Re: speedfreq: epia + longhaul + speedfreq + copying large files = freeze

Re: speedfreq: epia + longhaul + speedfreq + copying large files = freeze

[Jeremy Fitzhardinge]

On Tue, 2004-12-21 at 16:00 +0100, Sebastian wrote:
> Dear Jeremy,
> 
> thanks for your tool! I like it best of all cpufreq daemons,

Thanks!

This problem sounds like a bug in the longhaul driver though, which I
don't really know anything about.  (The system should never crash,
regardless of what speedfreqd does.)  

I've cc:d the cpufreq list to see if someone there can help.  Dave?

	J

> Anyway, the Epia board normally runs fine with speedfreq. Frequency gets
> adjusted between 400Mhz and 600MHz. But when I start cutting
> advertisements out of the DVB recordings with the VDR software (Video
> Disk Recorder from Klaus Schmiedinger, www.cadsoft.de/vdr) after a short
> while it freezes.  The cutting process is simply like copying large
> files from one folder to another folder on the same partition. Mostly
> 1.7 to 2.7GB. I tried the following:
> 
> 1. Cutting a recording, cpufreq enabled, speedfreq daemon running:
> 
> System freeze.
> 
> 2. Copying large files manually, cpufreq enabled, speedfreq daemon running:
> 
> System freeze.
> 
> 3. Cutting a recording, cpufreq enabled, speedfreq daemon disabled:
> 
> Works.
> 
> 4. Copying large files manually, cpufreq enabled, speedfreq daemon disabled:
> 
> Works.
> 
> I'd like to add that sometimes I get messages in syslog telling me that
> speedfreq thought the actual frequency was xxx and discovered that the
> real frequency was something else. Once that was the last thing in
> syslog before a freeze. I don't know if this is relevant, because 99% of
> the freezes the syslog stays like it is, nothing gets added.
> 
> The kernel is a vanilla 2.6.9. The Linux Distribution is Gentoo stable. 
> Flags are "-O2 -march=c3 -fomit-frame-pointer -mmmx -m3dnow -pipe". I 
> have no 4k stacks, no preemption, no nptl, no perl ithreads.
> 
> Maybe speedfreq changing frequencies too fast for the little Epia is the 
> reason for the freezes? What can I do?

Re: speedfreq: epia + longhaul + speedfreq + copying large files = freeze

[Dave Jones]

On Wed, Dec 22, 2004 at 11:02:39AM -0800, Jeremy Fitzhardinge wrote:
 > On Tue, 2004-12-21 at 16:00 +0100, Sebastian wrote:
 > > Dear Jeremy,
 > > 
 > > thanks for your tool! I like it best of all cpufreq daemons,
 > 
 > Thanks!
 > 
 > This problem sounds like a bug in the longhaul driver though, which I
 > don't really know anything about.  (The system should never crash,
 > regardless of what speedfreqd does.)  
 > 
 > I've cc:d the cpufreq list to see if someone there can help.  Dave?

Yeah, longhaul is badly broken right now, and tbh, I've no idea
what on earth is wrong. Speed transitions happen, and work correctly.
Then a few minutes later, system locks up.

I've been trying all sorts of things, from disabling PCI mastering,
disabling CPU caches, etc, but nothing seems to help. I'm truly puzzled.

		Dave

Re: speedfreq: epia + longhaul + speedfreq + copying large files = freeze

[Jeremy Fitzhardinge]

On Wed, 2004-12-22 at 15:31 -0500, Dave Jones wrote:
> Yeah, longhaul is badly broken right now, and tbh, I've no idea
> what on earth is wrong. Speed transitions happen, and work correctly.
> Then a few minutes later, system locks up.
> 
> I've been trying all sorts of things, from disabling PCI mastering,
> disabling CPU caches, etc, but nothing seems to help. I'm truly puzzled.

I remember old versions of the C3 used to forget to snoop DMA traffic
while in the hlt instruction, so it wouldn't surprise me if there's some
other bug like that related to longhaul.  

Do you mean that it will crash some minutes after a single transition,
or is it less deterministic than that?  Do lots of transitions make it
worse?  Is the delay between transition and crash a random or
approximately fixed length, and does it depend on either core or bus
speed?

The reason I ask is that I have the ear of Via engineering support at
the moment (we're using embedded C3's in a product at work), so I could
ask a few well-placed questions.

	J

Re: speedfreq: epia + longhaul + speedfreq + copying large files = freeze

[Sebastian]

[-- Attachment #1: Type: text/plain, Size: 1669 bytes --]

Dave Jones schrieb:
> On Wed, Dec 22, 2004 at 11:02:39AM -0800, Jeremy Fitzhardinge wrote:
>  > On Tue, 2004-12-21 at 16:00 +0100, Sebastian wrote:
>  > > Dear Jeremy,
>  > > 
>  > > thanks for your tool! I like it best of all cpufreq daemons,
>  > 
>  > Thanks!
>  > 
>  > This problem sounds like a bug in the longhaul driver though, which I
>  > don't really know anything about.  (The system should never crash,
>  > regardless of what speedfreqd does.)  
>  > 
>  > I've cc:d the cpufreq list to see if someone there can help.  Dave?
> 
> Yeah, longhaul is badly broken right now, and tbh, I've no idea
> what on earth is wrong. Speed transitions happen, and work correctly.
> Then a few minutes later, system locks up.
> 
> I've been trying all sorts of things, from disabling PCI mastering,
> disabling CPU caches, etc, but nothing seems to help. I'm truly puzzled.
> 
> 		Dave
> 
> 

Dear Jeremy, dear Dave, dear list,

I tried to collect more information using cpufreq.debug={1,2,3,4}. But I 
don't see anything helpful in the output. Maybe you can find somehing, 
so I'm going to attach it anyway. Would it be interesting to see if a 
previous kernel version like 2.6.5 worked? I sure could install an 
earlier kernel if you like. Maybe this has not so much to do with the 
cpufreq-longhaul driver but with changes to the IO-section of the kernel?
A few days ago I converted all partitions on the computer from reiserfs 
to xfs, but as you can see, this didn't change anything regarding the 
freezes :)

Regards

Sebastian

P.S.: I hope you don't mind me attaching the various output as a 
.bz-file. Unpacked it's ~24kB, bzipped only ~3kB.

