- 09 Jan, 2006 40 commits
-
-
David Gibson authored
Oops, forgot to compile the VMALLOCBASE/VMALLOC_START patch on iSeries. VMALLOC_START is defined in pgtable.h whereas previously VMALLOCBASE was previously defined in page.h. lparmap.c needs to be updated appropriately. Booted on iSeries RS64 (now). Signed-off-by:
David Gibson <dwg@au1.ibm.com> Signed-off-by:
Paul Mackerras <paulus@samba.org>
-
Andy Fleming authored
This patch adds oprofile support for the 7450 and all its multitudinous derivatives. * Added 7450 (and derivatives) support for oprofile * Changed e500 cputable to have oprofile model and cpu_type fields * Added support for classic 32-bit performance monitor interrupt * Cleaned up common powerpc oprofile code to be as common as possible * Cleaned up oprofile_impl.h to reflect 32 bit classic code * Added 32-bit MMCRx bitfield definitions and SPR numbers Signed-off-by:
Andy Fleming <afleming@freescale.com> Signed-off-by:
-
Kumar Gala authored
pci_address_to_pio is missing a closing curly brace Signed-off-by:
Kumar Gala <galak@kernel.crashing.org> Signed-off-by:
-
Benjamin Herrenschmidt authored
This fixes pci_address_to_pio() to return an unsigned long (to be safe) and fixes a bug in the implementation that caused it to return a bogus IO port number Signed-off-by:
Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by:
-
Kristian Mueller authored
Fix compilation of via-pmu.c without Power Management support. Signed-off-by: -
David Gibson authored
On ppc64, we independently define VMALLOCBASE and VMALLOC_START to be the same thing: the start of the vmalloc() area at 0xd000000000000000. VMALLOC_START is used much more widely, including in generic code, so this patch gets rid of the extraneous VMALLOCBASE. This does require moving the definitions of region IDs from page_64.h to pgtable.h, but they don't clearly belong in the former rather than the latter, anyway. While we're moving them, clean up the definitions of the REGION_IDs: - Abolish REGION_SIZE, it was only used once, to define REGION_MASK anyway - Define the specific region ids in terms of the REGION_ID() macro. - Define KERNEL_REGION_ID in terms of PAGE_OFFSET rather than KERNELBASE. It amounts to the same thing, but conceptually this is about the region of the linear mapping (which starts at PAGE_OFFSET) rather than of the kernel text itself (which is at KERNELBASE). Signed-off-by:
David Gibson <david@gibson.dropbear.id.au> Signed-off-by:
-
Benjamin Herrenschmidt authored
The patch enabling the new G5's with U4 broke initialization of the DART driver, causing it to trigger a BUG_ON for a case that is actually valid. This patch fixes it: Signed-off-by:
-
Benjamin Herrenschmidt authored
This adds some very basic support for the new machines, including the Quad G5 (tested), and other new dual core based machines and iMac G5 iSight (untested). This is still experimental ! There is no thermal control yet, there is no proper handing of MSIs, etc.. but it boots, I have all 4 cores up on my machine. Compared to the previous version of this patch, this one adds DART IOMMU support for the U4 chipset and thus should work fine on setups with more than 2Gb of RAM. Signed-off-by:
-
Kumar Gala authored
Added the ability to determine if an outbound window in the PCI host controller is for prefetchable memory and report it as such. Signed-off-by:
-
linas authored
There is code in the RPAPHP directory that is identical to this routine; I'll be removing that code in an upcoming patch, but this patch is needed to expose the function to make it callable. Signed-off-by:
Linas Vepstas <linas@austin.ibm.com> Signed-off-by:
-
Segher Boessenkool authored
Cleanup the MPIC IO-APIC workarounds, make them a bit more generic, smaller and faster. Signed-off-by:
Segher Boessenkool <segher@kernel.crashing.org> Signed-off-by:
-
Benjamin Herrenschmidt authored
The pre-parsed addrs/n_addrs fields in struct device_node are finally gone. Remove the dodgy heuristics that did that parsing at boot and remove the fields themselves since we now have a good replacement with the new OF parsing code. This patch also fixes a bunch of drivers to use the new code instead, so that at least pmac32, pseries, iseries and g5 defconfigs build. Signed-off-by:
