Re: F11-for-XO1.5 Release 10.1.1 Release Candidate 2
On 15.06.2010, at 16:47, Paul Fox wrote: bert wrote: On 14.06.2010, at 20:57, Paul Fox wrote: bert wrote: On 14.06.2010, at 03:47, James Cameron wrote: However, after letting the machine sit idle for a while (even just after booting, still in the Sugar home screen), the whole machine froze. Power LED was still on. Had to power-cycle. Does not happen all the time, but twice already. Maybe it's my machine (one of the first C-test ones)? your board is a B2. (OLPC board revision D1 -- confusing, i know.) so you'll need to keep power management off to keep your wlan alive in any case. your symptoms are definitely odd, though. can i assume you weren't having similar problems with previous releases? it might be worth reinstalling whatever you were running before -- do you know what that was? Well I didn't update in quite a while. Might have been from last year actually. there's one more technique for getting last gasp information in the case of a crash when you don't have serial. switch to a VT, and enter echo 9 /proc/sys/kernel/printk to enable full console logging. let the machine die -- if it's a kernel panic or oops you should get a stack backtrace on the VT. the bad news is that you'll only get the tail end of the backtrace, because the screen is small relative to the amount of data. paul Strange. Haven't been able to reproduce now for two days ... - Bert - ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: F11-for-XO1.5 Release 10.1.1 Release Candidate 2
On 17.06.2010, at 11:23, Bert Freudenberg wrote: On 15.06.2010, at 16:47, Paul Fox wrote: bert wrote: On 14.06.2010, at 20:57, Paul Fox wrote: bert wrote: On 14.06.2010, at 03:47, James Cameron wrote: However, after letting the machine sit idle for a while (even just after booting, still in the Sugar home screen), the whole machine froze. Power LED was still on. Had to power-cycle. Does not happen all the time, but twice already. Maybe it's my machine (one of the first C-test ones)? your board is a B2. (OLPC board revision D1 -- confusing, i know.) so you'll need to keep power management off to keep your wlan alive in any case. your symptoms are definitely odd, though. can i assume you weren't having similar problems with previous releases? it might be worth reinstalling whatever you were running before -- do you know what that was? Well I didn't update in quite a while. Might have been from last year actually. there's one more technique for getting last gasp information in the case of a crash when you don't have serial. switch to a VT, and enter echo 9 /proc/sys/kernel/printk to enable full console logging. let the machine die -- if it's a kernel panic or oops you should get a stack backtrace on the VT. the bad news is that you'll only get the tail end of the backtrace, because the screen is small relative to the amount of data. paul Strange. Haven't been able to reproduce now for two days ... - Bert - And even stranger, now that I did not enable the printk it happened again. It was sleeping, I touched the pad, power led came on, but screen backlight did not, and it does not respond to keys or touch anymore. Also, it does not go back to sleep according to the power led, so it indeed is frozen I'd say. Mysterious. - Bert - ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: F11-for-XO1.5 Release 10.1.1 Release Candidate 2
On 14.06.2010, at 06:24, Chris Ball wrote: Hi Bert, Copied ~olpc/Activities/Sratch.activity/vm-sound-ALSA plugin to /usr/lib/squeak/3.10-5/. After that, in Etoys playback sounds a bit worse, but it's still okay (hard to tell actually). And no freeze after suspend, playing just resumes. So it would be a good idea to use that ALSA plugin for Squeak in general, me thinks. Could we get a squeak/etoys package with a fixed alsa-plugin ASAP, please? Someone would need to rebuild Fedora's squeak-vm-3.10-5 package with Derek's patched sqUnixSoundALSA.c file (attached to the forwarded msg below). The etoys package is unaffected. Unless you can/want to use his binary, that is. Yes, does sound a bit like hardware. Can I get a replacement? Preferably the HS version once it's ready? ;) - Bert - Begin forwarded message: From: Derek O'Connell d...