Re: mmap bug in 2.2.13 (or in documentation)
Richard Zidlicky (Richard.Zidlicky@stud.informatik.uni-erlangen.de)
Thu, 30 Dec 1999 21:49:14 +0100
On Wed, Dec 29, 1999 at 04:15:03PM +0100, Rik van Riel wrote:
> On Wed, 29 Dec 1999, Jamie Lokier wrote:
> > Abramo Bagnara wrote:
> > > As I'm unable to understand why a file opened WR_ONLY cannot be mmap'ed
> > > PROT_WRITE, MAP_SHARED I deduce that a break is missing at line 222.
> >
> > On some platforms (i386), PROT_WRITE is equivalent to PROT_READ |
> > PROT_WRITE. The processor provides no mechanism for write-only access.
>
> And on others you can only address memory in word sizes,
> whereas some languages (C, for example) mandate access
> on the byte or even bit level.
>
> This means that you have to read a word first before you
> can write a single byte, making the whole write-only
> thing a pointless excercise.
I would not call it pointless, despite these limitations. I was doing
something similar in an emulator to emulate HW registers. Turned out
to be awfully complicated but I haven't yet given up ;-)
Bye
Richard
-
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/