Opened 16 years ago
Closed 16 years ago
#5283 closed defect (fixed)
startup with non unity timestretch doesnt work properly
Reported by: | Mark Spieth | Owned by: | Isaac Richards |
---|---|---|---|
Priority: | minor | Milestone: | unknown |
Component: | mythtv | Version: | unknown |
Severity: | medium | Keywords: | |
Cc: | Ticket locked: | no |
Description
if a program/group has a non 1.0 timestretch associated with it, timestretch is not initialized correctly.
whilst this is primarily due to multiple calls to reconfigure which are probably not necessary, the underlying support function does not cater for removal of timestretch without the subsequent stretch value not being 1.0
attached patch allows correct reinitialization of timestretch within reconfigure.
Attachments (5)
Change History (8)
Changed 16 years ago by
Attachment: | mythtv_initialtimerstretch.1.patch added |
---|
comment:1 Changed 16 years ago by
comment:2 Changed 16 years ago by
Interesting. Using a kernel command line of "nosmp maxcpus=1" does *not* fix the segfault, but it does make the BT different. It still dies on a memcpy, but now it is at around line 1033 of audiooutputbase.cpp (SVN: 16252):
(gdb) up #1 0xb6a08033 in AudioOutputBase::_AddSamples (this=0xac5c0008, buffer=0x94aa680, interleaved=true, samples=1536, timecode=24836170) at audiooutputbase.cpp:1033 //len = WaitForFreeSpace(amount); char *ob = encoder->GetOutBuff(); if (amount >= bdiff) { memcpy(audiobuffer + org_waud, ob, bdiff); ob += bdiff; amount -= bdiff; org_waud = 0; } if (amount > 0) ----> memcpy(audiobuffer + org_waud, ob, amount); bdiff = AUDBUFSIZE - amount; org_waud += amount; } (gdb) print org_waud $2 = 1536000 (gdb) print ob $3 = 0x9549200 "rø\037N\001" (gdb) print amount $4 = 6144 URL: http://cvs.mythtv.org/svn/branches/release-0-21-fixes/mythtv Repository Root: http://cvs.mythtv.org/svn Repository UUID: 7dbf422c-18fa-0310-86e9-fd20926502f2 Revision: 17246
I am attaching a trace and the BT.
comment:3 Changed 16 years ago by
Resolution: | → fixed |
---|---|
Status: | new → closed |
Changed 16 years ago by
Attachment: | mythtv_initialtimerstretch.2.patch added |
---|
release-0-21-fixes version of patch, including fix for triggered overflow
Mark says my segfault is not part of the audio stuff. That being said, my mythfrontend randomly segfaults if "mythtv_initialtimerstretch.1.patch" is applied. I am attaching a BT.