- 29 Oct, 2005 37 commits
-
-
Paul Mackerras authored
Somehow we ended up with an #ifdef CONFIG_PPC64 around the export of cur_cpu_spec, but raid6 as a module needs it on ppc32 as well as ppc64. Signed-off-by:Paul Mackerras <paulus@samba.org>
-
Paul Mackerras authored
Signed-off-by: -
Paul Mackerras authored
Signed-off-by: -
Andy Whitcroft authored
Currently when we first select memory model (FLAT, DISCONTIG, SPARSE) then select whether the machine is NUMA. However NUMA systems may not be FLAT. This constraint it not honoured and we may configure a NUMA/FLAT system. Reorder the configuration such that we choose NUMA first which allows us to only list the memory models which are valid. We now default NUMA for known NUMA systems. Note that this new order also matches that used in x86. Signed-off-by:
Andy Whitcroft <apw@shadowen.org> Signed-off-by:
Joel Schopp <jschopp@austin.ibm.com> Signed-off-by:
Andrew Morton <akpm@osdl.org> Signed-off-by:
-
Olaf Hering authored
remove duplicate local variable, saves 2 asm instructions. Signed-off-by:
Olaf Hering <olh@suse.de> Signed-off-by:
-
Olaf Hering authored
'make install' creates a /boot/zImage[.vmode] file when the defconfig is used. It uses the second arg as file content, which is the vmlinux, and the 5th arg as file name, which is the BOOTIMAGE name. A comment in an earlier patch to install.sh states that yaboot can not load a zImage+initrd combo. This was true in kernel 2.6.5 because it did use bi_recs to pass the initrd info. But this concept was always broken. Register r3 holds the initrd address and r4 holds the initrd size. This works with all kernel versions. The current code in main.c leaves r3 and r4 alone, so the kernel should be able to see and use the memory range with the initrd content. If one wants to rerun mkinitrd, it is currently hard to get the uname -r value for the installed zImage. Without this info, mkinitrd can not know what modules to use. This would be fixable by including the /proc/version output of the new kernel. But it is simpler to just use the plain vmlinux. So all this patch does is to write to /boot/vmlinux instead to /boot/zImage Signed-off-by:
-
Olaf Hering authored
'make ARCH=ppc64 O=../O install' does not work with the defconfig. CONFIG_PPC_BPA is part of it, but the BPA bootimage variable is wrong: make[2]: *** No rule to make target `zImage', needed by `install'. Stop. Signed-off-by:
-
Olaf Hering authored
compile nls_cp437 and nls_iso8859_1 into the kernel in defconfig. This is already enabled in pSeries_defconfig. Reason: if one just boots the new shiny zImage and the root filesystem is on a filesystem not readable by yaboot (like jfs, raid or lvm) upgrading the bootloader will fail because the FAT bootpartition can not be mounted. Signed-off-by:
-
Olaf Hering authored
Currently, AC Power is 0 on a desktop G4. No batteries present should mean AC Power == 1. Signed-off-by:
-
Olaf Hering authored
Show firmware provided stackpointer during boot. This helps to find the "taboo" areas on the various boards. claim tends to fail for these memory areas, but some jokers return success anyway. Use %p to print the load address, its a pointer. Signed-off-by:
-
Olaf Hering authored
To prove that the relocation works, move the crt0.o away from the beginning. Move linker options from command line into linker script. rename entry point because '_start' is referenced in printf output. Signed-off-by:
-
Olaf Hering authored
Make the zImage relocateable. So yaboot could just load and run any ELF binary, without worrying about its load address. Signed-off-by:
-
Olaf Hering authored
No need to initialize global variables. Signed-off-by:
-
Olaf Hering authored
Update comment in memcpy, r7 contains the byte count. Signed-off-by:
-
Olaf Hering authored
Use memset to clear bss, instead of own version. Signed-off-by:
-
Olaf Hering authored
Remove userland related stuff from ld.script, they are not required for zImage use wildcards for some sections. Signed-off-by:
-
Olaf Hering authored
bootfiles must be relinked if linker script changes Signed-off-by:
-
Olaf Hering authored
Move the gunzip function up. Signed-off-by:
-
Olaf Hering authored
Compute the vmlinux size at runtime. Use Z_FULL_FLUSH instead of Z_FINISH, to extract only the ELF header and ELF program header. ->p_memsz is the required memory range for the executable, including bss ->p_filesz is the size of .text, .data and other runtime sections These values must be used for the claim call. All additional memory needed by the kernel is claimed in prom_init, remove the extra Mb. Pass the full memsize as target area to gunzip, otherwise not everything will be uncompressed. flush_cache has to flush all runtime sections, do not reduce the memrange by the ->p_offset value because its just that: an offset. Remove the Makefile code to produce an imagesize.c, its not needed anymore. Remove all FORCE flags, to not rebuild the zImage if vmlinux was not changed. Signed-off-by:
-
Olaf Hering authored
Switch ppc64 to the in-kernel zlib, it has less bugs than the current one. The code in arch/ppc64/boot is compiled as 32bit, so it can not use the includes from include/asm. Copy all zlib related header files and convert them with sed. Reduce the scratch size to 47k, check possible changes at runtime. Signed-off-by:
-
Olaf Hering authored
string.h needs definition of size_t, but not the one from linux/include Signed-off-by:
-
Olaf Hering authored
zutil.h does not need errno.h Signed-off-by:
-
Olaf Hering authored
There is no need to include module.h in inflate.c Signed-off-by:
-
Nicolas DET authored
