| View previous topic :: View next topic |
| Author |
Message |
VolcomPimp
Tux's lil' helper
Joined: 11 Mar 2004
Posts: 135
Location: Newark, CA
|
 Posted: Wed Apr 28, 2004 10:52 pm Post subject: |
 |
|
| how does win4lin work? |
|
| Back to top |
|
 |
nmcsween
Guru
Joined: 12 Nov 2003
Posts: 381
|
| Posted: Wed Apr 28, 2004 11:01 pm Post subject: |
|
|
| Just wondering steel300 what does the high res timers ( 1000hz, 500hz, etc ) effect? |
|
| Back to top |
|
|
nephros
Advocate
Joined: 07 Feb 2003
Posts: 2139
Location: Graz, Austria (Europe - no kangaroos.)
|
| Posted: Wed Apr 28, 2004 11:14 pm Post subject: |
|
|
| Ultraoctane.com wrote: | | Just wondering steel300 what does the high res timers ( 1000hz, 500hz, etc ) effect? |
As I understand it those are necessary for processes which require very accurate time measurement like some embedded applications. The standard kernel timers have not been fine-grained enough for those applications. I doubt there is any real use for desktop or server systems but I could of course be wrong, because other than this I have no clue what I'm talking about.
Anyway, I'm very happy that this is a NickSched release, because the nickless releases have always felt worse to me. Mouse and audio stuttering under load (make -j4) which never occured with "ye good olde" love patches.
Thanks Steel300, and perhaps you could keep up the scheduler switching every couple of releases?
EDIT: "foo->for" 
_________________
Please put [SOLVED] in your topic if you are a moron.
Last edited by nephros on Wed Apr 28, 2004 11:23 pm; edited 2 times in total |
|
| Back to top |
|
|
steel300
Veteran
Joined: 10 Jul 2003
Posts: 1155
|
| Posted: Wed Apr 28, 2004 11:17 pm Post subject: |
|
|
| nephros wrote: | | Ultraoctane.com wrote: | | Just wondering steel300 what does the high res timers ( 1000hz, 500hz, etc ) effect? |
As I understand it those are necessary foo processes which require very accurate time measurement like some embedded applications. The standard kernel timers have not been fine-grained enough for those applications. I doubt there is any real use for desktop of server systems but I could of course be wrong, because other than this I have no clue what I'm talking about. |
You got the Hi Res Timers part correct.
_________________
Rationality is the recognition of the fact that nothing can alter the truth and nothing can take precedence over that act of perceiving it. |
|
| Back to top |
|
|
Nutterpc
Tux's lil' helper
Joined: 02 Feb 2004
Posts: 83
|
| Posted: Wed Apr 28, 2004 11:22 pm Post subject: |
|
|
I've so far been keeping the nforce users optimised kernel config stuff up to date as I said I would steel, followin the releases 
Already got one for the latest, which has been smooth for me. No probs at all
_________________
If it isn't broke, you ain't tweaked it right
Registered Linux User 353232 |
|
| Back to top |
|
|
enzobelmont
Guru
Joined: 06 Apr 2004
Posts: 345
Location: Chiapas, Mexico
|
| Posted: Thu Apr 29, 2004 4:08 am Post subject: |
|
|
| steel300 wrote: | | lucida wrote: | Failed to compile when the win4lin support is enabled  |
Grrr. |
sorry man... it failed to compile when win4lin is enabled
| Code: |
arch/i386/kernel/signal.c: En la funci?n `do_notify_resume':
arch/i386/kernel/signal.c:644: `_TIF_MKI_RETUSER' no ha sido declarado aqu? (primero ?selo en esta funci?n)
arch/i386/kernel/signal.c:644: (Cada identificador no declarado solamente se reporta una vez
arch/i386/kernel/signal.c:644: para cada funcion en la que aparece.)
make[1]: *** [arch/i386/kernel/signal.o] Error 1
make: *** [arch/i386/kernel] Error 2
|
sorry again it is in spanish |
|
| Back to top |
|
|
not_registered
Tux's lil' helper
Joined: 04 Feb 2003
Posts: 148
|
| Posted: Thu Apr 29, 2004 5:04 am Post subject: |
|
|
AS is amazing compared to CFQ. Amazing.
_________________
It's Floam, it's Floam. It's flying foam! |
|
| Back to top |
|
|
aethyr
Veteran
Joined: 06 Apr 2003
Posts: 1085
Location: NYC
|
| Posted: Thu Apr 29, 2004 7:12 am Post subject: |
|
|
| not_registered wrote: | | AS is amazing compared to CFQ. Amazing. |
umm... you do know that AS is the default scheduler in the vanilla kernel?
