>On Sun, 15 Apr 2012 21:39:54 +0100 >Ken Moffat <zarniwh...@ntlworld.com> wrote: > > On Sun, Apr 15, 2012 at 03:27:23PM -0400, Michael Shell wrote: > > On Sun, 15 Apr 2012 17:50:25 +0100 > > Ken Moffat <zarniwh...@ntlworld.com> wrote: > > > > > i.e. it (slowly: the file is on a local disk) fills the > > > buffer, plays the buffer, then waits while it refills the > > > buffer. > > > > I doubt this is the problem, especially on a new machine, > > but in the past issues like this were a symptom that DMA > > mode was not enabled for the drive(s) in question. Are > > you sure the kernel is enabling DMA I/O for the DVD and > > the HD? > > > The disk is SATA, so using libata. All the googling I did in the > past suggests DMA is always on with libata for SATA devices (but see > below). > > BTW, is there a way to check for a drive's DMA mode in > > /proc on the newer (/dev/sda using) kernels? (In the old > > days it was in /proc/ide, but all that is routed through > > the scsi i/o system today.) > > I haven't found one. 'hdparm -t' reports > 112 MB/sec for the > disk, so I think DMA is definitely in use. However, it only reports > between 2 and 3 MB/sec for the DVD drive on /dev/sr0. BUT, that > plays fine (with minimal cpu load) if I use vlc without the DVD > menus, and similarly plays fine in xine. The wait for the buffer to > fill was on the hard disk. > > ĸen
Sounds to me like a problem with VLC. Perhaps you should focus on it. Maybe a different version, some changes in configure switches... or maybe some of the underlying libraries? It has been a long time since I last compiled or used VLC so I don't remember what things it links to. If it uses one of the libdvd* packages, maybe the problem is in that direction, especially if Xine does that part differently (Xine is also ancient history with me). -- Fourth law of programming: Anything that can go wrong wi sendmail: segmentation violation - core dumped -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page