Re: /proc/bus/pci odd behavior

Martin Mares (mj@ucw.cz)
Sun, 18 Oct 1998 12:38:41 +0200


Hello,

Alan wrote:

> Hardly /proc/ioports is a linked list of driver names and I/O reservations.

Aiee, I should not have slept while writing mail. I meant /dev/port, of course,
although /proc/kcore can also have side-effect on reads due to memory-mapped I/O.

> /dev/port is a better example. However being a device file find and tar and
> the like dont tend to drop you nasty suprises from it.

What about turning all side-effect-on-read files in /proc to character
devices?

Ingo wrote:

| maybe building a list of known broken card IDs and 'printing out something
| nasty' instead of 'locking up silently' prevents most of the problems?

I'd like to keep reading configuration registers on such broken cards
possible as I've already needed it several times.

| While it's not fundamentally different from using /proc/ioports, people
| feel uncomfortable because 'lspci -xx' is such a simple command and looks
| pretty harmless on the surface. (even if the man page warns about possible
| lockups). Maybe changing it to lspci --unsafe-query-all warns people.
| (also, it's just one typo away from 'lspci -x') [Admittedly, i use lspci
| -xx frequently and only recently learned about it's side effects. It looks
| too harmless IMHO.]

I'll change it to -xxx to prevent lockups due to typos and add the warning
to the help screen as well.

Have a nice fortnight

-- 
Martin `MJ' Mares   <mj@ucw.cz>   http://atrey.karlin.mff.cuni.cz/~mj/
Faculty of Math and Physics, Charles University, Prague, Czech Rep., Earth
"Anything is good and useful if it's made of chocolate."

- 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/