CFQ is the "new" scheduler. |
|
| Back to top |
|
|
Martigen
n00b
Joined: 04 Mar 2003
Posts: 59
|
| Posted: Thu Apr 29, 2004 4:26 pm Post subject: |
|
|
Regarding the new and alternate nforce fix: Prakesh's C1halt code seems less stable to me than the apic_tack fix. I've run with the new fix and got a total system lockup twice in an hour (and yes am passing 'idle=C1halt' to the kernel). Manually appying the apic_tack patch and passing apic_tack=2 I haven't had a freeze in the past day.
Could you return to the apic fix or include both (I think Prakesh said both could be applied)? |
|
| Back to top |
|
|
Ylin
n00b
Joined: 10 Apr 2004
Posts: 27
|
| Posted: Thu Apr 29, 2004 4:55 pm Post subject: |
|
|
| Martigen wrote: | Regarding the new and alternate nforce fix: Prakesh's C1halt code seems less stable to me than the apic_tack fix. I've run with the new fix and got a total system lockup twice in an hour (and yes am passing 'idle=C1halt' to the kernel). Manually appying the apic_tack patch and passing apic_tack=2 I haven't had a freeze in the past day.
Could you return to the apic fix or include both (I think Prakesh said both could be applied)? |
Just wanted to say the same.
"hdparm -Tt /dev/hdx" is a good stability test I think. And that C1halt patch always locks my system the second time I do that command. With the apic_tack patch I didn't experience any hardlock. So I agree with Martigen, please return to that apic_tack patch.
And yes, he said that both patches can be applied at the same time. |
|
| Back to top |
|
|
GoinBald
n00b
Joined: 19 Apr 2003
Posts: 39
Location: Canada
|
| Posted: Thu Apr 29, 2004 5:20 pm Post subject: |
|
|
Hello,
First Thanks
Fantastic patchset!
Was running her but I lose my 6th mouse button
With regards to the nforce patch do u have 1 of these mobos?
Albatron KM18G Pro
DFI NFII 400-AL
Epox 8RGA+
Shuttle AN35N
If not update the bios and you're set.
Baldy |
|
| Back to top |
|
|
boroshan
l33t
Joined: 16 Apr 2003
Posts: 730
Location: upside down
|
| Posted: Thu Apr 29, 2004 5:37 pm Post subject: good stuff |
|
|
Nvidia works, bootsplash works, seems smooth and responsive
Nice work. Thanks, Steel!
_________________
Don't let THEM immanentize the Eschaton! |
|
| Back to top |
|
|
dvc5
Guru
Joined: 06 Dec 2003
Posts: 433
Location: Sunnyvale, California
|
| Posted: Thu Apr 29, 2004 5:42 pm Post subject: |
|
|
| Code: | loz@thinkloz loz $ uname -a
Linux thinkloz.org 2.6.6-rc3-love1 #1 Thu Apr 29 10:32:03 PDT 2004 i686 Intel(R) Celeron(TM) CPU 1066MHz GenuineIntel GNU/Linux
loz@thinkloz loz $ |
Thanks steel!
_________________
#define NULL rand() /*heh heh heh */
Green Is Good |
|
| Back to top |
|
|
sawanv
Apprentice
Joined: 05 Dec 2002
Posts: 238
|
| Posted: Thu Apr 29, 2004 6:04 pm Post subject: |
|
|
Wooo...thanks steel300.
Any chance of getting the io-select patch in on 2.6.6-love1?
Sawan
_________________
Eir kahen "chalo ghar jayen",
Bir kahen "chalo ghar jayen",
Phatte kahen "chalo ghar jayen",
Cartman kahen....."Screw you guys, I am going home !!! " |
|
| Back to top |
|
|
steel300
Veteran
Joined: 10 Jul 2003
Posts: 1155
|
| Posted: Thu Apr 29, 2004 7:02 pm Post subject: |
|
|
| sawanv wrote: | Wooo...thanks steel300.
Any chance of getting the io-select patch in on 2.6.6-love1?
Sawan |
I'm working on it. It segfaulted USB things. So it needs some work. Nick and I are racing to see who can fix it first.
_________________
Rationality is the recognition of the fact that nothing can alter the truth and nothing can take precedence over that act of perceiving it. |
|
| Back to top |
|
|
PrakashP
Veteran
Joined: 27 Oct 2003
Posts: 1249
Location: C.C.A.A., Germania
|
| Posted: Thu Apr 29, 2004 7:27 pm Post subject: |
|
|
@nforce trouble:
Could you please do a "dmesg|grep halt"? Does then "using C1 halt disconnect friendly idle threads." appear? If yes, you need to report to Ross Dickson (just post at lkml), so he gets in contact with to update the patch. I am now using the C1halt patch for a couple of days and system runs fine (and of course survives hdparm -t).