@doconnel.f9.co.uk Date: 9. Juni 2010 21:33:39 MESZ To: Chris Ball c...@laptop.org Cc: John Maloney jmalo...@media.mit.edu, Bert Freudenberg b...@freudenbergs.de, Paul Fox p...@laptop.org Subject: Re: alsa fix4 fix5 attached. On 09/06/10 18:48, Chris Ball wrote: Hi John, Hi, Chris. I can build a new Scratch XO release. But first a few questions: a. Shouldn't this fixed ALSA go into the Etoys release, where it can be shared by both Scratch and Etoys? Ah, sounds like it. b. If I just put the new ALSA plugin file in the Scratch activity, are we sure it will get loaded? (i.e. does it override the version in the Squeak VM folder)? c. I confess, got busy with other things, so I haven't tested this. I know Derek has. Anyone else? Dunno about these. We can certainly help test a new build. d. Should I remove Bert's shell script code that suppresses power management? Yes, please. e. What's your time frame? We were hoping to release on Friday, but we can afford to delay until this fix is ready. Thanks! - Chris. alsa-xo-1.5-fix5.tar.gz Description: GNU Zip compressed data ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: F11-for-XO1.5 Release 10.1.1 Release Candidate 2
Hi, Bert. Re: Tried sound playback in Scratch, it's choppy, many clicks to hear. Suspended fine, resumed playing sound. Recording does not freeze the activity, but has bad quality. Derek says that the recorded sound is actually okay if you export it and play it with a good sound player. Does that fit with your experience? The choppy sound in Scratch is known; the fact that it works in EToys makes me think that Scratch would work, too, with the right buffer size and sampling rate. I'll try a few variations... -- John ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: F11-for-XO1.5 Release 10.1.1 Release Candidate 2
On Mon, Jun 14, 2010 at 9:58 AM, Bert Freudenberg b...@freudenbergs.de wrote: On 14.06.2010, at 06:24, Chris Ball wrote: Hi Bert, Copied ~olpc/Activities/Sratch.activity/vm-sound-ALSA plugin to /usr/lib/squeak/3.10-5/. After that, in Etoys playback sounds a bit worse, but it's still okay (hard to tell actually). And no freeze after suspend, playing just resumes. So it would be a good idea to use that ALSA plugin for Squeak in general, me thinks. Could we get a squeak/etoys package with a fixed alsa-plugin ASAP, please? Someone would need to rebuild Fedora's squeak-vm-3.10-5 package with Derek's patched sqUnixSoundALSA.c file (attached to the forwarded msg below). The etoys package is unaffected. I'm happy to apply a patch to the etoys package in Fedora if someone can provide me a patch rather than a complete new source file. Peter ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: F11-for-XO1.5 Release 10.1.1 Release Candidate 2
On 14.06.2010, at 13:01, John Maloney wrote: Hi, Bert. Re: Tried sound playback in Scratch, it's choppy, many clicks to hear. Suspended fine, resumed playing sound. Recording does not freeze the activity, but has bad quality. Derek says that the recorded sound is actually okay if you export it and play it with a good sound player. Does that fit with your experience? Yes, it seems to be playback-related. Actually, Sayamindu fixed that already, but it got reverted with Derek's anti-freeze patch. He just made a combined version: http://dev.laptop.org/ticket/9375 - Bert - ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: F11-for-XO1.5 Release 10.1.1 Release Candidate 2
