Recherche avancée

Sur d’autres sites (10424)

• avutil/threadmessage : split the pthread condition in two

  1er décembre 2015, par Clément Bœsch

  avutil/threadmessage : split the pthread condition in two
  Fix a dead lock under certain conditions. Let’s assume we have a queue of 1
  
  message max, 2 senders, and 1 receiver.
  Scenario (real record obtained with debug added) :
  
      [...]
  
      SENDER #0 : acquired lock
  
      SENDER #0 : queue is full, wait
  
      SENDER #1 : acquired lock
  
      SENDER #1 : queue is full, wait
  
      RECEIVER : acquired lock
  
      RECEIVER : reading a msg from the queue
  
      RECEIVER : signal the cond
  
      RECEIVER : acquired lock
  
      RECEIVER : queue is empty, wait
  
      SENDER #0 : writing a msg the queue
  
      SENDER #0 : signal the cond
  
      SENDER #0 : acquired lock
  
      SENDER #0 : queue is full, wait
  
      SENDER #1 : queue is full, wait
  Translated :
  
   - initially the queue contains 1/1 message with 2 senders blocking on
  
     it, waiting to push another message.
  
   - Meanwhile the receiver is obtaining the lock, read the message,
  
     signal & release the lock. For some reason it is able to acquire the
  
     lock again before the signal wakes up one of the sender. Since it
  
     just emptied the queue, the reader waits for the queue to fill up
  
     again.
  
   - The signal finally reaches one of the sender, which writes a message
  
     and then signal the condition. Unfortunately, instead of waking up
  
     the reader, it actually wakes up the other worker (signal = notify
  
     the condition just for 1 waiter), who can’t push another message in
  
     the queue because it’s full.
  
   - Meanwhile, the receiver is still waiting. Deadlock.
  This scenario can be triggered with for example :
  
      tests/api/api-threadmessage-test 1 2 100 100 1 1000 1000
  One working solution is to make av_thread_message_queue_send,recv()
  
  call pthread_cond_broadcast() instead of pthread_cond_signal() so both
  
  senders and receivers are unlocked when work is done (be it reading or
  
  writing).
  This second solution replaces the condition with two : one to notify the
  
  senders, and one to notify the receivers. This prevents senders from
  
  notifying other senders instead of a reader, and the other way around.
  
  It also avoid broadcasting to everyone like the first solution, and is,
  
  as a result in theory more optimized.
  

  • [DH] libavutil/threadmessage.c

• ffmpeg doesn't accept input in script

  21 octobre 2022, par Eberhardt

  this is a beginner's question but i can't figure out the answer after looking into it for several days :

  


  I want ffmpeg to extract the audio portion of a video and save it in an .ogg container. If i run the following command in terminal it works as expected :

  


  ffmpeg -i example.webm -vn -acodec copy example.ogg


  


  For convenience, i want to do this in a script. However, if i pass a variable to ffmpeg it apparently just considers the first word and produces the error "No such file or directory".

  


  I noticed that my terminal escapes spaces by a \ so i included this in my script. This doesn't solve the problem though.

  


  Can someone please explain to me, why ffmpeg doesn't consider the whole variable that is passed to it in a script while working correctly when getting passed the same content in the terminal ?

  


  This is my script that passes the filename with spaces escaped by \ to ffmpeg :

  


  #!/bin/bash

titelschr=$(echo $@ | sed "s/ /\\\ /g")
titelohne=$(echo $titelschr | cut -d. -f 1)
titelogg=$(echo -e ${titelohne}.ogg)  

ffmpeg -i $titelschr -vn -acodec copy $titelogg


  


  Thank you very much in advance !

  


• checkasm : arm : preserve the stack alignment in x264_checkasm_checked_call

  14 novembre 2016, par Janne Grunau

  checkasm : arm : preserve the stack alignment in x264_checkasm_checked_call
  The stack used by x264_checkasm_checked_call_neon was a multiple of 4
  
  when the checked function is called. AAPCS requires a double word (8 byte)
  
  aligned stack public interfaces. Since both calls are public interfaces
  
  the stack is misaligned when the checked is called.
  This can cause issues if code called within this (which includes
  
  the C implementations) relies on the stack alignment.
  

  • [DH] tools/checkasm-arm.S