Thanks

Sebastian

[-- Attachment #2: cpufreq_debug_epia_me6000.bz2 --]
[-- Type: application/octet-stream, Size: 2663 bytes --]

[-- Attachment #3: Type: text/plain, Size: 143 bytes --]

_______________________________________________
Cpufreq mailing list
Cpufreq@www.linux.org.uk
http://www.linux.org.uk/mailman/listinfo/cpufreq

Re: speedfreq: epia + longhaul + speedfreq + copying large files = freeze

[Sebastian]

Ok,

seems like attaching the output didn't really work. Here we go:

----------------------------------------------------------------
----------------------------------------------------------------

1. worked for the first time
2. freezing after 35 seconds
3. cpufreq.debug=1

dmesg output:

cpufreq-core: trying to register driver longhaul
cpufreq-core: adding CPU 0
longhaul: VIA C3 'Samuel 2' [C5B] CPU detected.  Longhaul v1 supported.
freq-table: table entry 0: 399000 kHz, 1 index
freq-table: table entry 1: 532000 kHz, 2 index
freq-table: table entry 2: 465000 kHz, 5 index
freq-table: table entry 3: 598000 kHz, 6 index
freq-table: setting show_table for cpu 0 to c1221e80
cpufreq-core: setting new policy for CPU 0: 399000 - 598000 kHz
freq-table: request for verification of policy (399000 - 598000 kHz) for 
cpu 0
freq-table: verification lead to (399000 - 598000 kHz) for cpu 0
freq-table: request for verification of policy (399000 - 598000 kHz) for 
cpu 0
freq-table: verification lead to (399000 - 598000 kHz) for cpu 0
cpufreq-core: new min and max freqs are 399000 - 598000 kHz
cpufreq-core: governor switch
cpufreq-core: __cpufreq_governor for CPU 0, event 1
cpufreq-core: governor: change or update limits
cpufreq-core: __cpufreq_governor for CPU 0, event 3
cpufreq-core: target for CPU 0: 598000 kHz, relation 0
freq-table: request for target 598000 kHz (relation: 0) for cpu 0
freq-table: target is 3 (598000 kHz, 6)
cpufreq-core: notification 0 of frequency transition to 598000 kHz
cpufreq-core: saving 591872 as reference value for loops_per_jiffy; freq 
is 598000 kHz
cpufreq-core: notification 1 of frequency transition to 598000 kHz
cpufreq-core: initialization complete
cpufreq-core: driver longhaul up and running
SGI XFS with no debug enabled
ACPI: Power Button (FF) [PWRF]
cpufreq-core: updating policy for CPU 0
cpufreq-core: setting new policy for CPU 0: 399000 - 598000 kHz
freq-table: request for verification of policy (399000 - 598000 kHz) for 
cpu 0
freq-table: verification lead to (399000 - 598000 kHz) for cpu 0
freq-table: request for verification of policy (399000 - 598000 kHz) for 
cpu 0
freq-table: verification lead to (399000 - 598000 kHz) for cpu 0
cpufreq-core: new min and max freqs are 399000 - 598000 kHz
cpufreq-core: governor: change or update limits
cpufreq-core: __cpufreq_governor for CPU 0, event 3
cpufreq-core: target for CPU 0: 598000 kHz, relation 0
freq-table: request for target 598000 kHz (relation: 0) for cpu 0
freq-table: target is 3 (598000 kHz, 6)
ACPI: Processor [CPU0] (supports C1 C2)
ACPI: Processor [CPU0] (supports 2 throttling states)

Cutting a recording: froze after 5 seconds.

vdr root # speedfreq -m
New policy: dynamic 399-598 MHz
CPU new speed 399 MHz, 95.0027% idle
CPU new speed 532 MHz, 13.1302% idle
CPU new speed 598 MHz, 28.5787% idle
CPU new speed 532 MHz, 99.9924% idle
CPU new speed 399 MHz, 94.7394% idle
CPU new speed 532 MHz, 25.0019% idle
CPU new speed 598 MHz, 26.309% idle

tail -f /var/log/messages:

Dec 23 11:35:38 vdr cpufreq-core: setting new policy for CPU 0: 399000 - 
598000 kHz
Dec 23 11:35:38 vdr freq-table: request for verification of policy 
(399000 - 598000 kHz) for cpu 0
Dec 23 11:35:38 vdr freq-table: verification lead to (399000 - 598000 
kHz) for cpu 0
Dec 23 11:35:38 vdr freq-table: request for verification of policy 
(399000 - 598000 kHz) for cpu 0
Dec 23 11:35:38 vdr freq-table: verification lead to (399000 - 598000 
kHz) for cpu 0
Dec 23 11:35:38 vdr cpufreq-core: new min and max freqs are 399000 - 
598000 kHz
Dec 23 11:35:38 vdr cpufreq-core: governor: change or update limits
Dec 23 11:35:38 vdr cpufreq-core: __cpufreq_governor for CPU 0, event 3
Dec 23 11:35:38 vdr cpufreq-core: target for CPU 0: 598000 kHz, relation 0
Dec 23 11:35:38 vdr freq-table: request for target 598000 kHz (relation: 
0) for cpu 0
Dec 23 11:35:38 vdr freq-table: target is 3 (598000 kHz, 6)
Dec 23 11:35:38 vdr cpufreq-core: target for CPU 0: 498000 kHz, relation 0
Dec 23 11:35:38 vdr freq-table: request for target 498000 kHz (relation: 
0) for cpu 0
Dec 23 11:35:38 vdr freq-table: target is 1 (532000 kHz, 2)
Dec 23 11:35:38 vdr cpufreq-core: notification 0 of frequency transition 
to 532000 kHz
Dec 23 11:35:38 vdr cpufreq-core: notification 1 of frequency transition 
to 532000 kHz
Dec 23 11:35:38 vdr cpufreq-core: scaling loops_per_jiffy to 526548 for 
frequency 532000 kHz
Dec 23 11:35:38 vdr cpufreq-core: target for CPU 0: 399000 kHz, relation 0
Dec 23 11:35:38 vdr freq-table: request for target 399000 kHz (relation: 
0) for cpu 0
Dec 23 11:35:38 vdr freq-table: target is 0 (399000 kHz, 1)
Dec 23 11:35:38 vdr cpufreq-core: notification 0 of frequency transition 
to 399000 kHz
Dec 23 11:36:29 vdr printk: 2 messages suppressed.
Dec 23 11:36:29 vdr cpufreq-core: target for CPU 0: 499000 kHz, relation 0
Dec 23 11:36:29 vdr freq-table: request for target 499000 kHz (relation: 
0) for cpu 0
Dec 23 11:36:29 vdr freq-table: target is 1 (532000 kHz, 2)
Dec 23 11:36:29 vdr cpufreq-core: notification 0 of frequency transition 
to 532000 kHz
Dec 23 11:36:29 vdr cpufreq-core: scaling loops_per_jiffy to 526548 for 
frequency 532000 kHz
Dec 23 11:36:29 vdr cpufreq-core: notification 1 of frequency transition 
to 532000 kHz
Dec 23 11:36:30 vdr cpufreq-core: target for CPU 0: 598000 kHz, relation 0
Dec 23 11:36:30 vdr freq-table: request for target 598000 kHz (relation: 
0) for cpu 0
Dec 23 11:36:30 vdr freq-table: target is 3 (598000 kHz, 6)
Dec 23 11:36:30 vdr cpufreq-core: notification 0 of frequency transition 
to 598000 kHz

4. cpufreq.debug=2

dmesg:

longhaul: VIA C3 'Samuel 2' [C5B] CPU detected.  Longhaul v1 supported.
longhaul: MinMult:3.0x MaxMult:4.5x
longhaul: FSB:133MHz  Lowest speed: 399MHz   Highest speed:399MHz
longhaul: Setting to FSB:133MHz Mult:4.5x (598MHz)

Cutting a recording froze system after 1 minute 33 seconds.