On Mon, Jun 14, 2010 at 12:29 PM, Derek O'Connell d...@doconnel.f9.co.uk wrote: Hi Peter, I don't have much experience creating patches so please let me know if the attached is the wrong format for your purposes. Created using: diff -cB /home/doc/std/dev/squeak/vm/3.10-5/unix-3.10-5/platforms/unix/vm-sound-ALSA/sqUnixSoundALSA.c /home/doc/std/dev/squeak/vm/3.11.3/src/squeak-svn/platforms/unix/vm-sound-ALSA/sqUnixSoundALSA.c sqUnixSoundALSA.patch If you could use 'diff -u origfile newfile output.patch' from the root of the source tree that would be better. Thanks, Peter On 14/06/10 11:58, Peter Robinson wrote: On Mon, Jun 14, 2010 at 9:58 AM, Bert Freudenbergb...@freudenbergs.de wrote: On 14.06.2010, at 06:24, Chris Ball wrote: Hi Bert, Copied ~olpc/Activities/Sratch.activity/vm-sound-ALSA plugin to /usr/lib/squeak/3.10-5/. After that, in Etoys playback sounds a bit worse, but it's still okay (hard to tell actually). And no freeze after suspend, playing just resumes. So it would be a good idea to use that ALSA plugin for Squeak in general, me thinks. Could we get a squeak/etoys package with a fixed alsa-plugin ASAP, please? Someone would need to rebuild Fedora's squeak-vm-3.10-5 package with Derek's patched sqUnixSoundALSA.c file (attached to the forwarded msg below). The etoys package is unaffected. I'm happy to apply a patch to the etoys package in Fedora if someone can provide me a patch rather than a complete new source file. Peter ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: F11-for-XO1.5 Release 10.1.1 Release Candidate 2
Hi, Bert and Derek. Derek just send me the combined version. I tested playback and recording on the XO 1.5 and it was perfect. Hooray! I did NOT test to verify that it survives suspect resume, as Derek's previous version did. Hopefully someone else has verified that. It sounds as though this fixed version of the ALSA plugin will get into the XO Squeak VM package. In that case, should I omit it from the next version of Scratch? (A new version of XO Scratch is coming soon to include the Lego WeDo plugin which Derek just finished.) Many thanks to Derek and Sayamindu for their hard work. -- John On Jun 14, 2010, at 7:08 AM, Bert Freudenberg wrote: On 14.06.2010, at 13:01, John Maloney wrote: Hi, Bert. Re: Tried sound playback in Scratch, it's choppy, many clicks to hear. Suspended fine, resumed playing sound. Recording does not freeze the activity, but has bad quality. Derek says that the recorded sound is actually okay if you export it and play it with a good sound player. Does that fit with your experience? Yes, it seems to be playback-related. Actually, Sayamindu fixed that already, but it got reverted with Derek's anti-freeze patch. He just made a combined version: http://dev.laptop.org/ticket/9375 - Bert - ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: F11-for-XO1.5 Release 10.1.1 Release Candidate 2
Hi, Which source tree root are you referring to? The sources for the current Fedora package. A source rpm has a tarball of the sources, and optionally a series of patch files relative to that upstream tarball. Not sure where to find those sources though, I'm not too familiar with Fedora procedures. Here's the latest build: http://koji.fedoraproject.org/koji/buildinfo?buildID=148721 The .src.rpm contains a specfile (that describes which patches to apply on the tarball) and the tarball and patches. If you want to unpack the .src.rpm on a machine that isn't running Fedora, I think you'll need rpm2cpio and cpio: rpm2cpio squeak-vm-3.10.5-2.fc11.src.rpm | cpio -idmv Thanks, - Chris. -- Chris Ball c...@laptop.org One Laptop Per Child ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: [Testing] F11-for-XO1.5 Release 10.1.1 Release Candidate 2
On 14.06.2010, at 03:47, James Cameron wrote: On Sun, Jun 13, 2010 at 05:32:12PM +0200, Bert Freudenberg wrote: Installed on my 1.5 after updating the firmware to q3a39. Typing boot in OFW right after fs-update froze the machine. Unpredictable behaviour is known to occur if you try to boot after fs-update, you should use bye. Ah, okay. But that shouldn't have done permanent damage, right? Should I reflash? Powered off (long-press on power button) and booted. Let it sit idle for a while in the first-time Sugar screen (name dialog) - machine froze. It should idle suspend. It should not freeze. It does idle suspend. When the LED starts blinking and I touch the pad it wakes up fine. Just when I come back after some time of idling, it doesn't wake up anymore. The freeze happens sometimes with the screen still lit, sometimes after it turned off. However, after letting the machine sit idle for a while (even just after booting, still in the Sugar home screen), the whole machine froze. Power LED was still on. Had to power-cycle. Does not happen all the time, but twice already. Maybe it's my machine (one of the first C-test ones)? Sounds bad. Do you have a serial port attached? I ask because I suspect a kernel panic and a serial port is a practical way to obtain more problem data. No, I don't have one. Might also be worth running memtest from OFW as well, just to exclude certain other causes. Ran memtest (from 30m up as Richard suggested), passed, no errors. I also took out the battery to make sure everything is reset. Still freezes when I let it sit long enough. - Bert - ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: [Testing] F11-for-XO1.5 Release 10.1.1 Release Candidate 2
bert wrote: On 14.06.2010, at 03:47, James Cameron wrote: On Sun, Jun 13, 2010 at 05:32:12PM +0200, Bert Freudenberg wrote: Installed on my 1.5 after updating the firmware to q3a39. Typing boot in OFW right after fs-update froze the machine. Unpredictable behaviour is known to occur if you try to boot after fs-update, you should use bye. Ah, okay. But that shouldn't have done permanent damage, right? Should I reflash? Powered off (long-press on power button) and booted. Let it sit idle for a while in the first-time Sugar screen (name dialog) - machine froze. It should idle suspend. It should not freeze. It does idle suspend. When the LED starts blinking and I touch the pad it wakes up fine. Just when I come back after some time of idling, it doesn't wake up anymore. The freeze happens sometimes with the screen still lit, sometimes after it turned off. However, after letting the machine sit idle for a while (even just after booting, still in the Sugar home screen), the whole machine froze. Power LED was still on. Had to power-cycle. Does not happen all the time, but twice already. Maybe it's my machine (one of the first C-test ones)? Sounds bad. Do you have a serial port attached? I ask because I suspect a kernel panic and a serial port is a practical way to obtain more problem data. No, I don't have one. the next best thing to having a serial port is to edit /etc/rsyslog.conf, and change the destination of all the logs from /var/log to somewhere nonvolatile, like /home/olpc/log (be sure to create the directory). then either reboot, or killall -HUP rsyslogd to make the config change take effect. this might give some information on what was going on before the hang. paul Might also be worth running memtest from OFW as well, just to exclude certain other causes. Ran memtest (from 30m up as Richard suggested), passed, no errors. I also took out the battery to make sure everything is reset. Still freezes when I let it sit long enough. - Bert - ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel =- paul fox, p...@laptop.org ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: F11-for-XO1.5 Release 10.1.1 Release Candidate 2
Hi Peter,
I don't have much experience creating patches so please let me know if
the attached is the wrong format for your purposes. Created using:
diff -cB
/home/doc/std/dev/squeak/vm/3.10-5/unix-3.10-5/platforms/unix/vm-sound-ALSA/sqUnixSoundALSA.c
/home/doc/std/dev/squeak/vm/3.11.3/src/squeak-svn/platforms/unix/vm-sound-ALSA/sqUnixSoundALSA.c
sqUnixSoundALSA.patch
-D
On 14/06/10 11:58, Peter Robinson wrote:
On Mon, Jun 14, 2010 at 9:58 AM, Bert Freudenbergb...@freudenbergs.de wrote:
On 14.06.2010, at 06:24, Chris Ball wrote:
Hi Bert,
Copied ~olpc/Activities/Sratch.activity/vm-sound-ALSA plugin to
/usr/lib/squeak/3.10-5/.
After that, in Etoys playback sounds a bit worse, but it's still
okay (hard to tell actually). And no freeze after suspend,
playing just resumes. So it would be a good idea to use that ALSA
plugin for Squeak in general, me thinks.
Could we get a squeak/etoys package with a fixed alsa-plugin ASAP, please?
Someone would need to rebuild Fedora's squeak-vm-3.10-5 package with Derek's
patched sqUnixSoundALSA.c file (attached to the forwarded msg below). The etoys
package is unaffected.
I'm happy to apply a patch to the etoys package in Fedora if someone
can provide me a patch rather than a complete new source file.
