2.6.5-rc1: compilation error on ppc32
Hi,
2.6.5-rc1 compilation fails at arch/ppc/syslib/indirect_pci.c lines 47 and 86: variable dev isn't defined.
The change breaking it is a change from out_be32() to PCI_CFG_OUT (line 17513 of the patch at
http://www.kernel.org/pub/linux/kernel/v2.6/testing/patch-2.6.5-rc1.bz2 , which strangely enough doesn't show in bkweb interface).
Signature · 3 lines
Colin This message represents the official view of the voices in my head.
Quoted 7 lines · Colin Leroy
Hi,
2.6.5-rc1 compilation fails at
arch/ppc/syslib/indirect_pci.clines 47 and 86: variable dev isn't defined.
The change breaking it is a change fromout_be32()toPCI_CFG_OUT(line 17513 of the patch at
http://www.kernel.org/pub/linux/kernel/v2.6/testing/patch-2.6.5-rc1.bz2 ,
The following is an uncleaned up (and slightly incomplete, there's changes to <asm-ppc/dma-mapping.h> needed) patch that gets things compiling for me. I'm compiling now for a Motorola LoPEC and I'll actually test the changes shortly. If it works, I'll split it up and send to akpm.
===== arch/ppc/defconfig 1.26 vs edited =====
arch/ppc/defconfig +1 -1
@@ -722,7 +722,7 @@ CONFIG_PMAC_APM_EMU=y CONFIG_PMAC_BACKLIGHT=y # CONFIG_MAC_FLOPPY is not set -CONFIG_MAC_SERIAL=m +CONFIG_MAC_SERIAL=n CONFIG_ADB=y CONFIG_ADB_MACIO=y CONFIG_INPUT_ADBHID=y
===== arch/ppc/configs/common_defconfig 1.26 vs edited =====
arch/ppc/configs/common_defconfig +1 -1
@@ -721,7 +721,7 @@ CONFIG_PMAC_APM_EMU=y CONFIG_PMAC_BACKLIGHT=y # CONFIG_MAC_FLOPPY is not set -CONFIG_MAC_SERIAL=m +CONFIG_MAC_SERIAL=n CONFIG_ADB=y CONFIG_ADB_MACIO=y CONFIG_INPUT_ADBHID=y
===== arch/ppc/syslib/indirect_pci.c 1.13 vs edited =====
arch/ppc/syslib/indirect_pci.c +4 -4
@@ -44,8 +44,8 @@ cfg_type =1 ;PCI_CFG_OUT (hose->cfg_addr, - (0x80000000 | ((dev->bus->number - hose->bus_offset) <<16 ) - | (dev->devfn <<8 ) | ((offset &0xfc ) | cfg_type))); + (0x80000000 | ((bus->number - hose->bus_offset) <<16 ) + | (devfn <<8 ) | ((offset &0xfc ) | cfg_type)));/* * Note: the caller has already checked that offset is
@@ -83,8 +83,8 @@ cfg_type =1 ;PCI_CFG_OUT (hose->cfg_addr, - (0x80000000 | ((dev->bus->number - hose->bus_offset) <<16 ) - | (dev->devfn <<8 ) | ((offset &0xfc ) | cfg_type))); + (0x80000000 | ((bus->number - hose->bus_offset) <<16 ) + | (devfn <<8 ) | ((offset &0xfc ) | cfg_type)));/* * Note: the caller has already checked that offset is
===== include/asm-ppc/pci.h 1.29 vs edited =====
include/asm-ppc/pci.h +6 -6
@@ -200,7 +200,7 @@ {BUG_ON (direction ==PCI_DMA_NONE ); -consistent_sync_for_cpu (bus_to_virt (dma_handle), size, direction); +consistent_sync (bus_to_virt (dma_handle), size, direction); } static inline void pci_dma_sync_single_for_device(struct pci_dev *hwdev,
@@ -209,7 +209,7 @@ {BUG_ON (direction ==PCI_DMA_NONE ); -consistent_sync_for_device (bus_to_virt (dma_handle), size, direction); +consistent_sync (bus_to_virt (dma_handle), size, direction); }/* Make physical memory consistent for a set of streaming
@@ -227,8 +227,8 @@BUG_ON (direction ==PCI_DMA_NONE );for (i =0 ; i < nelems; i++, sg++) -consistent_sync_page_for_cpu (sg->page, sg->offset, - sg->length, direction); +consistent_sync_page (sg->page, sg->offset, + sg->length, direction); } static inline void pci_dma_sync_sg_for_device(struct pci_dev *hwdev,
@@ -240,8 +240,8 @@BUG_ON (direction ==PCI_DMA_NONE );for (i =0 ; i < nelems; i++, sg++) -consistent_sync_page_for_device (sg->page, sg->offset, - sg->length, direction); +consistent_sync_page (sg->page, sg->offset, + sg->length, direction); }/* Return whether the given PCI device DMA address mask can
===== include/asm-ppc/unistd.h 1.32 vs edited =====
include/asm-ppc/unistd.h +4 -4
@@ -415,10 +415,10 @@int sys_pipe (int __user *fildes);int sys_ptrace (long request,long pid,long addr,long data);struct sigaction ; -asmlinkage long sys_rt_sigaction (int sig, -const struct sigaction __user *act, -struct sigaction __user *oact, -size_t sigsetsize); +long sys_rt_sigaction (int sig, +const struct sigaction __user *act, +struct sigaction __user *oact, +size_t sigsetsize);#endif /* __KERNEL_SYSCALLS__ */
Quoted 1 line
which strangely enough doesn't show in bkweb interface).
It does, you just have to go back quite a ways. We all hate bk patches now :)
Signature · 2 lines
Tom Rini http://gate.crashing.org/~trini/