vdr root # speedfreq -m
New policy: dynamic 399-598 MHz
CPU new speed 399 MHz, 100% idle
CPU new speed 532 MHz, 14.1518% idle
CPU new speed 399 MHz, 94.995% idle
CPU new speed 532 MHz, 14.9996% idle
CPU new speed 598 MHz, 25.0019% idle
CPU new speed 532 MHz, 95.0027% idle
CPU new speed 399 MHz, 95.2336% idle
CPU new speed 532 MHz, 14.9996% idle
CPU new speed 598 MHz, 4.99733% idle
CPU new speed 532 MHz, 95.0027% idle
CPU new speed 399 MHz, 95.2409% idle
CPU new speed 532 MHz, 0% idle
CPU new speed 598 MHz, 4.75948% idle
CPU new speed 532 MHz, 99.9924% idle
CPU new speed 399 MHz, 90.481% idle
CPU new speed 532 MHz, 0% idle
CPU new speed 598 MHz, 5.00458% idle
CPU new speed 532 MHz, 90.0053% idle
CPU new speed 399 MHz, 99.9924% idle
CPU new speed 532 MHz, 14.9996% idle
CPU new speed 598 MHz, 4.99733% idle
CPU new speed 532 MHz, 90.4745% idle
CPU new speed 399 MHz, 100% idle
CPU new speed 532 MHz, 0% idle
CPU new speed 598 MHz, 4.99733% idle
CPU new speed 532 MHz, 89.9977% idle
CPU new speed 399 MHz, 95.2405% idle
CPU new speed 532 MHz, 5.00496% idle
CPU new speed 598 MHz, 10.0023% idle
CPU new speed 532 MHz, 100% idle
CPU new speed 399 MHz, 90.0053% idle
CPU new speed 532 MHz, 10.0023% idle
CPU new speed 598 MHz, 0% idle
CPU new speed 532 MHz, 95.2405% idle
CPU new speed 399 MHz, 90.0053% idle
CPU new speed 532 MHz, 0% idle
CPU new speed 598 MHz, 5.2602% idle
CPU new speed 532 MHz, 100% idle
CPU new speed 399 MHz, 89.9977% idle
CPU new speed 532 MHz, 0% idle
CPU new speed 598 MHz, 0% idle
CPU new speed 532 MHz, 100% idle
CPU new speed 399 MHz, 90.0053% idle
CPU new speed 532 MHz, 0% idle
CPU new speed 598 MHz, 4.99733% idle
CPU new speed 532 MHz, 94.9954% idle
CPU new speed 399 MHz, 85.0004% idle
CPU new speed 532 MHz, 14.9996% idle
CPU new speed 598 MHz, 15.7886% idle
CPU new speed 532 MHz, 95.2336% idle
CPU new speed 399 MHz, 95.0027% idle
CPU new speed 532 MHz, 0% idle
CPU new speed 598 MHz, 4.99695% idle
CPU new speed 532 MHz, 95.2409% idle
CPU new speed 399 MHz, 94.7318% idle
CPU new speed 532 MHz, 10.0015% idle
CPU new speed 598 MHz, 29.9992% idle
CPU new speed 532 MHz, 95.2336% idle
CPU new speed 399 MHz, 100% idle
CPU new speed 532 MHz, 0% idle
CPU new speed 598 MHz, 0% idle
CPU new speed 532 MHz, 95.0027% idle
CPU new speed 399 MHz, 90.4745% idle
CPU new speed 532 MHz, 24.9924% idle
CPU new speed 598 MHz, 29.9992% idle
CPU new speed 532 MHz, 90.4745% idle
CPU new speed 399 MHz, 95.0027% idle
CPU new speed 532 MHz, 10.0023% idle
CPU new speed 598 MHz, 0% idle
CPU new speed 532 MHz, 90.0053% idle
CPU new speed 399 MHz, 89.9977% idle
CPU new speed 532 MHz, 0% idle
CPU new speed 598 MHz, 0% idle
CPU new speed 532 MHz, 100% idle
CPU new speed 399 MHz, 95.2336% idle
CPU new speed 532 MHz, 5.2602% idle
CPU new speed 598 MHz, 20.0046% idle
CPU new speed 532 MHz, 94.995% idle
CPU new speed 399 MHz, 95.0027% idle
CPU new speed 532 MHz, 29.9992% idle
CPU new speed 598 MHz, 5.26823% idle
CPU new speed 532 MHz, 100% idle
CPU new speed 399 MHz, 95.2405% idle
CPU new speed 532 MHz, 10.0023% idle
CPU new speed 598 MHz, 26.3171% idle
CPU new speed 532 MHz, 90.4745% idle
CPU new speed 399 MHz, 99.992% idle
CPU new speed 532 MHz, 4.99733% idle
CPU new speed 598 MHz, 20.0046% idle
CPU new speed 532 MHz, 94.9878% idle
CPU new speed 399 MHz, 95.0027% idle
CPU new speed 532 MHz, 0% idle
CPU new speed 598 MHz, 5.2602% idle
CPU new speed 532 MHz, 94.9954% idle
CPU new speed 399 MHz, 95.2405% idle
CPU new speed 532 MHz, 0% idle
CPU new speed 598 MHz, 5.00496% idle
CPU new speed 532 MHz, 90.481% idle
CPU new speed 399 MHz, 94.9954% idle
CPU new speed 532 MHz, 0% idle
CPU new speed 598 MHz, 0% idle
CPU new speed 532 MHz, 89.9908% idle
CPU new speed 399 MHz, 100% idle
CPU new speed 532 MHz, 28.5766% idle
CPU new speed 598 MHz, 0% idle
CPU new speed 532 MHz, 95.0027% idle
CPU new speed 399 MHz, 89.9977% idle
CPU new speed 532 MHz, 0% idle
CPU new speed 598 MHz, 4.99695% idle
CPU new speed 532 MHz, 95.0027% idle
CPU new speed 399 MHz, 90.481% idle
CPU new speed 532 MHz, 0% idle
CPU new speed 598 MHz, 4.99733% idle
CPU new speed 532 MHz, 95.2405% idle
CPU new speed 399 MHz, 89.9977% idle
CPU new speed 532 MHz, 4.7664% idle
CPU new speed 598 MHz, 0% idle
CPU new speed 532 MHz, 95.0027% idle
CPU new speed 399 MHz, 94.995% idle
CPU new speed 532 MHz, 19.9969% idle
CPU new speed 598 MHz, 20.0046% idle
CPU new speed 532 MHz, 94.7318% idle
CPU new speed 399 MHz, 95.2409% idle
CPU new speed 532 MHz, 5.26823% idle
CPU new speed 598 MHz, 0% idle
CPU new speed 532 MHz, 95.0027% idle
CPU new speed 399 MHz, 90.0053% idle
CPU new speed 532 MHz, 0% idle
CPU new speed 598 MHz, 4.99733% idle
CPU new speed 532 MHz, 90.0053% idle
CPU new speed 399 MHz, 95.0027% idle
CPU new speed 532 MHz, 28.5694% idle
CPU new speed 598 MHz, 4.76675% idle
CPU new speed 532 MHz, 94.7474% idle
CPU new speed 399 MHz, 90.4745% idle
CPU new speed 532 MHz, 14.9996% idle
CPU new speed 598 MHz, 29.9992% idle
CPU new speed 532 MHz, 100% idle
CPU new speed 399 MHz, 90.0053% idle
CPU new speed 532 MHz, 10.5284% idle
CPU new speed 598 MHz, 4.99733% idle
CPU new speed 532 MHz, 100% idle
CPU new speed 399 MHz, 89.9908% idle
CPU new speed 532 MHz, 28.5714% idle
CPU new speed 598 MHz, 5.26823% idle
CPU new speed 532 MHz, 100% idle
CPU new speed 399 MHz, 90.481% idle
CPU new speed 532 MHz, 0% idle
CPU new speed 598 MHz, 0% idle
CPU new speed 532 MHz, 90.4738% idle
CPU new speed 399 MHz, 94.9954% idle
CPU new speed 532 MHz, 0% idle
CPU new speed 598 MHz, 0% idle
CPU new speed 532 MHz, 95.2405% idle
CPU new speed 399 MHz, 90.4745% idle
CPU new speed 532 MHz, 9.9939% idle
CPU new speed 598 MHz, 20.0046% idle
CPU new speed 532 MHz, 100% idle
CPU new speed 399 MHz, 94.9954% idle
CPU new speed 532 MHz, 10.0015% idle
CPU new speed 598 MHz, 19.0452% idle
CPU new speed 532 MHz, 100% idle
CPU new speed 399 MHz, 94.7398% idle
CPU new speed 532 MHz, 15.0072% idle
CPU new speed 598 MHz, 4.75914% idle
CPU new speed 532 MHz, 95.0027% idle
CPU new speed 399 MHz, 90.4745% idle
CPU new speed 532 MHz, 20.0046% idle
CPU new speed 598 MHz, 25.0019% idle
CPU new speed 532 MHz, 90.4672% idle
CPU new speed 399 MHz, 95.0027% idle
CPU new speed 532 MHz, 19.9969% idle
CPU new speed 598 MHz, 21.0569% idle
CPU new speed 532 MHz, 94.7318% idle
CPU new speed 399 MHz, 89.9985% idle
CPU new speed 532 MHz, 25% idle
CPU new speed 598 MHz, 4.99733% idle
CPU new speed 532 MHz, 95.0099% idle
CPU new speed 399 MHz, 95.2405% idle
CPU new speed 532 MHz, 0% idle
CPU new speed 598 MHz, 0% idle
CPU new speed 532 MHz, 100% idle
CPU new speed 399 MHz, 95.2405% idle
CPU new speed 532 MHz, 19.0525% idle
CPU new speed 598 MHz, 4.75948% idle
CPU new speed 532 MHz, 94.7318% idle
CPU new speed 399 MHz, 95.2405% idle
CPU new speed 532 MHz, 15.0008% idle
CPU new speed 598 MHz, 14.9985% idle
CPU new speed 532 MHz, 95.2405% idle
CPU new speed 399 MHz, 94.9954% idle
CPU new speed 532 MHz, 5.2602% idle
CPU new speed 598 MHz, 20.0031% idle
CPU new speed 532 MHz, 90.0122% idle
CPU new speed 399 MHz, 89.9908% idle