Peter
*** /home/doc/std/dev/squeak/vm/3.10-5/unix-3.10-5/platforms/unix/vm-sound-ALSA/sqUnixSoundALSA.c 2008-04-28 20:04:11.0 +0100
--- /home/doc/std/dev/squeak/vm/3.11.3/src/squeak-svn/platforms/unix/vm-sound-ALSA/sqUnixSoundALSA.c 2010-06-14 12:22:13.0 +0100
***
*** 135,141
if ((err= snd_pcm_start(output_handle)) 0)
{
if (err != -EPIPE)
! {
fprintf(stderr, snd_pcm_start(1): %s\n, snd_strerror(err));
success(false);
return 0;
--- 136,143
if ((err= snd_pcm_start(output_handle)) 0)
{
if (err != -EPIPE)
! /* if ((err != -EPIPE) (err != -ESTRPIPE))
! */ {
fprintf(stderr, snd_pcm_start(1): %s\n, snd_strerror(err));
success(false);
return 0;
***
*** 148,154
if ((err= snd_pcm_start(output_handle)) 0)
{
if (err != -EPIPE)
! {
fprintf(stderr, snd_pcm_start(2): %s\n, snd_strerror(err));
success(false);
return 0;
--- 150,157
if ((err= snd_pcm_start(output_handle)) 0)
{
if (err != -EPIPE)
! /* if ((err != -EPIPE) (err != -ESTRPIPE))
! */ {
fprintf(stderr, snd_pcm_start(2): %s\n, snd_strerror(err));
success(false);
return 0;
***
*** 168,173
--- 171,177
if (!output_handle) return 0;
snd_pcm_delay(output_handle, delay);
+ snd_pcm_avail_update(output_handle);
state= snd_pcm_state (output_handle);
/* if underrun causes, max delay is loosened */
***
*** 199,226
static sqInt sound_PlaySamplesFromAtLength(sqInt frameCount, sqInt arrayIndex, sqInt startIndex)
{
! if (output_handle)
! {
! void *samples= (void *)arrayIndex + startIndex * output_channels * 2;
! int count= snd_pcm_writei(output_handle, samples, frameCount);
! if (count frameCount / 2)
! {
! output_buffer_frames_available= 0;
! }
! if (count 0)
! {
! if (count == -EPIPE)/* underrun */
! {
! int err;
! snd(pcm_prepare(output_handle), sound_PlaySamples: snd_pcm_prepare);
! return 0;
! }
! fprintf(stderr, snd_pcm_writei returned %i\n, count);
! return 0;
! }
! return count;
}
! success(false);
return 0;
}
--- 203,246
static sqInt sound_PlaySamplesFromAtLength(sqInt frameCount, sqInt arrayIndex, sqInt startIndex)
{
! if (!output_handle)
! {
! success(false);
! return 0;
! }
!
! void *samples= (void *)arrayIndex + startIndex * output_channels * 2;
! int count= snd_pcm_writei(output_handle, samples, frameCount);
! if (count frameCount / 2)
! output_buffer_frames_available= 0;
!
! if (count = 0)
! return count;
!
! if (count != -EPIPE count != -ESTRPIPE)
! {
! fprintf(stderr, snd_pcm_writei returned %i\n, count);
! return 0;
! }
!
! int err;
! if (count == -EPIPE) { /* under-run */
! err = snd_pcm_prepare (output_handle);
! if (err 0)
! printf(Can't recovery from underrun, prepare failed: %s, snd_strerror (err));
! return 0;
! } else if (count == -ESTRPIPE) {
! while ((err = snd_pcm_resume (output_handle)) == -EAGAIN)
! sleep(1); /* wait until the suspend flag is released */
!
! if (err 0) {
! err = snd_pcm_prepare (output_handle);
! if (err 0)
! printf(Can't recovery from suspend, prepare failed: %s, snd_strerror (err));
}
! return 0;
! }
!
return 0;
}
***
*** 306,318
int frameCount= ((bufferSizeInBytes / 2) - startSliceIndex) / input_channels;
int count= snd_pcm_readi(input_handle,
Re: F11-for-XO1.5 Release 10.1.1 Release Candidate 2
On 14/06/10 12:56, Peter Robinson wrote: On Mon, Jun 14, 2010 at 12:29 PM, Derek O'Connelld...@doconnel.f9.co.uk wrote: Hi Peter, I don't have much experience creating patches so please let me know if the attached is the wrong format for your purposes. Created using: diff -cB /home/doc/std/dev/squeak/vm/3.10-5/unix-3.10-5/platforms/unix/vm-sound-ALSA/sqUnixSoundALSA.c /home/doc/std/dev/squeak/vm/3.11.3/src/squeak-svn/platforms/unix/vm-sound-ALSA/sqUnixSoundALSA.c sqUnixSoundALSA.patch If you could use 'diff -u origfile newfile output.patch' from the root of the source tree that would be better. Which source tree root are you referring to? Chris mentioned 3.10-5 and I do most my development on 3.11.3 so I diff'ed between them. It's a bad state of affairs but I keep several versions of the VM source around and would normally simply email the complete source file to Ian Piumarta (as he requests). -D Thanks, Peter On 14/06/10 11:58, Peter Robinson wrote: On Mon, Jun 14, 2010 at 9:58 AM, Bert Freudenbergb...