Added information message when using the SRAM in mv643xx_eth_probe() Signed-off-by:
Nicolas DET <det.nicolas@free.fr> Signed-off-by:
Sven Luther <sl@bplan-gmbh.de> Signed-off-by:
-
Vitaly Bordug authored
This patch fixes a numbers of issues regarding to that both 8xx and 82xx began to use ppc_sys model: - Platform is now identified by default deviceless SOC, if no BOARD_CHIP_NAME is specified in the bard-specific header. For the list of supported names refer to (arch/ppc/syslib/) mpc8xx_sys.c and mpc82xx_sys.c for 8xx and 82xx respectively. - Fixed a bug in identification by name - if the name was not found, it returned -1 instead of default deviceless ppc_spec. - fixed devices amount in the 8xx platform system descriptions Signed-off-by:
Vitaly Bordug <vbordug@ru.mvista.com> Signed-off-by:
Marcelo Tosatti <marcelo.tosatti@cyclades.com> Signed-off-by:
Kumar Gala <kumar.gala@freescale.com> Signed-off-by:
-
Andy Fleming authored
This patch updates the 85xx platform code to support the new PHY Layer. Signed-off-by:
Andy Fleming <afleming@freescale.com> Signed-off-by:
-
Nicolas DET authored
Add proper entry to support the Marvell MV64361 (Marvell Discovery II) SRAM. This feature may be used by the mv643xx_eth driver. Signed-off-by:
-
Olaf Hering authored
mesh, mac53c94 and airport already have an entry. Add the network drivers for pmac. Signed-off-by:
Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by:
-
scwhab@suse.de authored
Provide a "compatible" entry in /sys/bus/macio/devices/*/ This can be used to load drivers via the modules.alias file. Signed-off-by:
-
Olaf Hering authored
This implements a nvram acccess method, similar to arch/ppc64/kernel/pSeries_nvram.c tested on CHRP B50. Signed-off-by:
-
Roland Dreier authored
Change the phys_mem_access_prot() function to take a pfn instead of an address. This allows mmap64() to work on /dev/mem for addresses above 4G on 32-bit architectures. We start with a pfn in mmap_mem(), so there's no need to convert to an address; in fact, it's actively bad, since the conversion can overflow when the address is above 4G. Similarly fix the ppc32 page_is_ram() function to avoid a conversion to an address by directly comparing to max_pfn. Working with max_pfn instead of high_memory fixes page_is_ram() to give the right answer for highmem pages. Signed-off-by:
Roland Dreier <rolandd@cisco.com> Cc: Anton Blanchard <anton@samba.org> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by:
-
Olaf Hering authored
Mention a few more commands in xmon. System.map processing was replaced with kallsyms. Signed-off-by:
-
Marcelo Tosatti authored
#ifdef out an ALTIVEC specific tweak in __switch_to() Signed-off-by:
-
Matt Porter authored
Cleanup PPC440 eval boards (bamboo, ebony, luan and ocotea) to better support U-Boot as bootloader. Signed-off-by:
Stefan Roese <sr@denx.de> Signed-off-by:
Matt Porter <mporter@kernel.crashing.org> Signed-off-by:
-
Marcelo Tosatti authored
Here is an uptodated version of the MPC8xx PCMCIA driver for v2.6, addressing comments by Jeff and Dominik: - use IO accessors instead of direct device memory referencing - avoid usage of non-standard "uint/uchar" data types - kill struct typedef's Will submit it for inclusion once v2.6.14 is out. Testing on 8xx platforms is more than welcome! Works like a charm on our custom hardware (CONFIG_PRxK). Cc: Dominik Brodowski <linux@dominikbrodowski.net> Signed-off-by:
-
Marcelo Tosatti authored
Convert core 8xx drivers to use in_xxxbe/in_xxx macros instead of direct memory references. Other than making IO accesses explicit (which is a plus for readability), a common set of macros provides a unified place for the volatile flag to constraint compiler code reordering. There are several unlucky places at the moment which lack the volatile flag. Signed-off-by:
-
Lee Nicks authored
Depending on how GCC is built, GCC 4 may generate altivec instructions without user explicitly requesting vector operations in the code. Although this is a performance booster for user applications, it is a problem for kernel. This patch explicitly instruct GCC to NOT generate altivec instructions while building the kernel. Here are some test cases I ran. (1) build gcc 4.0.1 with '--with-cpu=7450 --enable-altivec --enable-cxx-flags=-mcpu=7450', and use this gcc to build kernel WITHOUT this kernel patch. Kernel fail to boot up on a 7450 board because of altivec instructions in kernel. (2) build gcc 4.0.1 with "--with-cpu=7450 --enable-altivec --enable-cxx-flags=-mcpu=7450", and use this gcc to build kernel WITH this kernel patch. Kernel boot up on a 7450 board without any problem. (3) build gcc 4.0.1 with "--with-cpu=750 --enable-cxx-flags=-mcpu=750", and use this gcc to build kernel with or without this kernel patch. Kernel boot up on a 7450 board without any problem. This patch should also work with GCC 3 or even earlier GCC 2.95.3. Signed-off-by:Lee Nicks <allinux@gmail.com> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by:
-
- 28 Oct, 2005 3 commits
-
-
Paul Mackerras authored
The merged version follows the ppc64 version pretty closely mostly, and in fact ARCH=ppc64 now uses the arch/powerpc/xmon version. The main difference for ppc64 is that the 'p' command to call show_state (which was always pretty dodgy) has been replaced by the ppc32 'p' command, which calls a given procedure (so in fact the old 'p' command behaviour can be achieved with 'p $show_state'). Signed-off-by: -
Paul Mackerras authored
The sc instruction emulation can't be done the same way on 32-bit as 64-bit yet, but this should work OK. Signed-off-by: -
Paul Mackerras authored
... for consistency with 64-bit. Signed-off-by:
-