tail -f /var/log/messages:

Dec 23 11:42:36 vdr sshd[6282]: Accepted publickey for root from 
192.168.0.2 port 32955 ssh2
Dec 23 11:43:31 vdr longhaul: Setting to FSB:133MHz Mult:4.0x (532MHz)
Dec 23 11:43:31 vdr longhaul: Setting to FSB:133MHz Mult:3.0x (399MHz)
Dec 23 11:43:47 vdr sshd[6355]: Accepted publickey for root from 
192.168.0.2 port 32956 ssh2
Dec 23 11:44:39 vdr longhaul: Setting to FSB:133MHz Mult:4.0x (532MHz)
Dec 23 11:44:39 vdr longhaul: Setting to FSB:133MHz Mult:3.0x (399MHz)
Dec 23 11:44:40 vdr longhaul: Setting to FSB:133MHz Mult:4.0x (532MHz)
Dec 23 11:44:40 vdr longhaul: Setting to FSB:133MHz Mult:4.5x (598MHz)
Dec 23 11:44:41 vdr longhaul: Setting to FSB:133MHz Mult:4.0x (532MHz)
Dec 23 11:44:41 vdr longhaul: Setting to FSB:133MHz Mult:3.0x (399MHz)
Dec 23 11:44:41 vdr longhaul: Setting to FSB:133MHz Mult:4.0x (532MHz)
Dec 23 11:44:42 vdr longhaul: Setting to FSB:133MHz Mult:4.5x (598MHz)
Dec 23 11:44:43 vdr longhaul: Setting to FSB:133MHz Mult:4.0x (532MHz)
Dec 23 11:44:43 vdr longhaul: Setting to FSB:133MHz Mult:3.0x (399MHz)
Dec 23 11:44:44 vdr printk: 2 messages suppressed.
Dec 23 11:44:44 vdr longhaul: Setting to FSB:133MHz Mult:4.0x (532MHz)
Dec 23 11:44:49 vdr printk: 11 messages suppressed.
Dec 23 11:44:49 vdr longhaul: Setting to FSB:133MHz Mult:4.0x (532MHz)
Dec 23 11:44:54 vdr printk: 10 messages suppressed.
Dec 23 11:44:54 vdr longhaul: Setting to FSB:133MHz Mult:4.5x (598MHz)
Dec 23 11:44:59 vdr printk: 10 messages suppressed.
Dec 23 11:44:59 vdr longhaul: Setting to FSB:133MHz Mult:4.0x (532MHz)
Dec 23 11:45:04 vdr printk: 9 messages suppressed.
Dec 23 11:45:04 vdr longhaul: Setting to FSB:133MHz Mult:4.0x (532MHz)
Dec 23 11:45:09 vdr printk: 11 messages suppressed.
Dec 23 11:45:09 vdr longhaul: Setting to FSB:133MHz Mult:4.0x (532MHz)
Dec 23 11:45:15 vdr printk: 11 messages suppressed.
Dec 23 11:45:15 vdr longhaul: Setting to FSB:133MHz Mult:4.0x (532MHz)
Dec 23 11:45:19 vdr printk: 8 messages suppressed.
Dec 23 11:45:19 vdr longhaul: Setting to FSB:133MHz Mult:3.0x (399MHz)
Dec 23 11:45:24 vdr printk: 10 messages suppressed.
Dec 23 11:45:24 vdr longhaul: Setting to FSB:133MHz Mult:4.0x (532MHz)
Dec 23 11:45:29 vdr printk: 11 messages suppressed.
Dec 23 11:45:29 vdr longhaul: Setting to FSB:133MHz Mult:4.0x (532MHz)
Dec 23 11:45:34 vdr printk: 10 messages suppressed.
Dec 23 11:45:34 vdr longhaul: Setting to FSB:133MHz Mult:4.5x (598MHz)
Dec 23 11:45:39 vdr printk: 10 messages suppressed.
Dec 23 11:45:39 vdr longhaul: Setting to FSB:133MHz Mult:4.0x (532MHz)
Dec 23 11:45:44 vdr printk: 10 messages suppressed.
Dec 23 11:45:44 vdr longhaul: Setting to FSB:133MHz Mult:3.0x (399MHz)
Dec 23 11:45:49 vdr printk: 10 messages suppressed.
Dec 23 11:45:49 vdr longhaul: Setting to FSB:133MHz Mult:4.0x (532MHz)
Dec 23 11:45:54 vdr printk: 9 messages suppressed.
Dec 23 11:45:54 vdr longhaul: Setting to FSB:133MHz Mult:4.0x (532MHz)
Dec 23 11:45:59 vdr printk: 9 messages suppressed.
Dec 23 11:45:59 vdr longhaul: Setting to FSB:133MHz Mult:4.0x (532MHz)
Dec 23 11:46:04 vdr printk: 11 messages suppressed.
Dec 23 11:46:04 vdr longhaul: Setting to FSB:133MHz Mult:4.0x (532MHz)