@freudenbergs.de wrote: On 14.06.2010, at 06:24, Chris Ball wrote: Hi Bert, Copied ~olpc/Activities/Sratch.activity/vm-sound-ALSA plugin to /usr/lib/squeak/3.10-5/. After that, in Etoys playback sounds a bit worse, but it's still okay (hard to tell actually). And no freeze after suspend, playing just resumes. So it would be a good idea to use that ALSA plugin for Squeak in general, me thinks. Could we get a squeak/etoys package with a fixed alsa-plugin ASAP, please? Someone would need to rebuild Fedora's squeak-vm-3.10-5 package with Derek's patched sqUnixSoundALSA.c file (attached to the forwarded msg below). The etoys package is unaffected. I'm happy to apply a patch to the etoys package in Fedora if someone can provide me a patch rather than a complete new source file. Peter ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: F11-for-XO1.5 Release 10.1.1 Release Candidate 2
On 14/06/10 13:21, Bert Freudenberg wrote: On 14.06.2010, at 14:15, Derek O'Connell wrote: On 14/06/10 12:56, Peter Robinson wrote: On Mon, Jun 14, 2010 at 12:29 PM, Derek O'Connelld...@doconnel.f9.co.uk wrote: Hi Peter, I don't have much experience creating patches so please let me know if the attached is the wrong format for your purposes. Created using: diff -cB /home/doc/std/dev/squeak/vm/3.10-5/unix-3.10-5/platforms/unix/vm-sound-ALSA/sqUnixSoundALSA.c /home/doc/std/dev/squeak/vm/3.11.3/src/squeak-svn/platforms/unix/vm-sound-ALSA/sqUnixSoundALSA.c sqUnixSoundALSA.patch If you could use 'diff -u origfile newfile output.patch' from the root of the source tree that would be better. Which source tree root are you referring to? The sources for the current Fedora package. A source rpm has a tarball of the sources, and optionally a series of patch files relative to that upstream tarball. Not sure where to find those sources though, I'm not too familiar with Fedora procedures. Me neither but thanks for the pointer. - Bert - ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: F11-for-XO1.5 Release 10.1.1 Release Candidate 2
Hi Peter,
sorry for the delay, hopefully the attach is what you want.
-D
On 14/06/10 12:56, Peter Robinson wrote:
On Mon, Jun 14, 2010 at 12:29 PM, Derek O'Connelld...@doconnel.f9.co.uk
wrote:
Hi Peter,
I don't have much experience creating patches so please let me know if the
attached is the wrong format for your purposes. Created using:
diff -cB
/home/doc/std/dev/squeak/vm/3.10-5/unix-3.10-5/platforms/unix/vm-sound-ALSA/sqUnixSoundALSA.c
/home/doc/std/dev/squeak/vm/3.11.3/src/squeak-svn/platforms/unix/vm-sound-ALSA/sqUnixSoundALSA.c
sqUnixSoundALSA.patch
If you could use 'diff -u origfile newfile output.patch' from the
root of the source tree that would be better.
Thanks,
Peter
On 14/06/10 11:58, Peter Robinson wrote:
On Mon, Jun 14, 2010 at 9:58 AM, Bert Freudenbergb...@freudenbergs.de
wrote:
On 14.06.2010, at 06:24, Chris Ball wrote:
Hi Bert,
Copied ~olpc/Activities/Sratch.activity/vm-sound-ALSA plugin to
/usr/lib/squeak/3.10-5/.
After that, in Etoys playback sounds a bit worse, but it's still
okay (hard to tell actually). And no freeze after suspend,
playing just resumes. So it would be a good idea to use that ALSA
plugin for Squeak in general, me thinks.
Could we get a squeak/etoys package with a fixed alsa-plugin ASAP,
please?
Someone would need to rebuild Fedora's squeak-vm-3.10-5 package with
Derek's patched sqUnixSoundALSA.c file (attached to the forwarded msg
below). The etoys package is unaffected.
I'm happy to apply a patch to the etoys package in Fedora if someone
can provide me a patch rather than a complete new source file.
