On 11 Jun 99, at 9:13, Christophe Kumsta wrote:
> Hi all,
>
> just a word to tell that the last gcc-egcs 2.95 seems to be buggy
> with the io.
>
> I have a driver for an owner video acquisition PCI card and this one
> worked well. Last day I've installed the last
> [gcc version gcc-2.95 19990524 (prerelease)]
> and then my driver did no more works !
>
> After a lot of debug , I realized that the function 'ioremap' returned
> me a valid virt. address but all reading on this remapped io did
> 0xFFFFFFFF !!
>
> Now I've reinstalled egcs [gcc version egcs-2.90.29 980515 (egcs-1.0.3
> release)]
> and the same recompilation make the driver working !!!!! with the same
> source !!
>
> Also , if you plan to use IO , don't use gcc-2.95 now ( next release
> perhaps ... )
At first gcc-2.95 19990524 (prerelease) is prerelease as it is clearly
stated and may contain bugs (perhaps less than experimental
versions but still so). I myself would prefer stable version for kernel
related things (imagine some problem with development snapshot
make IDE drawer "crazy"). I'm using egcs-1.1.2 for most work in
Linux (including compiling kernel), but I also am experimenting with
current snapshots from gcc_2.95_branch for my own applications
(problem here is less likely to do much harm there)
So I don't think we should give recommendations about compiler
versions which are not yet officially released
If You are ready to take a risk You can try out later snapshots
(currently 19990608 is latest) and study the problem to prepare
FULL bug report to be sent to egcs-bugs mailing list.
Andris
--- [rtl] ---
To unsubscribe:
echo "unsubscribe rtl" | mail [EMAIL PROTECTED] OR
echo "unsubscribe rtl <Your_email>" | mail [EMAIL PROTECTED]
----
For more information on Real-Time Linux see:
http://www.rtlinux.org/~rtlinux/