5. cpufreq.debug=3

dmesg:

cpufreq-core: trying to register driver longhaul
cpufreq-core: adding CPU 0
longhaul: VIA C3 'Samuel 2' [C5B] CPU detected.  Longhaul v1 supported.
longhaul: MinMult:3.0x MaxMult:4.5x
longhaul: FSB:133MHz  Lowest speed: 399MHz   Highest speed:399MHz
freq-table: table entry 0: 399000 kHz, 1 index
freq-table: table entry 1: 532000 kHz, 2 index
freq-table: table entry 2: 465000 kHz, 5 index
freq-table: table entry 3: 598000 kHz, 6 index
freq-table: setting show_table for cpu 0 to c1221e80
cpufreq-core: setting new policy for CPU 0: 399000 - 598000 kHz
freq-table: request for verification of policy (399000 - 598000 kHz) for 
cpu 0
freq-table: verification lead to (399000 - 598000 kHz) for cpu 0
freq-table: request for verification of policy (399000 - 598000 kHz) for 
cpu 0
freq-table: verification lead to (399000 - 598000 kHz) for cpu 0
cpufreq-core: new min and max freqs are 399000 - 598000 kHz
cpufreq-core: governor switch
cpufreq-core: __cpufreq_governor for CPU 0, event 1
cpufreq-core: governor: change or update limits
cpufreq-core: __cpufreq_governor for CPU 0, event 3
cpufreq-core: target for CPU 0: 598000 kHz, relation 0
freq-table: request for target 598000 kHz (relation: 0) for cpu 0
freq-table: target is 3 (598000 kHz, 6)
cpufreq-core: notification 0 of frequency transition to 598000 kHz
cpufreq-core: saving 591872 as reference value for loops_per_jiffy; freq 
is 598000 kHz
longhaul: Setting to FSB:133MHz Mult:4.5x (598MHz)
cpufreq-core: notification 1 of frequency transition to 598000 kHz
cpufreq-core: initialization complete
cpufreq-core: driver longhaul up and running
SGI XFS with no debug enabled
ACPI: Power Button (FF) [PWRF]
cpufreq-core: updating policy for CPU 0
cpufreq-core: setting new policy for CPU 0: 399000 - 598000 kHz
freq-table: request for verification of policy (399000 - 598000 kHz) for 
cpu 0
freq-table: verification lead to (399000 - 598000 kHz) for cpu 0
freq-table: request for verification of policy (399000 - 598000 kHz) for 
cpu 0
freq-table: verification lead to (399000 - 598000 kHz) for cpu 0
cpufreq-core: new min and max freqs are 399000 - 598000 kHz
cpufreq-core: governor: change or update limits
cpufreq-core: __cpufreq_governor for CPU 0, event 3
cpufreq-core: target for CPU 0: 598000 kHz, relation 0
freq-table: request for target 598000 kHz (relation: 0) for cpu 0
freq-table: target is 3 (598000 kHz, 6)
ACPI: Processor [CPU0] (supports C1 C2)
ACPI: Processor [CPU0] (supports 2 throttling states)

Cutting a recording froze the computer after 3 seconds

vdr root # speedfreq -m
New policy: dynamic 399-598 MHz
CPU new speed 399 MHz, 94.995% idle
CPU new speed 532 MHz, 13.9999% idle
CPU new speed 399 MHz, 100% idle

tail -f /vag/log/messages:

