because I'd rather own a free OS than steal one that's not worth paying for.
Gentoo because I'm a masochist
AthlonXP-M on A7N8X. Portage ~x86 |
|
Back to top |
|
|
fallow Bodhisattva
Joined: 08 Jan 2004 Posts: 2208 Location: Poland
|
Posted: Sun Dec 19, 2004 10:35 pm Post subject: |
|
|
Thank You for Your effort Rommel
ok - split
Code: |
Version 1.93c ------Sequential Output------ --Sequential Input- --Random-
Concurrency 1 -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP
JaiBaba-nitro 4G 235 88 15381 21 17763 5 535 99 50945 6 196.6 10
JaiBaba-vvmm 4G 201 93 13914 2 16539 6 306 97 30974 8 167.3 7
Latency-nitro 63115us 85420ms 46072ms 52295us 99348us 3202ms
Latency-vvmm 73864us 128s 11185ms 55976us 25152us 4506ms
Version 1.93c ------Sequential Create------ --------Random Create--------
JaiBaba -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
files /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
nitro 16 30272 52 +++++ +++ 20549 99 27118 94 +++++ +++ 21314 98
vvmm 16 +++++ +++ +++++ +++ 20879 98 28600 91 +++++ +++ 21636 98
Latency-nitro 11333us 2947us 2948us 103us 93us 138us
Latency-vivid 163us 2712us 2743us 92us 29us 136us
|
so in first segment nitro is better ( sequential input/output )
but in second vvmm is better ( sequential create / random create )
also imho this test is not really correct as my sched latency test because : is only one repeat . reiser4 has random big latencies in acces to the i/o subsystem so it`s depend randomly.
nitro is better in first phase , vvmm is better in second phase.
also what I/O Scheduler did You use for this test ?
it`s interesting also what would be results of vv_e not vvmm_e
Gentree wrote: |
If someone invovled in developing this patch set tells me I'm not likely to get anything better I'll stop digging, but I'd like to give it a fair test.
|
I`m happy of fact that You test vv it`s good
with my and Troll`s emu10k1 there is no problem with sampling on audacity, I think that is 2 possibilities :
- the drivers in 2.6.10-rc3 for You card has some bugs.
- hm....the drivers has some bugs ( hehe) I like monthy python
- or bonus third option - maybe the drivers has some bugs ?
I`m really don`t know - there is no alsa patches in vv_e1
maybe with 2.6.10-rc4 or I hope that with 2.6.10-stable will be ok
cheers _________________ "Time is a companion that goes with us on a journey. It reminds us to cherish each moment, because it will never come again. What we leave behind is not as important as how we have lived" J-L. Picard |
|
Back to top |
|
|
Gentree Watchman
Joined: 01 Jul 2003 Posts: 5350 Location: France, Old Europe
|
Posted: Mon Dec 20, 2004 3:18 am Post subject: |
|
|
Quote: | with my and Troll`s emu10k1 there is no problem with sampling on audacity, I think that is 2 possibilities :- the drivers in 2.6.10-rc3 for You card has some bugs. |
That's odd because the system that gave me such poor sound also runs audigy (firewire) sound card.
The good news : I tried the same thing on my main box with an Ensoniq 128 PCI and the sound was clean with both kernels.
The systems are basically fairly similar , the major difference being the problem one has an Radeon card the other a GeForce.
I'll go back to verify the ATI based system now.
Hopefully I will be able to use the vv kernel. _________________ Linux, because I'd rather own a free OS than steal one that's not worth paying for.
Gentoo because I'm a masochist
AthlonXP-M on A7N8X. Portage ~x86 |
|
Back to top |
|
|
fallow Bodhisattva
Joined: 08 Jan 2004 Posts: 2208 Location: Poland
|
Posted: Mon Dec 20, 2004 3:20 pm Post subject: |
|
|
Gentree wrote: |
Hopefully I will be able to use the vv kernel. |
nice I`m waiting for 2.6.10-rc4 ( bk13 today ) to do next vv_e release
I wad done bonnie++ tests again on my machine :
AMD Athlon 1.0 , 256 MB Ram, HDD ST3120026A , VIA KT133
kernels : 2.6.9-nitro4, 2.6.10-rc3-vv_e1 , 2.6.10-rc3-vvmm_e1
every test has 2 repeations ( to see randomization of latencies,reiser4 especially) t1 - first repeat , t2 - second repeat.
so it`s easy to see that results of first and second test for the same kernel can be much different hehe
selected I/O Scheduler in all kernels : Deadline
results here : text formated with Opera7.60p4
Code: |
Version 1.93c ------Sequential Output------ --Sequential Input- --Random-
Concurrency 1 -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP
nitro-t1 2G 83 98 28119 28 13343 12 188 99 31990 17 129.9 7
nitro-t2 2G 82 98 29816 20 13202 13 192 99 32522 15 128.5 7
vv_e1-t1 2G 81 97 26637 19 12873 18 149 99 31427 19 123.6 7
vv_e1-t2 2G 78 97 26555 22 12899 19 148 99 31066 19 121.9 7
vvmm_e1-t1 2G 74 97 27244 20 17004 23 156 98 20545 31 106.0 8
vvmm_e1-t2 2G 81 97 28423 26 16958 22 169 98 20526 31 104.5 8
nitro-t1 158ms 11370ms 2638ms 48835us 57986us 3692ms
nitro-t2 162ms 4340ms 4828ms 51951us 59802us 3952ms
vv_e1-t1 181ms 10241ms 3724ms 66494us 47453us 3937ms
vv_e1-t2 182ms 15744ms 2953ms 67635us 62251us 3586ms
vvmm_e1-t1 218ms 8854ms 4560ms 78992us 23805us 4356ms
vvmm_e1-t2 185ms 9475ms 4659ms 77441us 20931us 4036ms
--------------------------Sequential Create---------------Random Create--------
Enterprise -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
files /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
nitro-t1 16 14455 79 +++++ +++ 7426 99 7975 96 +++++ +++ 7416 99
nitro-t2 16 14956 83 +++++ +++ 7364 99 7943 96 +++++ +++ 7529 99
vv_e1-t1 16 15672 77 +++++ +++ 7532 99 8069 96 +++++ +++ 7740 99
vv_e1-t2 16 15609 78 +++++ +++ 7606 99 8096 96 +++++ +++ 7700 99
vvmm_e1-t1 16 15958 76 +++++ +++ 7205 99 7900 96 +++++ +++ 7390 99
vvmm_e1-t2 16 16286 75 +++++ +++ 7281 99 8097 96 +++++ +++ 7518 99
nitro-t1 8627us 7945us 8256us 10059us 27us 343us
nitro-t2 391us 8841us 8822us 6592us 23us 349us
vv_e1-t1 306us 8347us 8460us 8036us 25us 344us
vv_e1-t2 301us 8205us 8307us 6508us 26us 349us
vvmm_e1-t1 648us 8102us 8393us 8523us 21us 351us
vvmm_e1-t2 205us 7645us 8023us 13561us 22us 356us
|
cheers _________________ "Time is a companion that goes with us on a journey. It reminds us to cherish each moment, because it will never come again. What we leave behind is not as important as how we have lived" J-L. Picard |
|
Back to top |
|
|
|
|
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
|
|