- 28 Jun, 2007 6 commits
-
-
Geert Uytterhoeven authored
Fix some PS3 build warnings reported by `make C=1'. You need to install sparse: git://git.kernel.org/pub/scm/devel/sparse/sparse.gitSigned-off-by:
Geert Uytterhoeven <Geert.Uytterhoeven@sonycom.com> Signed-off-by:
Geoff Levand <geoffrey.levand@am.sony.com> Signed-off-by:
Paul Mackerras <paulus@samba.org>
-
Geoff Levand authored
Use ioremap_flags() to map SPU regions as non-guarded. Change the use of _ioremap() to ioremap_flags(). CC: Arnd Bergmann <arnd.bergmann@de.ibm.com> CC: Masato Noguchi <Masato.Noguchi@jp.sony.com> CC: Takao Shinohara <shin@sm.sony.co.jp> Signed-off-by:
-
Masakazu Mokuno authored
Add a utility routine ps3_compare_firmware_version() to compare system firmware versions. Uses the existing ps3_get_firmware_version() routine. Signed-off-by:
Masakazu Mokuno <mokuno@sm.sony.co.jp> Signed-off-by:
-
Geoff Levand authored
Change the PS3 debug routines from using the GCC specific '__attribute__ ((unused))' to the preprocessor macro __maybe_unused. Signed-off-by:
-
Geoff Levand authored
Rename the PS3 static symbol virqs to ps3_ipi_virqs to aid in debugging. Signed-off-by:
-
Geoff Levand authored
Add a shutdown method to spu_sysdev_class to allow proper spu resource cleanup on system shutdown. This is needed to support kexec on the PS3 platform. Signed-off-by:
Arnd Bergmann <arnd.bergmann@de.ibm.com> Signed-off-by:
-
- 25 Jun, 2007 18 commits
-
-
Sachin P. Sant authored
On Power machines supporting VRMA, Kexec/Kdump does not work. VRMA (virtual real-mode area) means that accesses with IR/DR = 0 (i.e. the MMU "off") actually still go through the hash table, using entries put there by the hypervisor. This means that when we clear out the hash table on kexec, we need to make sure these entries are left untouched. This also adds plpar_pte_read_raw() on the lines of plpar_pte_remove_raw(). Signed-off-by : Sachin Sant <sachinp@in.ibm.com> Signed-off-by : Mohan Kumar M <mohan@in.ibm.com> Acked-by:
Benjamin Herrenschmidt <benh@kernel.crashing.org> Acked-by:
Olof Johansson <olof@lixom.net> Signed-off-by:
-
Mohan Kumar M authored
In some of the PPC970 based systems, interrupt would be distributed to offline cpus also even when booted with "maxcpus=1". So check whether cpu online map and cpu present map are equal or not. If they are equal default_distrib_server is used as interrupt server otherwise boot cpu (default_server) used as interrupt server. In addition to this, if an interrupt is assigned to a specific cpu (ie smp affinity) and if that cpu is not online, the earlier code used to return the default_distrib_server as interrupt server. This introduces an additional parameter to the get_irq function, called strict_check. Based on this parameter, if the cpu is not online either default_distrib_server or -1 is returned. Signed-off-by:
Mohan Kumar M <mohan@in.ibm.com> Cc: Michael Ellerman <michael@ellerman.id.au> Acked-by:
Milton Miller <miltonm@bga.com> Signed-off-by:
-
Michael Neuling authored
On pSeries the firmware features are not setup until ppc_md.init_early, so we can't do the firmware feature sections fixups till after this. Currently firmware feature sections is only used on iSeries which inits the firmware features much earlier. This is a bug in waiting on pSeries. Signed-off-by:
Michael Neuling <mikey@neuling.org> Signed-off-by:
-
Michael Neuling authored
For POWERPC, stolen time accounts for cycles lost to the hypervisor or PURR cycles attributed to the other SMT thread. Hence, when a PURR is available, we should still calculate stolen time, irrespective of being virtualised. Signed-off-by:
-
Mark A. Greer authored
Specifying 'console=ttyMM0' on the cmdline for the prmpc2800 is no longer necessary. Signed-off-by:
Mark A. Greer <mgreer@mvista.com> Signed-off-by:
-
Mark A. Greer authored
When a Marvell MPSC (serial controller) port is the specified /chosen/stdout-path device, call 'add_preferred_console()' so the user doesn't have to specify a 'console=ttyMMx' cmdline argument. Signed-off-by:
Dale Farnsworth <dale@farnsworth.org> Signed-off-by:
-
Michael Ellerman authored
Signed-off-by:
Michael Ellerman <michael@ellerman.id.au> Acked-by:
-
Michael Ellerman authored
This patch adds irq_create_direct_mapping(). This routine is an alternative to irq_create_mapping(), for irq controllers that can use linux virq numbers directly as hardware numbers. Signed-off-by:
-
Michael Ellerman authored
A future patch will need the logic at the end of irq_create_mapping() which setups a virq and installs it in the irq_map. So split it out into a new function irq_setup_virq(). Signed-off-by:
-
Olof Johansson authored
