Messages in this thread | | | Date | Fri, 16 Mar 2001 14:09:38 -0500 | From | David Mansfield <> | Subject | Re: [PATCH] fix a bug in ioctl(CDROMREADAUDIO) in cdrom.c in 2.2 |
| |
Pierre Etchemaite wrote: > > Le 14-Mar-2001, Jani Jaakkola écrivait : > > > > Using ioctl(CDROMREADAUDIO) with nframes argument being larger than 8 and > > not divisible by 8 causes kernel to read and return more audio data than > > was requested. This is bad since it clobbers up processes memory > > (I noticed this when my patched cdparanoia segfaulted). > > Same thing for 2.4.2. > > Is my allocation loop "over engineering", or just plain bad thing to do ? >
I've been running this (or close: my version tries 8 frames, then jumps immediately to 1, without trying 4 and 2 in between if the kmalloc fails) since it was changed. Without such a patch, my CDDA read speed drops to 25% the original rate. You also have the fix that started the thread!
Jens (cdrom maintainer) said he was working on a more elegant solution, but to me, such a simple fix as yours should go in the kernel in the meantime. Jens?
-- David Mansfield (718) 963-2020 david@ultramaster.com Ultramaster Group, LLC www.ultramaster.com - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |