Re: (XFree86) mmap of /proc/XXX/mem
Egbert Eich (eich@hilbert.ikp.physik.tu-darmstadt.de)
Thu, 2 Dec 1999 03:34:40 +0100 (MET)
Alexander Viro writes:
>
>
> On Wed, 1 Dec 1999, Egbert Eich wrote:
>
> > startup for each card that requests it. When a card wants to
> > make a BIOS call its real mode image is mapped to address 0
> > by mmapping /proc/self/mem. However I have reports mmapping
> > /proc/XXX/mem was removed from later versions of the 2.3 branch.
>
> In short: mmap on /proc/XXX/mem is bloody huge overkill. With
> ill-specified semantics. If all you need is an alias on private data that
> belongs to the same process - fine, but then it should be done in a
Yes, but how? We need a solution otherwise we won't be able to
use the vm86 mode even on Intel. The emulator makes everything
even slower.
> different way. If you want file-backed stuff - better yet, just mmap() it
> twice. But mmap() on /proc/<pid>/mem was simply broken. It gave a
mmap() what twice? I'm sorry I don't understand what you mean.
> way to screw swapper. Badly.
>
>
Regards,
Egbert.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/