Uninline virq_to_hw and export it so modules can use it. The alternative would be to export the irq_map array instead, but it's an infrequently called function, and keeping the array unexported seems considerably cleaner. Signed-off-by:
-
Mark A. Greer authored
The holly support currently has separate rules to wrap its device tree with its zImage. This can now be done automatically without the extra rules so update holly support to use the automatic feature. Signed-off-by:
Mark A. Greer <mgreer@mista.com> Acked-by:
David Gibson <david@gibson.dropbear.id.au> Signed-off-by:
-
Mark A. Greer authored
There are 2 config options that indicate whether the platform being built has a device tree source file associated with it. Namely, CONFIG_WANT_DEVICE_TREE and CONFIG_DEVICE_TREE. When CONFIG_WANT_DEVICE_TREE is 'y' and CONFIG_DEVICE_TREE isn't an empty string, automatically wrap the specified device tree with the zImage being built. To achieve this, the 'dts' variable will only be set when the conditions above are true. The changes to the zImage.initrd.% and zImage.% rules cause the device tree to be wrapped when 'dts' is set; otherwise, they will work as they previosly did (i.e., build a zImage with no device tree). Signed-off-by:
-
Mark A. Greer authored
Being able to selectively wrap a device tree with the zIimage at build time has been deemed unnecessary, so this removes Makefile support for that feature. Signed-off-by:
-
Nathan Lynch authored
cpu_purr_data is a per-cpu array used to account for stolen time on partitioned systems. It used to be the case that cpus accessed each others' cpu_purr_data, so each entry was protected by a spinlock. However, the code was reworked ("Simplify stolen time calculation") with the result that each cpu accesses its own cpu_purr_data and not those of other cpus. This means we can get rid of the spinlock as long as we're careful to disable interrupts when accessing cpu_purr_data in process context. Signed-off-by:Nathan Lynch <ntl@pobox.com> Signed-off-by:
-
Guennadi Liakhovetski authored
With both generic rtc and powerpc timer suspend / resume code now in the (powerpc.git) tree, powerpc platforms using the generic timer and enabling power management will have timer.o linked in the kernel, which they don't need. Moreover, it will likely WARN_ON(!ppc_md.get_rtc_time), save zero-time and return no error on suspend... As a possible solution we can choose not to build timer.o when RTC_CLASS is enabled. However, I can imagine systems with 2 rtc's, one served by the ppc-rtc, another one generic built as a module, in which case using the ppc-rtc for suspend / resume will be impossible. Not to say, that such a configuration would be ugly... Signed-off-by:
G. Liakhovetski <g.liakhovetski@gmx.de> Signed-off-by:
-
Olof Johansson authored
Glue code to hook up the pata_platform on the PA Semi Electra eval board. CFE sets up device tree entries for the IDE interface, with device type 'ide' and compatible field 'electra-ide'. We unfortunately need to modify the resources before calling the generic platform driver, since the device tree only has one register window in it and the driver expects two. Adding this as an of_platform driver instead doesn't give us any benefit, it just adds one more layer of register/probe functions. Since CONFIG_PATA_PLATFORM depends on CONFIG_EMBEDDED, add that as a default for PPC_PASEMI. Signed-off-by:
-
Johannes Berg authored
This makes the timer sysdev use mktime instead of rtc_tm_to_time, since rtc_tm_to_time just calls mktime anyway, and this means we don't have a dependency on rtc-lib. Signed-off-by:
Johannes Berg <johannes@sipsolutions.net> Signed-off-by:
-
Milton Miller authored
With this, when kexec-ing, we copy the code and start the slaves on their journey to the next kernel's spin loop as soon as we copy the kexec image into place. The kernel doesn't know exactly which slaves are spinning in kexec_wait. This allows us to pass more than max-cpus to the next kernel. But it also means that we might leave some behind. Moving the code here means they have the time it takes us to clear the hash table to wake up and move on. Moving the code any earlier would reuqire walking the image description to search for the code, which could span multiple pages. Signed-off-by:
-
- 14 Jun, 2007 16 commits
-
-
David Gibson authored
This fixes some problems with the way the some things represented in the device tree for the Holly and Taiga boards. This means changes both to the dts files, and to the code which instantiates the tsi108 ethernet platform devices based on the device tree. - First, and most importantly, the ethernet PHYs are given with an identical 'reg' property. This reg currently encodes the accessible register used to initiate mdio interaction with the PHYs, rather than a meaningful address on the parent bus (mdio in this case), which is incorrect. Instead we give the address of these registers as 'reg' in the mdio node itself, and encode the ID of each phy in their 'reg' propertyies. - Currently the platform device constructor enables a workaround in the tsi108 ethernet driver based on the compatible property of the PHY. This is incorrect, because the workaround in question is necessary due to the board's wiring of the PHY, not the model of PHY itself. This patch alters the constructor to instead enable the workaround based on a new special property in the PHY node. - The compatible properties on a number of nodes in the device tree are insufficiently precise. In particular the PHYs give only "bcm54xx", which is broken, since there are many bcm54xx PHY models, and they have differences which matter. The mdio had a compatible property of "tsi-ethernet" identical to the ethernet MAC nodes, which doesn't make sense. The ethernet, i2c, bridge and PCI nodes were given only as "tsi-*" which is somewhat inprecise, we replace with "tsi108-*" in the case of Taiga (which has a TSI108 bridge), and "tsi109-*", "tsi108-*" in the case of Holly (which has a TSI109 bridge). - We remove some "model" properties from the ethernets on Taiga board which were neither useful nor adequately precise. - On Holly we change to using a dtc label instead of a full path to reference the MPIC node, which makes the dts a little more readable. Signed-off-by:
-
David Gibson authored
Currently, the Holly device tree includes a bootargs property in /chosen, which gives a commandline. This is somewhat inconvenient, because it means an alternative default command line can't be given in the kernel config - the value obtained from the dts via the bootwrapper will always override CONFIG_CMDLINE. This removes the command line from the dts, and instead puts the same command line as a default in holly_defconfig. Signed-off-by:
-
David Gibson authored
The various cuboot platforms (i.e. pre-device tree aware u-boot for 83xx, 85xx and Ebony) share a certain amount of code for parsing the boot parameters. To a certain extent that's inevitable, since they platforms have different definitions of the bd_t structure. However, with some macro work and a helper function, this patch improves the situation a bit. In the process, this fixes a bug on Ebony, which was incorrectly handling the parameters passed form u-boot for the command line (the bug was copied from 83xx and 85xx which have subsequently been fixed). Signed-off-by:
-
David Gibson authored
In the device tree for Ebony, the 'ranges' property in the node for the EBC bridge shows the mappings from the chip select / address lines actually used for the EBC peripherals into the address space of the OPB. At present, these mappings are hardcoded in ebony.dts for the mappings set up by the OpenBIOS firmware when it configures the EBC bridge. This replaces the hardcoded mappings with code in the zImage to read the EBC configuration registers and create an appropriate ranges property based on them. This should make the zImage and kernel more robust to changes in firmware configuration. In particular, some of the Ebony's DIP switches can change the effective address of the Flash and other peripherals in OPB space. With this patch, the kernel will be able to cope with at least some of the possible variations. Signed-off-by:
-
David Gibson authored
The ebony_exit() function which resets the Ebony board should in fact be common to most if not all 44x boards. This moves the function out into 44x.c, renaming it, so it can be used by other 44x platforms. Signed-off-by:
-
David Gibson authored
Currently the powerpc kernel has a 64-bit only feature, COHERENT_ICACHE used for those CPUS which maintain icache/dcache coherency in hardware (POWER5, essentially). It also has a feature, SPLIT_ID_CACHE, which is used on CPUs which have separate i and d-caches, which is to say everything except 601 and Freescale E200. In nearly all the places we check the SPLIT_ID_CACHE, what we actually care about is whether the i and d-caches are coherent (which they will be, trivially, if they're the same cache). This tries to clarify the situation a little. The COHERENT_ICACHE feature becomes availble on 32-bit and is set for all CPUs where i and d-cache are effectively coherent, whether this is due to special logic (POWER5) or because they're unified. We check this, instead of SPLIT_ID_CACHE nearly everywhere. The SPLIT_ID_CACHE feature itself is replaced by a UNIFIED_ID_CACHE feature with reversed sense, set only on 601 and Freescale E200. In the two places (one Freescale BookE specific) where we really care whether it's a unified cache, not whether they're coherent, we check this feature. The CPUs with unified cache are so few, we could consider replacing this feature bit with explicit checks against the PVR. This will make unifying the 32-bit and 64-bit cache flush code a little more straightforward. Signed-off-by:
