September 2011 - Wine-devel

What is the PulseAudio underrun bug?
by Joerg-Cyril.Hoehle＠t-systems.com 09 Sep '11

09 Sep '11

Hi, As I see that the PA team is preparing release candidates for 1.0, I wonder whether the Wine team communicated them the issues we face. Browsing open tickets http://pulseaudio.org/report/1 I found nothing obvious. The PA bug, as I face it in Ubuntu Lucid, is not something that happens randomly after N minutes or hours. It is *immediately* detectable. Symptoms: 1. winecfg audio test hangs. But Wine is a complex program, so I wrote a trivial one as follows: - set generic hw_params - prefill snd_pcm_write with e.g. 100ms of data; - start ALSA or use hw_params_set_start_threshold; - print out snd_pcm_state, snd_pcm_avail_update and snd_pcm_delay in a loop. 2. avail_update and avail_delay will freeze approx. at the time underrun would have happened with plug:dmix or plughw:0. In particular, delay will not decrease below < 0. By contrast, dmix or hw would see avail_update grow continuously, far above buffer_size, while delay decreases towards minus infinity -- regardless of whether XRUN state was entered or still RUNNING playing silence. 3. Optionally, XRUN mode is never entered, PA claims to be RUNNING. (XRUN is not a necessity, depending on hw_params_set_silence_*) >From there on, more snd_pcm_write(N frames) will not be heard, and avail_update() and delay() will solely bump to reflect that N frames have been queued, and remain frozen. snd_pcm_reset() won't help. snd_pcm_drain() will surprisingly play the buffered audio. Note that in non-blocking mode, this function will return -11 with "dmix" or "hw" ALSA devices, so here again PA differs from others. It looks like PA's front end (claiming to be running) has lost sync with its backend (which apparently stopped). 4. I don't know why and when, but the damage is not always limited to this one process. Sometimes the PA server will need to be restarted (pactl exit) otherwise no more sound will play for more than a click's time. It obviously corrupted its data structures. Andrew Eikum found out that activating the handle_underrun configuration in the newest alsa plugin package (released 2011-01-31!) allows to work around the bug. However when I read the comment to the patch, http://git.alsa-project.org/?p=alsa-plugins.git;a=commit;h=c20d516e22962012… I've the impression that using handle_underrun is backwards and is *not* the recommended way to use PulseAudio. IOW, PA without handle_underrun is subject to the underrun bug and that is abnormal behaviour. Or was it? Does the brandnew PA RC exhibit sane behaviour? Which is the "need handle_underrun" ticket in PA's bug tracker? http://pulseaudio.org/report/1 Where on Ubuntu's launchpad? What do they now about the issue? Regards, Jörg Höhle

2 1

user32: Avoid drawing the caret out of the window
by Bruno Jesus 09 Sep '11

09 Sep '11

Please, ignore this patch. It's overkill to check the bounding box during every caret timer callback. I have found a better approach to this by checking inside SetCaretPos/ShowCaret. That's because the caret doesn't move by itself so it's useless to check everytime. Thanks, Bruno -- universe* god::bigbang (void); //and then it all began...

1 0

What is the PulseAudio underrun bug?
by Joerg-Cyril.Hoehle＠t-systems.com 09 Sep '11

09 Sep '11

Hi again, Something that I still have trouble understanding is: Why did I find out only 2 weeks ago that Ubuntu Lucid is affected? Perhaps because I'm mostly using Intrepid which works better as far as the dreaded Intel Xorg screen flicker, suspend/resume and audio are concerned. Why do the Wine users only observe random drop outs after some time? Why does it look like the new mmdevapi code accentuated the problem (does it?)? Perhaps because most apps use dsound and dsound does efforts to send a continuous stream of data to the audio device, without underruns. Likewise, MCIWAVE/MCIAVI/winmm:PlaySound send data in a tight loop. Perhaps because the current GetPosition in both mmdevapi and winmm is not trustworthy and may cause underruns when a sound ends? Perhaps because "underrun at end" is almost designed into winmm? (unless you end with waveOutReset) Also, there may be more than underruns which is broken in PA. Underruns do not explain why I sometimes need to restart the PA server, as mentioned in my prior point 4. Do underruns really happen on my xGHz machine? On my Ubuntu Lucid system, the dreaded Intel XOrg screen flicker is enough to cause some apps running at that time to lose audio. Moving from one virtual desktop or screen to another may cause paging. That's no proof, but underruns may happen more often than one would like. Apps must robustly handle them. Regards, Jörg Höhle

