Aaron Culliney d4adabda93 SNEAKY BUGFIX race condition between UI thread and CPU thread on disk insertion
- Previously we would start the CPU thread and then insert, which has been a long standing race condition, but
      possibly better exposed since the recent disk.c rephactor to use mmap I/O
    - This directly manifested on a Kindle Fire 1st Gen as a crash, and, anecdotally on other devices as a "stalled"
      disk read requiring a reboot of the virtual machine.  Yay for crappy devices helping to expose crappy code! (git
      blame me). =P
2015-10-30 23:19:33 -07:00
..
2015-10-25 10:56:08 -07:00
2015-02-16 08:47:26 -08:00
2015-02-16 08:47:26 -08:00
2015-03-10 13:28:27 -07:00
res
2015-03-10 14:32:48 -07:00
2015-02-16 08:47:26 -08:00