From mboxrd@z Thu Jan 1 00:00:00 1970 From: Michael Buesch Subject: Re: b44: high ping times with wireless-dev Date: Sun, 17 Jun 2007 14:12:44 +0200 Message-ID: <200706171412.45074.mb@bu3sch.de> References: <200706162327.47413.maxi@daemonizer.de> <200706171255.40155.mb@bu3sch.de> <200706171403.34299.maxi@daemonizer.de> Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-15" Content-Transfer-Encoding: 7bit Cc: netdev , "linux-kernel" , "linux-wireless" , Gary Zambrano , Stephen Hemminger , "Jeff Garzik" , "Arnaldo Carvalho de Melo" To: Maximilian Engelhardt Return-path: In-Reply-To: <200706171403.34299.maxi-OwNUvPV92VfddJNmlsFzeA@public.gmane.org> Content-Disposition: inline Sender: linux-wireless-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org List-Id: netdev.vger.kernel.org On Sunday 17 June 2007 14:03:30 Maximilian Engelhardt wrote: > On Sunday 17 June 2007, Michael Buesch wrote: > > On Saturday 16 June 2007 23:27:43 Maximilian Engelhardt wrote: > > > [...] > > > ACPI: PCI Interrupt Link [LNKD] enabled at IRQ 10 > > > ACPI: PCI Interrupt 0000:02:02.0[A] -> Link [LNKD] -> GSI 10 (level, low) > > > -> IRQ 10 > > > ssb: Sonics Silicon Backplane found on PCI device 0000:02:02.0 > > > b44.c:v2.0 > > > eth0: Broadcom 44xx/47xx 10/100BaseT Ethernet 00:c0:9f:29:99:a7 > > > [...] > > > > Ok, I prepared two debugging patches. > > > > Please enable SonicsSiliconBackplane Debugging in the kernel kconfig, > > so I can get more detail information about your card. > > Device Drivers/Sonics Silicon Backplane/SSB debugging > > (Must disable "No SSB kernel messages") > > > > Please apply and test the attached debugging patches in a row. > > So apply patch 1 and test if it works again. If not, apply > > patch 2 and test if it works. > > Always save complete dmesg log on each test run and send it to me. > > > > Thanks for testing. > > (This time it seems we are actually getting somewhere, when > > dealing with sane people. :D ) > > I did the tests with my kernel where only the card is on interrupt 10. dmesg > is attached. > With the first patch applied networking does work again. I also additionally > tried patch2 and it also does work. Great! To me this seems to be a silicon bug. The IRQ routing value, which is read from the chip here, is hardcoded in the old b44 driver. I'll implement a workaround for this and submit a patch. -- Greetings Michael.