>>>>> "Paul" == Paul Gortmaker <p_gortmaker@yahoo.com> writes:
Paul> Uwe Bonnes wrote:
>> appended patch registers the IO Memory region used by the smc-ultra
>> driver. It is against 2.4.2, but applies to 2.2.18 too.
>>
>> Could anybody have a look at the patch and either forward it to Linux
>> and Alan or give me the "go" to do that?
Paul> It leaks memory (dev->priv) and mixes exit styles with what is in
Paul> use in that driver (i.e. return -EBUSY vs. retval=-EBUSY; goto
Paul> out;) and the mem region should be requested conditionally on
Paul> whether the card is in PIO or sh-mem mode.
Paul> I originally had a patch to fix three things that are rather
Paul> related. They are 1) adding request_mem_region, 2) ioremap
Paul> instead of isa_r/w and 3) purge of isa-isms like dev->rmem_* from
Paul> the netdev struct. It didn't go into 2.4 and I didn't push it,
Paul> since it wasn't a showstopper by any means (and my patch probably
Paul> had bugs too ;)
Paul> If people really want another shiny number in /proc to gawk at
Paul> (pun intended) then I'll post a 2.4.x patch for the common ISA
Paul> sh-mem 8390 cards that just adds the request_mem_region stuff.
Hallo Paul,
thanks for your answer. However I don't request another "shiny' number in
/proc, but I want to avoid problems with my PCMCIA-ISA card with the
SMC. This was a showstopper for my, as net connectivity broke when I
inserted a card.
So the patch was only a quick hack, but I'd appriciate if you
try to force your patches in the kernel and want we to test things.
Bye
--
Uwe Bonnes bon@elektron.ikp.physik.tu-darmstadt.de
Institut fuer Kernphysik Schlossgartenstrasse 9 64289 Darmstadt
--------- Tel. 06151 162516 -------- Fax. 06151 164321 ----------
-
: send the line "unsubscribe linux-net" in
the body of a message to majordomo@vger.kernel.org