Dec 23 11:50:26 vdr sshd[6323]: Accepted publickey for root from 
192.168.0.2 port 32964 ssh2
Dec 23 11:51:55 vdr cpufreq-core: setting new policy for CPU 0: 399000 - 
598000 kHz
Dec 23 11:51:55 vdr freq-table: request for verification of policy 
(399000 - 598000 kHz) for cpu 0
Dec 23 11:51:55 vdr freq-table: verification lead to (399000 - 598000 
kHz) for cpu 0
Dec 23 11:51:55 vdr freq-table: request for verification of policy 
(399000 - 598000 kHz) for cpu 0
Dec 23 11:51:55 vdr freq-table: verification lead to (399000 - 598000 
kHz) for cpu 0
Dec 23 11:51:55 vdr cpufreq-core: new min and max freqs are 399000 - 
598000 kHz
Dec 23 11:51:55 vdr cpufreq-core: governor: change or update limits
Dec 23 11:51:55 vdr cpufreq-core: __cpufreq_governor for CPU 0, event 3
Dec 23 11:51:55 vdr cpufreq-core: target for CPU 0: 598000 kHz, relation 0
Dec 23 11:51:55 vdr freq-table: request for target 598000 kHz (relation: 
0) for cpu 0
Dec 23 11:51:55 vdr freq-table: target is 3 (598000 kHz, 6)
Dec 23 11:51:55 vdr cpufreq-core: target for CPU 0: 498000 kHz, relation 0
Dec 23 11:51:55 vdr freq-table: request for target 498000 kHz (relation: 
0) for cpu 0
Dec 23 11:51:55 vdr freq-table: target is 1 (532000 kHz, 2)
Dec 23 11:51:55 vdr cpufreq-core: notification 0 of frequency transition 
to 532000 kHz
Dec 23 11:51:55 vdr longhaul: Setting to FSB:133MHz Mult:4.0x (532MHz)
Dec 23 11:51:55 vdr cpufreq-core: notification 1 of frequency transition 
to 532000 kHz
Dec 23 11:51:55 vdr cpufreq-core: scaling loops_per_jiffy to 526548 for 
frequency 532000 kHz
Dec 23 11:51:55 vdr cpufreq-core: target for CPU 0: 399000 kHz, relation 0
Dec 23 11:51:55 vdr freq-table: request for target 399000 kHz (relation: 
0) for cpu 0
Dec 23 11:51:55 vdr freq-table: target is 0 (399000 kHz, 1)
Dec 23 11:52:27 vdr printk: 4 messages suppressed.
Dec 23 11:52:27 vdr cpufreq-core: target for CPU 0: 499000 kHz, relation 0
Dec 23 11:52:27 vdr freq-table: request for target 499000 kHz (relation: 
0) for cpu 0
Dec 23 11:52:27 vdr freq-table: target is 1 (532000 kHz, 2)
Dec 23 11:52:27 vdr cpufreq-core: notification 0 of frequency transition 
to 532000 kHz
Dec 23 11:52:27 vdr cpufreq-core: scaling loops_per_jiffy to 526548 for 
frequency 532000 kHz
Dec 23 11:52:27 vdr longhaul: Setting to FSB:133MHz Mult:4.0x (532MHz)

6. cpufreq.debug=4

dmesg:

longhaul: VIA C3 'Samuel 2' [C5B] CPU detected.  Longhaul v1 supported.
userspace: managing cpu 0 started (399000 - 598000 kHz, currently 598000 
kHz)
userspace: limit event for cpu 0: 399000 - 598000 kHz, currently 598000 
kHz, last set to 598000 kHz
userspace: saving cpu_cur_freq of cpu 0 to be 598000 kHz
userspace: saving cpu_cur_freq of cpu 0 to be 598000 kHz

Cutting a recording froze the computer after 24 seconds.

vdr root # speedfreq -m
New policy: dynamic 399-598 MHz
CPU new speed 399 MHz, 94.9989% idle
CPU new speed 532 MHz, 29.0001% idle
CPU new speed 399 MHz, 90.0015% idle
CPU new speed 532 MHz, 23.8102% idle
CPU new speed 598 MHz, 5.00114% idle
CPU new speed 532 MHz, 90.0015% idle
CPU new speed 399 MHz, 100% idle
CPU new speed 532 MHz, 9.52191% idle
CPU new speed 598 MHz, 5.26421% idle
CPU new speed 532 MHz, 90.4745% idle
CPU new speed 399 MHz, 94.7358% idle
CPU new speed 532 MHz, 19.0452% idle
CPU new speed 598 MHz, 5.00114% idle
CPU new speed 532 MHz, 89.4828% idle
CPU new speed 399 MHz, 99.996% idle
CPU new speed 532 MHz, 28.573% idle
CPU new speed 598 MHz, 0% idle
CPU new speed 532 MHz, 100% idle
CPU new speed 399 MHz, 90.0015% idle
CPU new speed 532 MHz, 9.99847% idle
CPU new speed 598 MHz, 29.9969% idle
CPU new speed 532 MHz, 95.2405% idle
CPU new speed 399 MHz, 100% idle
CPU new speed 532 MHz, 19.9992% idle
CPU new speed 598 MHz, 29.9992% idle
CPU new speed 532 MHz, 90.0015% idle
CPU new speed 399 MHz, 100% idle
CPU new speed 532 MHz, 0% idle
CPU new speed 598 MHz, 4.76277% idle
CPU new speed 532 MHz, 100% idle
CPU new speed 399 MHz, 90.481% idle
CPU new speed 532 MHz, 0% idle
CPU new speed 598 MHz, 5.00114% idle
CPU new speed 532 MHz, 100% idle
CPU new speed 399 MHz, 90.0015% idle
CPU new speed 532 MHz, 9.99847% idle
CPU new speed 598 MHz, 0% idle
CPU new speed 532 MHz, 99.996% idle
CPU new speed 399 MHz, 90.4745% idle

tail -f /var/log/messages:

Dec 23 11:57:20 vdr 0.7.0[5382]: accepted new client on /dev/lircd
Dec 23 11:58:21 vdr userspace: limit event for cpu 0: 399000 - 598000 
kHz, currently 598000 kHz, last set to 598000 kHz
Dec 23 11:58:22 vdr userspace: _cpufreq_set for cpu 0, freq 498000 kHz
Dec 23 11:58:22 vdr userspace: saving cpu_cur_freq of cpu 0 to be 532000 kHz
Dec 23 11:58:22 vdr userspace: saving cpu_cur_freq of cpu 0 to be 532000 kHz
Dec 23 11:58:22 vdr userspace: _cpufreq_set for cpu 0, freq 399000 kHz
Dec 23 11:58:22 vdr userspace: saving cpu_cur_freq of cpu 0 to be 399000 kHz
Dec 23 11:58:22 vdr userspace: saving cpu_cur_freq of cpu 0 to be 399000 kHz
Dec 23 11:58:55 vdr userspace: _cpufreq_set for cpu 0, freq 499000 kHz
Dec 23 11:58:55 vdr userspace: saving cpu_cur_freq of cpu 0 to be 532000 kHz
Dec 23 11:58:55 vdr userspace: saving cpu_cur_freq of cpu 0 to be 532000 kHz
Dec 23 11:58:55 vdr userspace: _cpufreq_set for cpu 0, freq 399000 kHz
Dec 23 11:58:55 vdr userspace: saving cpu_cur_freq of cpu 0 to be 399000 kHz
Dec 23 11:58:55 vdr userspace: saving cpu_cur_freq of cpu 0 to be 399000 kHz
Dec 23 11:58:55 vdr userspace: _cpufreq_set for cpu 0, freq 499000 kHz
Dec 23 11:58:55 vdr userspace: saving cpu_cur_freq of cpu 0 to be 532000 kHz
Dec 23 11:58:55 vdr userspace: saving cpu_cur_freq of cpu 0 to be 532000 kHz
Dec 23 11:58:56 vdr userspace: _cpufreq_set for cpu 0, freq 598000 kHz
Dec 23 11:59:00 vdr printk: 29 messages suppressed.
Dec 23 11:59:00 vdr userspace: _cpufreq_set for cpu 0, freq 399000 kHz
Dec 23 11:59:01 vdr Warning: CPU frequency is 399000, cpufreq assumed 
532000 kHz.
Dec 23 11:59:05 vdr printk: 32 messages suppressed.
Dec 23 11:59:05 vdr userspace: _cpufreq_set for cpu 0, freq 498000 kHz

---------------------------------------------------------------------------
---------------------------------------------------------------------------

Regards

Sebastian

Re: speedfreq: epia + longhaul + speedfreq + copying large files = freeze

[Dave Jones]

On Wed, Dec 22, 2004 at 06:38:52PM -0800, Jeremy Fitzhardinge wrote:

 > > I've been trying all sorts of things, from disabling PCI mastering,
 > > disabling CPU caches, etc, but nothing seems to help. I'm truly puzzled.
 > 
 > I remember old versions of the C3 used to forget to snoop DMA traffic
 > while in the hlt instruction, so it wouldn't surprise me if there's some
 > other bug like that related to longhaul.  

That's why I was hoping disablign mastering would do the trick.

 > Do you mean that it will crash some minutes after a single transition,
 > or is it less deterministic than that?

Usually takes a few. I've noticed if I run cpuspeed, it steps it down
from 1.2GHz to 500MHz through a half dozen or so steps. Whilst it does
this, its stable. If I then kill cpuspeed, its stable.
But, if I leave cpuspeed run, and then do something CPU intensive, it
tries to jump up to fullspeed quickly (through less steps than
it took to scale down), when it does this, it hangs shortly afterwards.

We're not doing voltage scaling in longhaul yet, but that shouldn't
matter, as we should be booting at max voltage (as we boot at max speed)

 > The reason I ask is that I have the ear of Via engineering support at
 > the moment (we're using embedded C3's in a product at work), so I could
 > ask a few well-placed questions.
 
I speak with the centaur folks regularly, but they haven't got any
ideas what the problem could be either.

		Dave

Re: speedfreq: epia + longhaul + speedfreq + copying large files = freeze

[Jeremy Fitzhardinge]

On Thu, 2004-12-23 at 13:24 -0500, Dave Jones wrote:
> Usually takes a few. I've noticed if I run cpuspeed, it steps it down
> from 1.2GHz to 500MHz through a half dozen or so steps. Whilst it does
> this, its stable. If I then kill cpuspeed, its stable.
> But, if I leave cpuspeed run, and then do something CPU intensive, it
> tries to jump up to fullspeed quickly (through less steps than
> it took to scale down), when it does this, it hangs shortly afterwards.

Sebastian is reporting the problem with speedfreqd, which will use the
same number of steps both ways (and, depending on the CPU load curve, at
the same rate).

Does rate-limiting the transitions help stability?

> We're not doing voltage scaling in longhaul yet, but that shouldn't
> matter, as we should be booting at max voltage (as we boot at max speed)

Is voltage switching complex?  It might be worth trying - its not
something which normally improves stability, but maybe there's some
regulator issue under load changes, or something...

	J

Re: speedfreq: epia + longhaul + speedfreq + copying large files = freeze

[Dave Jones]

On Thu, Dec 23, 2004 at 12:05:03PM -0800, Jeremy Fitzhardinge wrote:

 > Sebastian is reporting the problem with speedfreqd, which will use the
 > same number of steps both ways (and, depending on the CPU load curve, at
 > the same rate).
 > 
 > Does rate-limiting the transitions help stability?

that's something worth trying maybe.

 > > We're not doing voltage scaling in longhaul yet, but that shouldn't
 > > matter, as we should be booting at max voltage (as we boot at max speed)
 > 
 > Is voltage switching complex?  It might be worth trying - its not
 > something which normally improves stability, but maybe there's some
 > regulator issue under load changes, or something...

With longhaul, its a little more complicated than it is with powernow,
as theres no bios table telling us which voltage should be used
for each frequency, and they vary depending on stepping & speed of the part.
The first version of the longhaul driver did actually do voltage scaling,
but the 'match the voltage to the speed' code was very crude, and
assumed that the voltage table was 1:1 with the frequency, which
worked for the most part, but was far from being 'the right thing'.

I'm not convinced that adding it back in would fix the stability problems.
If anything, it would add another variable.

		Dave