- 20 Dec, 2007 30 commits
-
-
Benjamin Herrenschmidt authored
This adds flags the platforms can use to enable domain numbers in /proc/bus/pci. Signed-off-by:
Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by:
Paul Mackerras <paulus@samba.org>
-
Benjamin Herrenschmidt authored
The 32 bits PCI code carries an old hack that was only useful for G5 machines. Nowdays, the 32 bits kernel doesn't support any of those machines anymore so the hack is basically never used, so remove it. Signed-off-by:
-
Benjamin Herrenschmidt authored
This adds to the 32 bits PCI code some flags, replacing the old pci_assign_all_busses global, that allow us to control various aspects of the PCI probing, such as whether to re-assign all resources or not, or to not try to assign anything at all. This also adds the flag x86 already has to avoid ISA alignment on bridges that don't have ISA forwarding enabled (no legacy devices on the top level bus) and sets it for PowerMacs. Signed-off-by:
-
Benjamin Herrenschmidt authored
The 32 bits PowerPC PCI code has a hack for use by some PowerMacs to try to re-open PCI<->PCI bridge IO resources that were closed by the firmware. This is no longer necessary as the generic code will now do that for us. Signed-off-by:
-
Benjamin Herrenschmidt authored
This makes the 32 bits PowerPC PCI code use the generic code to assign resources to devices that had unassigned or conflicting resources. This allow us to remove the local implementation that was incomplete and could not assign for example a PCI<->PCI bridge from scratch, which is needed on various embedded platforms. Signed-off-by:
-
Benjamin Herrenschmidt authored
There's a stale & bogus piece of code in 32 bits PCI code that complains about ISA related alignment issues. Just remove it. Signed-off-by:
-
David Gibson authored
This patch alters the kernel makefiles to build dtc from the sources embedded in the previous patch. It also changes the arch/powerpc/boot/wrapper script to use the embedded dtc, rather than expecting a copy of dtc already installed on the system. Signed-off-by:
David Gibson <david@gibson.dropbear.id.au> Signed-off-by:
-
David Gibson authored
This incorporates a copy of dtc into the kernel source, in arch/powerpc/boot/dtc-src. This commit only imports the upstream sources verbatim, a later commit will actually link it into the kernel Makefiles and use the embedded code during the kernel build. Signed-off-by:
-
Olof Johansson authored
There's nothing in holly.c that needs linux/ide.h, just remove it from the list of includes. Signed-off-by:
Olof Johansson <olof@lixom.net> Acked-by:
Josh Boyer <jwboyer@linux.vnet.ibm.com> Signed-off-by:
-
joe@perches.com authored
Signed-off-by:
Joe Perches <joe@perches.com> Signed-off-by:
-
joe@perches.com authored
Signed-off-by:
-
joe@perches.com authored
Signed-off-by:
-
joe@perches.com authored
Signed-off-by:
-
Michael Ellerman authored
PowerPC currently doesn't implement pci_set_dma_mask(), which means drivers calling it will get the generic version in drivers/pci/pci.c. The powerpc dma mapping ops include a dma_set_mask() hook, which luckily is not implemented by anyone - so there is no bug in the fact that the hook is currently never called. However in future we'll add implementation(s) of dma_set_mask(), and so we need pci_set_dma_mask() to call the hook. To save adding a hook to the dma mapping ops, pci-set_consistent_dma_mask() simply calls the dma_set_mask() hook and then copies the new mask into dev.coherenet_dma_mask. Signed-off-by:
Michael Ellerman <michael@ellerman.id.au> Signed-off-by:
-
Milton Miller authored
We have multiple calls to has_feature being inlined, but gcc can't be sure that the store via get_paca() doesn't alias the path to cur_cpu_spec->feature. Reorder to put the calls to read_purr and read_spurr adjacent to each other. To add a sense of consistency, reorder the remaining lines to perform parallel steps on purr and scaled purr of each line instead of calculating and then using one value before going on to the next. In addition, we can tell gcc that no SPURR means no PURR. The test is completely hidden in the PURR case, and in the !PURR case the second test is eliminated resulting in the simple register copy in the out-of-line branch. Further, gcc sees get_paca()->system_time referenced several times and allocates a register to address it (shadowing r13) instead of caching its value. Reading into a local varable saves the shadow of r13 and removes a potentially duplicate load (between the nested if and its parent). Signed-off-by:
Milton Miller <miltonm@bga.com> Signed-off-by:
-
Milton Miller authored
If CPU_FTR_PURR is not set, we will never set cpu_purr_data->initialized. Checking via __get_cpu_var on 64 bit avoids one dependent load compared to cpu_has_feature in the not-present case, and is always required when it is present. The code is under CONFIG_VIRT_CPU_ACCOUNTING so 32 bit will not be affected. Signed-off-by:
-
Milton Miller authored
timer_interrupt() was calculating per_cpu_offset several times, having to start from the toc because of potential aliasing issues. Placing both decrementer per_cpu varables in a struct and calculating the address once with __get_cpu_var results in better code on both 32 and 64 bit. Signed-off-by:
-
Milton Miller authored
Use __get_cpu_var(x) instead of per_cpu(x, smp_processor_id()), as it is optimized on ppc64 to access the current cpu's per-cpu offset directly; it's local_paca.offset instead of TOC->paca[local_paca->processor_id].offset. This is the trivial portion, two functions with one use each. Signed-off-by:
-
Milton Miller authored
as its only called from time_init, which is __init. Also remove unneeded forward declaration. Signed-off-by:
-
Milton Miller authored
The per-processor interrupt request register and current processor priority register are only accessed on the current cpu. In fact the hypervisor doesn't even let us choose which cpu's registers to access. The only function to use cpu twice is xics_migrate_irqs_away, not a fast path. But we can cache the result of get_hard_processor_id() instead of calling get_hard_smp_processor_id(cpu) in a loop across the call to rtas. Years ago the irq code passed smp_processor_id into get_irq, I thought we might initialize the CPPR third party at boot as an extra measure of saftey, and it made the code symmetric with the qirr (queued interrupt for software generated interrupts), but now it is just extra and sometimes unneeded work to pass it down. Signed-off-by:
-
Ishizaki Kou authored
This splits the machine definition for celleb into two definitions, one for celleb_beat, and the other for celleb_native. Though this looks complex because of sorting some functions, there are no more semantic changes than that for the splitting. Signed-off-by:
Kou Ishizaki <Kou.Ishizaki@toshiba.co.jp> Acked-by:
Arnd Bergmann <arnd@arndb.de> Signed-off-by:
-
Ishizaki Kou authored
This makes mmio_nvram_init() callable unconditionally by providing a dummy definition when CONFIG_MMIO_NVRAM is not defined. Signed-off-by:
-
Olof Johansson authored
Implement MSI support for PA Semi PWRficient platforms. MSI is done through a special range of sources on the openpic controller, and they're unfortunately breaking the usual concepts of how sources are programmed: * The source is calculated as 512 + the value written into the MSI register * The vector for this source is added to the source and reported through IACK This means that for simplicity, it makes much more sense to just set the vector to 0 for the source, since that's really the vector we expect to see from IACK. Also, the affinity/priority registers will affect 16 sources at a time. To avoid most (simple) users from being limited by this, allocate 16 sources per device but use only one. This means that there's a total of 32 sources. If we get usage scenarions that need more sources, the allocator should probably be revised to take an alignment argument and size, not just do natural alignment. Finally, since I'm already touching the MPIC names on pasemi, rename the base one from the somewhat odd " PAS-OPIC " to "PASEMI-OPIC". Signed-off-by:
-
Stephen Rothwell authored
Commit fbd568a3 ("Change synchronize_kernel to _rcu and _sched") changed the deprecated synchronize_kernel() in HvLpEvent_unregisterHandler() to synchronize_rcu(). It turns out that it should have been synchronize_sched(). Signed-off-by:
Stephen Rothwell <sfr@canb.auug.org.au> Acked-by:
Paul E. McKenney <paulmck@linux.vnet.ibm.com> Signed-off-by:
-
Scott Wood authored
Many operations, as currently used in the wrapper, assume they can pass NULL and have it be treated as the root node. However, libfdt-wrapper converts NULL to -1, which is only appropriate when searching for nodes, and will cause an error otherwise. Signed-off-by:
Scott Wood <scottwood@freescale.com> Acked-by:
-
Scott Wood authored
fdt_wrapper_create_node passes a variable called offset to offset_devp(), which uses said parameter to initialize a local variable called offset. Due to one of the odder aspects of the C language, the result is an undefined variable, with no error or warning. Signed-off-by:
-
Balbir Singh authored
Here's a dumb simple implementation of fake NUMA nodes for PowerPC. Fake NUMA nodes can be specified using the following command line option numa=fake=<node range> node range is of the format <range1>,<range2>,...<rangeN> Each of the rangeX parameters is passed using memparse(). I find this useful for fake NUMA emulation on my simple PowerPC machine. I've tested it on a non-numa box with the following arguments: numa=fake=1G numa=fake=1G,2G name=fake=1G,512M,2G numa=fake=1500M,2800M mem=3500M numa=fake=1G mem=512M numa=fake=1G mem=1G Signed-off-by:
Balbir Singh <balbir@linux.vnet.ibm.com> Acked-by:
-
Grant Likely authored
The machine initcall macros allow initcalls to be registered which test machine_is() before executing the initcall. Signed-off-by:
Grant Likely <grant.likely@secretlab.ca> Signed-off-by:
-
Olof Johansson authored
Optimize MPIC IPIs, by passing in the IPI number as the argument to the handler, since all we did was translate it back based on which mpic the interrupt came though on (and that was always the primary mpic). Signed-off-by:
-
Paul Mackerras authored
Merge branch 'for-2.6.25' of git://git.kernel.org/pub/scm/linux/kernel/git/arnd/cell-2.6 into for-2.6.25
-
- 19 Dec, 2007 10 commits
-
-
Johannes Berg authored
This adds platform_suspend_ops for PMU based machines, directly in the PMU driver. This allows suspending via /sys/power/state on powerbooks. The patch also replaces the PMU ioctl with a simple call to pm_suspend(PM_SUSPEND_MEM). Additionally, it cleans up some debug code. Signed-off-by:
Johannes Berg <johannes@sipsolutions.net> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by:
-
Paul Mackerras authored
Sleep on the powerbook 3400 has been broken since the change that made powerbook_sleep_3400 call pmac_suspend_devices(), which disables interrupts. There are a couple of loops in powerbook_sleep_3400 that depend on interrupts being enabled, and in fact it has to have interrupts enabled at the point of going to sleep since it is an interrupt from the PMU that wakes it up. This fixes it by using pmu_wait_complete() instead of a spinloop, and by explicitly enabling interrupts before putting the CPU into sleep mode (which is OK since all interrupts except the PMU interrupt have been disabled at the interrupt controller by this stage). This changes the logic so that it keeps putting the CPU into sleep mode until the completion of the interrupt transaction from the PMU that signals the end of sleep. Also, we now call pmu_unlock() before sleep so that the via_pmu_interrupt() code can process the interrupt event from the PMU properly. Now that generic code saves and restores PCI state, it is no longer necessary to do that here. Thus pbook_pci_save/restore and related functions are no longer necessary, so this removes them. Lastly, this moves the ioremap of the memory controller to init code rather than doing it on every sleep/wakeup cycle. Signed-off-by: -
Paul Mackerras authored
This is fairly straightforward, and lets us get rid of x.completion as well. Signed-off-by: -
Paul Mackerras authored
This converts the therm_pm72.c driver to use the kthread API. I thought about making it use kthread_stop() instead of the `state' variable and the `ctrl_complete' completion, but that isn't simple and will require changing the way that `state' is used. Signed-off-by: -
Paul Mackerras authored
This converts adb.c to use the kthread API. It also changes adb_request so that if the ADBREQ_SYNC flag is specified, we now sleep waiting for the request to finish using an on-stack completion rather than spinning. To implement this, we now require that if the ADBREQ_SYNC flag is set, the `done' parameter must be NULL. All of the existing callers of adb_request that pass ADBREQ_SYNC appear to be in process context and have done == NULL. Doing this allows us to get rid of an awful hack in adb_request() where we used to test whether the request was coming from the adb probe task and use a completion if it was, and otherwise spin. This also gets rid of a static request block that was used if the req parameter to adb_request was NULL. None of the callers do that any more, so the static request block is no longer necessary. Signed-off-by: -
Paul Mackerras authored
We aren't supposed to use kernel_thread directly in drivers any more, and in fact using kthread_run is a bit simpler. Signed-off-by: -
Johannes Berg authored
This kills off the remnants of the old sleep notifiers now that they are no longer used. Signed-off-by:
-
Johannes Berg authored
This replaces the pmu sleep notifier that adb had with suspend/resume hooks in a new platform driver/device. Signed-off-by:
-
Jeremy Kerr authored
We're currently getting a warning from not checking the result of sysfs_create_group, which is declared as __must_check. This change introduces appropriate error-handling for spu_add_sysdev_attr_group() Signed-off-by:
Jeremy Kerr <jk@ozlabs.org> Signed-off-by:
-
Jeremy Kerr authored
Currently, we have a possibilty that the SLBs setup during context switch don't cover the entirety of the necessary lscsa and code regions, if these regions cross a segment boundary. This change checks the start and end of each region, and inserts a SLB entry for each, if unique. We also remove the assumption that the spu_save_code and spu_restore_code reside in the same segment, by using the specific code array for save and restore. Signed-off-by:
-