David Gibson <dwg@au1.ibm.com> Signed-off-by:
-
David Gibson authored
Using typedefs to rename structure types if frowned on by CodingStyle. However, we do so for the hash PTE structure on both ppc32 (where it's called "PTE") and ppc64 (where it's called "hpte_t"). On ppc32 we also have such a typedef for the BATs ("BAT"). This removes this unhelpful use of typedefs, in the process bringing ppc32 and ppc64 closer together, by using the name "struct hash_pte" in both cases. Signed-off-by: -
David Gibson authored
This factors some things defined in both pgtable-ppc32.h and pgtable-ppc64.h into the common part of asm-powerpc/pgtable.h. These are all things which have essentially identical definitions, and which by their nature are very unlikely ever to need different definitions in the two cases. Signed-off-by:
-
David Gibson authored
In arch/powerpc/mm/pgtable_32.c, the variable io_bat_index and the macro is_power_of_4() no longer have any users. This removes them. Signed-off-by:
-
David Gibson authored
APUS (the Amiga Power-Up System) is not supported under arch/powerpc and it's unlikely it ever will be. Therefore, this patch removes the fragments of APUS support code from arch/powerpc which have been copied from arch/ppc. A few APUS references are left in asm-powerpc in .h files which are still used from arch/ppc. Signed-off-by:
-
David Gibson authored
These old-fashioned IO mapping functions no longer have any callers in code which remains relevant on arch/powerpc. Therefore, this removes them from arch/powerpc. Signed-off-by:
-
David Gibson authored
arch/powerpc still relies on asm-ppc/mmu.h for most 32-bit MMU types. This is another step towards fixing this. It takes the portions of asm-ppc/mmu.h related to the "classic" 32-bit hash page table MMU which are still relevant in arch/powerpc and puts them in a new asm-powerpc/mmu-hash32.h, included when appropriate from asm-powerpc/mmu.h. Signed-off-by:
-
David Gibson authored
Currently, all OF-related code in the bootloader is contained in of.c. of.c also provides the platform specific things necessary to boot on an OF platform. However, there are platforms (such as PReP) which can include an OF implementation, but are not bootable as pure OF systems. For use by such platforms, this patch splits out the low-level parts of the OF code (call_prom() and various wrappers thereof) into a new oflib.c file. In addition, the code related to bootwrapper console output via OF are moved to a new ofconsole.c file. Both these files are included in the wrapper.a library where they can be used by both full-OF and partial OF platforms. Signed-off-by:
-
Arnd Bergmann authored
A lot of the options in arch/powerpc/Kconfig deal with the CPU menu, and my next patches add more to them. Moving them to a new arch/powerpc/platforms/Kconfig.cputype file makes it easier to follow. There are no functional changes in here. Signed-off-by:
Arnd Bergmann <arnd@arndb.de> Signed-off-by:
-
will schmidt authored
We have had complaints where a threaded application is left in a bad state after one of it's threads is killed when we hit a VM: out_of_memory condition. Killing just one of the process threads can leave the application in a bad state, whereas killing the entire process group would allow for the application to restart, or be otherwise handled, and makes it very obvious that something has gone wrong. This change allows the entire process group to be taken down, rather than just the one thread. lightly tested on powerpc Signed-off-by:
Will <will_schmidt@vnet.ibm.com> Signed-off-by:
-
Jake Moilanen authored
A Power6 can give up CPU cycles on a dedicated CPU (as opposed to a shared CPU) to other shared processors if the administrator asks for it (via the HMC). This enables that to work properly on P6. This just involves setting a bit in the CAS structure as well as the VPA. To donate cycles, a CPU has to have all SMT threads idle and have the donate bit set in the VPA. Then call H_CEDE. The reason why shared processors just aren't used is because dedicated CPUs are guaranteed an actual processor, yet the system is still able to increase the capacity of the shared CPU pool. Also rename the VPA's cpuctls_task_attrs field to a more accurate name. Signed-off-by:
Jake Moilanen <moilanen@austin.ibm.com> Signed-off-by:
-