Peter
--- Squeak-3.10-5/platforms/unix/vm-sound-ALSA/sqUnixSoundALSA.c 2008-04-21 22:48:26.0 +0100
+++ Squeak-3.10-5/platforms/unix/vm-sound-ALSA/sqUnixSoundALSA-100614.c 2010-06-14 12:22:13.0 +0100
@@ -77,6 +77,7 @@
static snd_pcm_uframes_t output_buffer_size= 0;
static double max_delay_frames= 0;
+
static void output_callback(snd_async_handler_t *handler)
{
signalSemaphoreWithIndex(output_semaphore);
@@ -135,7 +136,8 @@
if ((err= snd_pcm_start(output_handle)) 0)
{
if (err != -EPIPE)
- {
+/* if ((err != -EPIPE) (err != -ESTRPIPE))
+*/ {
fprintf(stderr, snd_pcm_start(1): %s\n, snd_strerror(err));
success(false);
return 0;
@@ -148,7 +150,8 @@
if ((err= snd_pcm_start(output_handle)) 0)
{
if (err != -EPIPE)
- {
+/* if ((err != -EPIPE) (err != -ESTRPIPE))
+*/ {
fprintf(stderr, snd_pcm_start(2): %s\n, snd_strerror(err));
success(false);
return 0;
@@ -168,6 +171,7 @@
if (!output_handle) return 0;
snd_pcm_delay(output_handle, delay);
+ snd_pcm_avail_update(output_handle);
state= snd_pcm_state (output_handle);
/* if underrun causes, max delay is loosened */
@@ -199,28 +203,44 @@
static sqInt sound_PlaySamplesFromAtLength(sqInt frameCount, sqInt arrayIndex, sqInt startIndex)
{
- if (output_handle)
-{
- void *samples= (void *)arrayIndex + startIndex * output_channels * 2;
- int count= snd_pcm_writei(output_handle, samples, frameCount);
- if (count frameCount / 2)
- {
- output_buffer_frames_available= 0;
- }
- if (count 0)
- {
- if (count == -EPIPE)/* underrun */
- {
- int err;
- snd(pcm_prepare(output_handle), sound_PlaySamples: snd_pcm_prepare);
- return 0;
- }
- fprintf(stderr, snd_pcm_writei returned %i\n, count);
- return 0;
- }
- return count;
+ if (!output_handle)
+ {
+success(false);
+return 0;
+ }
+
+ void *samples= (void *)arrayIndex + startIndex * output_channels * 2;
+ int count= snd_pcm_writei(output_handle, samples, frameCount);
+ if (count frameCount / 2)
+output_buffer_frames_available= 0;
+
+ if (count = 0)
+return count;
+
+ if (count != -EPIPE count != -ESTRPIPE)
+ {
+fprintf(stderr, snd_pcm_writei returned %i\n, count);
+return 0;
+ }
+
+ int err;
+ if (count == -EPIPE) { /* under-run */
+err = snd_pcm_prepare (output_handle);
+if (err 0)
+ printf(Can't recovery from underrun, prepare failed: %s, snd_strerror (err));
+return 0;
+ } else if (count == -ESTRPIPE) {
+while ((err = snd_pcm_resume (output_handle)) == -EAGAIN)
+ sleep(1); /* wait until the suspend flag is released */
+
+if (err 0) {
+ err = snd_pcm_prepare (output_handle);
+ if (err 0)
+ printf(Can't recovery from suspend, prepare failed: %s, snd_strerror (err));
}
- success(false);
+return 0;
+ }
+
return 0;
}
@@ -306,13 +326,28 @@
int frameCount= ((bufferSizeInBytes / 2) - startSliceIndex) / input_channels;
int count= snd_pcm_readi(input_handle, samples, frameCount);
if (count 0)
- {
- if (count == -EPIPE)
- snd_pcm_prepare(input_handle);
- else if (count != -EAGAIN)
- fprintf(stderr, snd_pcm_readi
Re: F11-for-XO1.5 Release 10.1.1 Release Candidate 2
On Sat, Jun 12, 2010 at 12:46:05AM -0400, Chris Ball wrote: This build is the second RC for the 10.1.1 release; if no new blockers appear, it will become the final release after a period of testing. http://wiki.laptop.org/go/Release_notes/10.1.1 has been drafted, Further edits are welcome. In particular: 0. are there any interesting new features or notable bugs that should be listed? (e.g. does my sense of interest and notability match yours?) 1. is the installation process correct in other circumstances? (it works for me, requires less keyboard use on the XO-1.5, and has less risk of bad experience caused by issuing commands after fs-update), 2. will the build be on download.laptop.org or build.laptop.org? (presuming the latter for the time being), 3. are there any pictures available of holding down all four game keys? 4. has anyone recently tested olpc-update from os64 to os20x? -- James Cameron http://quozl.linux.org.au/ ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: [Testing] F11-for-XO1.5 Release 10.1.1 Release Candidate 2
