Skip to contentSkip to conversation

b44: regression in 2.6.22

DISCUSSION Uwe Bugla 1 participant 1 message May 28, 2007

lore
Share & tools
lore

Uwe Bugla

author
#0 lore raw
Quoted 19 lines · Michael Buesch

Can you also test the following patch?
I think there's a bug in b44 that is doesn't properly discard shared IRQs, so it might possibly generate a NAPI storm, dunno. Worth a try.

Index: linux-2.6.22-rc3/drivers/net/b44.c
===================================================================
--- linux-2.6.22-rc3.orig/drivers/net/b44.c 2007-05-27 23:01:44.000000000 +0200 +++ linux-2.6.22-rc3/drivers/net/b44.c 2007-05-28 12:48:27.000000000 +0200 @@ -911,6 +911,8 @@ static irqreturn_t b44_interrupt(int irq spin_lock(&bp->lock);

istat = br32(bp, B44_ISTAT);
+ if (istat == 0xFFFFFFFF)
+ goto out; /* Shared IRQ not for us */
imask = br32(bp, B44_IMASK);

/* The interrupt mask register controls which interrupt bits

Review irq_ack: · line 950

Quoted 7 lines
(quoted diff) +1 -0
@@ -942,6 +944,7 @@ irq_ack:
 bw32(bp, B44_ISTAT, istat);
 br32(bp, B44_ISTAT);
 }
+out:
 spin_unlock(&bp->lock);
 return IRQ_RETVAL(handled);
 }

I did try this patch on a affected kernel, but I didn't notice any big difference. Perhaps the kernel is a bit less slow during the test, but It's hard to tell.

Maxi

I tried that one also, but it did not help to solve the real problem.

The real problem is the b44 -----> b44-ssb-port in 2.6.22-rc2-mm1.
This port was not done correctly, as my bcm4401 runs fine with all 2.6 kernels except 2.6.22-rc2-mm1.

I also tried to load the two modules that way:

modprobe ssb debug=4
modprobe b44 debug=4

Unfortunately those two modules do not understand those command line parameters.

I even reverted significant parts of module b44.h (i. e. the backplane definitions). But that did not help either.

In so far the changes for b44.c in kernel 2.6.22-rc2-mm1 are definitely responsible for the fact that my bcm4401 controller refuses to work.

Cheers

Uwe

P. S.: I am basically not against those ports like the above mentioned one. But if they are done they at least should be done more severe and correct.

As this port is a deep cut into the architecture, it would be very helpful to make the two new modules understand debug command line parameters.
Just to somplify troubleshooting.

You cannot rip out code of one module, attempt to do a port, and afterwards jump around crying:
"For me it works - what can I do?"

You moreover cannot chase bug reporters to test 3 different developper kernels containing the same code expecting that there will be a real result to help resolve the real issue.

You moreover cannot accuse bug reporters for typo errors in DNS server, for the ACPI system to be buggy etc.

This is no professional bug handling, but it is just wasting time and nerves due to lack of concept how to resolve the real issue (i. BAD b44 ------> b44-ssb port).

Hint: Please everyone reading this see the following threads for positive examples how bug reports are to be resolved:

A. the sis900.c Oops (before kernel 2.6.21.2)
B. The "alsa architecture broken in parts" issue

Those are only 2 examples on how to resolve bug report issues in a real sovereign and sophisticated manner.

_______________________________________________________________
SMS schreiben mit WEB.DE FreeMail - einfach, schnell und kostenguenstig. Jetzt gleich testen! http://f.web.de/?mc=021192