1 0

Re: winspool.drv: Correct test for Device Capabilities
by Marvin 09 Sep '11

09 Sep '11

Hi, While running your changed tests on Windows, I think I found new failures. Being a bot and all I'm not very good at pattern recognition, so I might be wrong, but could you please double-check? Full results can be found at http://testbot.winehq.org/JobDetails.pl?Key=14060 Your paranoid android. === WINEBUILD (build) === Can't determine base name

3 2

Auto-selecting local variable/parameters when 'p'rinting values using winedbg?
by Frédéric Delanoy 09 Sep '11

09 Sep '11

While debugging with winedbg, I sometimes stumble upon situations like those (especially for widely used names) Many symbols with name 'foo', choose the one you want (<cr> to abort): [1]: 0x1234567 foo in libwhatever.so.1 [2]: Local variable (in a register) of foo => 2 This can be cumbersome, e.g. when you don't want to see foo val in libc.so. Is it somehow possible to auto-select the "local value"? Frédéric

1 0

Re: [2/8] msi/test: Mark some test results as broken.
by Marvin 09 Sep '11

09 Sep '11

Hi, While running your changed tests on Windows, I think I found new failures. Being a bot and all I'm not very good at pattern recognition, so I might be wrong, but could you please double-check? Full results can be found at http://testbot.winehq.org/JobDetails.pl?Key=14087 Your paranoid android. === WNT4WSSP6 (32 bit install) === install.c:6000: Test failed: Per-user icon file isn't where it's expected (C:\WINNT\Profiles\Default User\Application Data\Microsoft\Installer\{7DF88A49-996F-4EC8-A022-BF956F9B2CBB}\testicon)

1 0

Re: cmd: GOTO is a no-op and doesn't output anything in interactive mode
by Frédéric Delanoy 09 Sep '11

09 Sep '11

2011/9/9 Frédéric Delanoy <frederic.delanoy(a)gmail.com>: > --- > programs/cmd/builtins.c | 9 +++++---- > 1 files changed, 5 insertions(+), 4 deletions(-) > This patch depends on my other non-yet-commited goto patch (Avoid rereading batch file for every call/goto executed) Shouldn't probably have sent it yet. Sorry for the inconvenience. Frédéric

1 0

crypt32: Test/correct CertGetNameString with NULL pvTypePara
by Juan Lang 08 Sep '11

08 Sep '11

Fix for bug 23287. --Juan

1 0

[PATCH 1/2] mmdevapi: IsFormatSupported fills closest match iff it returns S_FALSE.
by Joerg-Cyril.Hoehle＠t-systems.com 08 Sep '11

08 Sep '11

Andrew Eikum asked: >It would be nice to have tests for this, >Are these part of the tests you're >planning on submitting after you get the fixes in? That is part of my tests, the latest testbot job is https://testbot.winehq.org/JobDetails.pl?Key=14067 You can see that I still need to compute position estimates based on the high performance timer as Sleep() is not accurate enough. >and a fix for coreaudio if needed. Likewise for patch 2/2. The Mac never returns S_FALSE, so there was nothing to change. I've yet another patch about IsFormatSupported in my queue, but it needs more testing. For instance, I've not yet tested inconsistent formats (bad blockalign) or non PCM formats (e.g. A-law, ADPCM). In short, it looks like IsFormatSupported simply suggests GetMixFormat() in shared mode, but would it suggest linear PCM when given A-law? Regards, Jörg Höhle

1 0

Re: [PATCH 1/2] mmdevapi: IsFormatSupported fills closest match iff it returns S_FALSE.
by Octavian Voicu 08 Sep '11

08 Sep '11

--- a/dlls/winealsa.drv/mmdevdrv.c+++ b/dlls/winealsa.drv/mmdevdrv.c> @@ -1310,17 +1313,14 @@ exit: > LeaveCriticalSection(&This->lock); > HeapFree(GetProcessHeap(), 0, formats); > > - if(hr == S_OK || !out){ > - CoTaskMemFree(closest); > - if(out) > - *out = NULL; > - }else if(closest){ > + if(hr == S_FALSE && out) { > closest->nBlockAlign = > closest->nChannels * closest->wBitsPerSample / 8; > closest->nAvgBytesPerSec = > closest->nBlockAlign * closest->nSamplesPerSec; > *out = closest; > - } > + } else > + CoTaskMemFree(closest); You are dereferencing 'closest' without checking if it's non-NULL. It can be NULL if control jumps to the 'exit' label for several cases. This only applies to winealsa; for wineoss there are no gotos. Octavian

1 1