Might also be worth running memtest from OFW as well, just to exclude certain other causes. memtest will puke errors unless you set the lower address range to 30M. Dunno why. Guess a bug needs top be filed for sorting it out. If you run memtest hit 'c' right after it starts to go into config then '2' for address range '1' for lower limit. Enter '30m' press return then press '0' to restart the tests. -- Richard A. Smith ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: F11-for-XO1.5 Release 10.1.1 Release Candidate 2
Hi Bert, Copied ~olpc/Activities/Sratch.activity/vm-sound-ALSA plugin to /usr/lib/squeak/3.10-5/. After that, in Etoys playback sounds a bit worse, but it's still okay (hard to tell actually). And no freeze after suspend, playing just resumes. So it would be a good idea to use that ALSA plugin for Squeak in general, me thinks. Could we get a squeak/etoys package with a fixed alsa-plugin ASAP, please? Had to power-cycle. Does not happen all the time, but twice already. Maybe it's my machine (one of the first C-test ones)? Yes, does sound a bit like hardware. Thanks, - Chris. -- Chris Ball c...@laptop.org One Laptop Per Child ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: F11-for-XO1.5 Release 10.1.1 Release Candidate 2
--- On Sat, 6/12/10, Chris Ball c...@laptop.org wrote: From: Chris Ball c...@laptop.org Subject: F11-for-XO1.5 Release 10.1.1 Release Candidate 2 To: Fedora OLPC fedora-olpc-l...@redhat.com Cc: test...@lists.laptop.org, Devel devel@lists.laptop.org Date: Saturday, June 12, 2010, 12:46 AM http://wiki.laptop.org/go/F11_for_1.5 http://build.laptop.org/10.1.1/os203 Compressed image size: 705.34mb (+13.94mb since build 202) This build is the second RC for the 10.1.1 release; if no new blockers appear, it will become the final release after a period of testing. Some bugs that we'd like particular help testing: * http://dev.laptop.org/ticket/10045 : - Fix loss of audio/video sync in Record activity. Some instances of sync loss may remain, especially if you're running other activities at the same time as Record; we'd like to hear your feedback. * http://dev.laptop.org/ticket/10168 : - Fix Scratch can freeze up after playing sound. See reproduction instructions in the ticket. Thanks! Wow 2 builds in a day! You must be working late ;- At this stage the only difference between os203 and os126 looks like some newer f11 packages, correct? They both work fine (clean installed one and olpc-update the other) and Record-81 is a BIG improvement. Syncing is almost perfect (in my finger snap test :) up to a minute that I tested, operation is snappy, while image display while recording is pretty smooth. However, besides the annoying 5sec silent audio added in the beginning of audio recordings that Mikus described and persists, there is still an old nasty one with data-loss potential. In short trying to record a clip after you recored, played and deleted one, quits the application. Filed here http://bugs.sugarlabs.org/ticket/2041 Other activities worked fine (even Measure 29) Sharing with an XO-1/os15 through an AP was smooth and snappy, initiating activities either from the XO-1.5 or the XO-1. ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: F11-for-XO1.5 Release 10.1.1 Release Candidate 2
Hi, Wow 2 builds in a day! You must be working late ;- At this stage the only difference between os203 and os126 looks like some newer f11 packages, correct? Yes, that's exactly right. I didn't take the newer F11 updates into the stable branch. http://bugs.sugarlabs.org/ticket/2041 Interesting, thanks. Could you run olpc-log and attach the generated tarball to the bug? (I'm particularly interested in dmesg output.) - Chris. -- Chris Ball c...@laptop.org One Laptop Per Child ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
