From mboxrd@z Thu Jan 1 00:00:00 1970 From: arnd@arndb.de (Arnd Bergmann) Date: Sat, 16 Oct 2010 23:54:20 +0200 Subject: [PATCH 0/7] arm/versatile PCI support, second try In-Reply-To: References: <1287072657-32477-1-git-send-email-arnd@arndb.de> Message-ID: <201010162354.20418.arnd@arndb.de> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Saturday 16 October 2010, Linus Walleij wrote: > 2010/10/14 Arnd Bergmann : > > > With lots of help and testing from Peter Maydell, I've completed > > a second version of the PCI support for the versatile platform, > > which works much better now on real hardware, although it does > > not improve existing qemu setups. > > I've got nothing better to do tonite than playing with the PB1176 > board so tried this out... > > Some fixup is needed in 6/7, the EB patch because I get this: > arch/arm/mach-realview/realview_eb.c:94: error: > 'REALVIEW_PCI_CORE_BASE' undeclared here (not in a function) > arch/arm/mach-realview/realview_eb.c:110: error: > 'REALVIEW_PCI_IO_BASE' undeclared here (not in a function) > arch/arm/mach-realview/realview_eb.c:111: error: > 'REALVIEW_PCI_IO_BASE_SIZE' undeclared here (not in a function) > make[3]: *** [arch/arm/mach-realview/realview_eb.o] Error 1 > make[2]: *** [arch/arm/mach-realview] Error 2 > make[2]: *** Waiting for unfinished jobs.... > > No these are indeed named > REALVIEW_EB_PCI_CORE_BASE > REALVIEW_EB_PCI_IO_BASE0 > etc... Ah right, I removed these when I prepared the patches for submission, initially I was using the common definitions for the realview boards but I figured it would be bad since newer realviews have a different controller. Apparently I forgot to move over the EB code to use the new definitions I added later. I'll fix these up. > If I disable the EB support it compiles and boots. Nothing much > else happens, not even if I enable PCI debugging. But I guess > that's normal since I don't have any PCI enclosure expansion > board for this card? Yes, without a PCI card to plug into it it's pretty pointless. > The only visible change is that /sys/bus/pci now exists with no > slots nor devices in it. > > But it's not causing any regressions for me so with a fix > for the compile problem it's > Tested-by: Linus Walleij Thanks, Arnd