Some clarifications about the Linux throughput problem
This commit is contained in:
parent
07f535c72b
commit
18b72f1602
@ -1 +1 @@
|
|||||||
#define Cdrskin_timestamP "2009.12.06.073404"
|
#define Cdrskin_timestamP "2009.12.06.074622"
|
||||||
|
@ -1285,6 +1285,12 @@ int burn_disc_close_track_dvd_plus_r(struct burn_write_opts *o,
|
|||||||
MAP_SHARED | MAP_ANONYMOUS, -1, (off_t) 0);
|
MAP_SHARED | MAP_ANONYMOUS, -1, (off_t) 0);
|
||||||
is supposed to allocate a properly aligned buffer.
|
is supposed to allocate a properly aligned buffer.
|
||||||
64 KB is supposed to be a safe size.
|
64 KB is supposed to be a safe size.
|
||||||
|
Actually mmap() seems to be the main cause for a positive effect
|
||||||
|
of O_DIRECT.
|
||||||
|
|
||||||
|
This simplified transmission function did not bring visible benefit.
|
||||||
|
So for now it is not worth to teach it all applicable details of old
|
||||||
|
CD sector oriented transmission.
|
||||||
|
|
||||||
@return 1= ok, go on , 2= no input with track->open_ended = nothing written
|
@return 1= ok, go on , 2= no input with track->open_ended = nothing written
|
||||||
<= 0 = error
|
<= 0 = error
|
||||||
|
Loading…
Reference in New Issue
Block a user