BTW, was the patch stable for you with past love kernel? I mean _I don't know whether Len's patch (which remaps timer irq to APIC, check with cat /proc/interrupts) is in vanilla kernel. |
|
| Back to top |
|
|
jonnii
n00b
Joined: 10 Feb 2004
Posts: 33
|
| Posted: Thu Apr 29, 2004 8:16 pm Post subject: |
|
|
wait, so this kernel cant be used with resiserfs or what?
i dont get it. |
|
| Back to top |
|
|
Rylan
n00b
Joined: 08 May 2002
Posts: 41
|
| Posted: Thu Apr 29, 2004 8:28 pm Post subject: |
|
|
| Does the lirc patch have lirc_i2c? I need that for my hauppauge card. |
|
| Back to top |
|
|
dvc5
Guru
Joined: 06 Dec 2003
Posts: 433
Location: Sunnyvale, California
|
| Posted: Thu Apr 29, 2004 8:32 pm Post subject: |
|
|
| jonnii wrote: | wait, so this kernel cant be used with resiserfs or what?
i dont get it. |
I think it's just reiser4 that's causing problems. I'm running reiser3 and it runs smoothly so far.
_________________
#define NULL rand() /*heh heh heh */
Green Is Good |
|
| Back to top |
|
|
scout
Veteran
Joined: 08 Mar 2003
Posts: 1991
Location: France, Paris en Semaine / Metz le W-E
|
| Posted: Thu Apr 29, 2004 9:16 pm Post subject: |
|
|
| aethyr wrote: | | not_registered wrote: | | AS is amazing compared to CFQ. Amazing. |
umm... you do know that AS is the default scheduler in the vanilla kernel?
CFQ is the "new" scheduler. |
CFQ isn't new (google a bit and you'll find patches more than a year old) ... cfq is better only for low latency, apart from that it is outperformed by AS which is the default. If you want to understand how AS works, It's explained in the Doc included in the kernel, however I didn't find any doc for cfq ...
_________________
http://petition.eurolinux.org/ - Petition against ePatents
L'essence de la finesse |
|
| Back to top |
|
|
aethyr
Veteran
Joined: 06 Apr 2003
Posts: 1085
Location: NYC
|
| Posted: Fri Apr 30, 2004 1:10 am Post subject: |
|
|
| scout wrote: | | aethyr wrote: | | not_registered wrote: | | AS is amazing compared to CFQ. Amazing. |
umm... you do know that AS is the default scheduler in the vanilla kernel?
CFQ is the "new" scheduler. |
CFQ isn't new (google a bit and you'll find patches more than a year old) ... cfq is better only for low latency, apart from that it is outperformed by AS which is the default. If you want to understand how AS works, It's explained in the Doc included in the kernel, however I didn't find any doc for cfq ... |
I know it's not new (as in just released), this is why I put "new" in quotes. Anyway, it's new compared to AS, and it's still not available in the vanilla kernel (which only has deadline and AS AFAIK). There's also no newer scheduler available that I'm aware of.
That is what qualified it as "new" in my book. |
|
| Back to top |
|
|
OneOfOne
Guru
Joined: 28 May 2003
Posts: 368
|
|
| Back to top |
|
|
Ylin
n00b
Joined: 10 Apr 2004
Posts: 27
|
| Posted: Fri Apr 30, 2004 1:05 pm Post subject: |
|
|
Is it just the bootsplash that changed?
I'll give it a try later. |
|
| Back to top |
|
|
scaba
Apprentice
Joined: 10 Jan 2004
Posts: 252
Location: Zurich, Switzerland
|
| Posted: Fri Apr 30, 2004 1:51 pm Post subject: |
|
|
thanks a lot. compiling now...
_________________
Out of silence, a story. Out of chaos, order. Out of nothingness, love...
-- Tad Williams, Otherland |
|
| Back to top |
|
|
dr3amwalker
n00b
Joined: 28 Apr 2004
Posts: 8
Location: Lisbon, Portugal
|
| Posted: Fri Apr 30, 2004 3:14 pm Post subject: |
|
|
thanks for this release, compiled sucessfully
bootsplash working smoothly 
_________________
still walking dreams
2.6.6-rc3-love4 | linux: #353780 |
|
| Back to top |
|
|
|
Unsupported Software