-
Benjamin Herrenschmidt authored
The platinumfb driver used only on some powermacs has an issue with some video modes & limited VRAM. This patch fixes it. Signed-off-by:
-
Benjamin Herrenschmidt authored
This adds a defconfig for PowerMac with ARCH=powerpc Signed-off-by:
-
Anton Blanchard authored
Remove a comment in head.S which is no longer relevant. Signed-off-by:
Anton Blanchard <anton@samba.org> Signed-off-by:
-
Anton Blanchard authored
Milton and I were looking at the cputable code and it looks like we can set spurious bits on 64bit. Signed-off-by:
-
Anton Blanchard authored
We used to print a NUMA cpu summary at boot before the hotplug cpu code was added. This has been useful for catching machine configuration as well as firmware bugs in the past. This patch restores that functionality. An example of the output is: Node 0 CPUs: 0-7 Node 1 CPUs: 8-15 Signed-off-by:
-
Nicolas Kaiser authored
Add three files not mentioned in Documentation/powerpc/00-INDEX. Sort alphabetically. Signed-off-by:
Nicolas Kaiser <nikai@nikai.net> Signed-off-by:
-
Arnd Bergmann authored
/dev/nvram uses the user-provided read/write size for kmalloc, which fails, if a large number is passed. This will always use a single page at most, which can be expected to succeed. Signed-off-by:
Arnd Bergmann <arndb@de.ibm.com> Signed-off-by:
-
David Woodhouse authored
We currently crash in the fedora installer because the keyboard driver tries to access I/O space that is not there on our hardware. This uses the same solution as powermac by just marking all legacy i/o as invalid. Signed-off-by:
-
Jens.Osterkamp@de.ibm.com authored
So far, the iommu code was hardwired to a linear mapping between 0x20000000 and 0x40000000, so it could only support 512MB of RAM. This patch still keeps the linear mapping, but looks for proper ibm,dma-window properties to set up larger windows, this makes the maximum supported RAM size 2GB. If there is anything unusual about the dma-window properties, we fall back to the old behavior. We also support switching off the iommu completely now with the regular iommu=off command line option. Signed-off-by:
-
Arnd Bergmann authored
Because of always clearing DSISR at spu class 1 interrupt handler, kernel may lose Class1[Mf] interrupt. Signed-off-by:
Masato Noguchi <Masato.Noguchi@jp.sony.com> Signed-off-by:
Geoff Levand <geoff.levand@am.sony.com> Signed-off-by:
-
Arnd Bergmann authored
Output from hexdump with "%08x" depends on HOST platform's endian. When building linux by cross toolchain, that difference makes errors. Signed-off-by:
-
Arnd Bergmann authored
One of my last patches contained a broken line from splitting out some other changes, this restores a working version. Signed-off-by:
-
Arnd Bergmann authored
One of the two users of spufs_calls.owner still has a race when calling try_module_get while the module is removed. This makes it use the correct instance of owner. Noticed by Milton Miller. Signed-off-by:
-
Kumar Gala authored
This patch adds support for the TQ Components TQM85xx modules. Currently the modules TQM8540/8541/8555/8560 are supported. Signed-off-by:
Stefan Roese <sr@denx.de> Signed-off-by:
-
Arnd Bergmann authored
I can't really get a conclusive answer from the firmware people what to check for, so I just try scanning for anything that starts with "IBM,CPB", which should be correct for all hardware produced so far and for systemsim. Signed-off-by:
-
Arnd Bergmann authored
Handling mailbox interrupts was broken in multiple respects, the combination of which was hiding the bugs most of the time. - The ibox interrupt mask was open initially even though there are no waiters on a newly created SPU. - Acknowledging the mailbox interrupt did not work because it is level triggered and the mailbox data is never retrieved from inside the interrupt handler. - The interrupt handler delivered interrupts with a disabled mask if another interrupt is triggered for the same class but a different mask. - The poll function did not enable the interrupt if it had not been enabled, so we might run into the poll timeout if none of the other bugs saved us and no signal was delivered. We probably still have a similar problem with blocking read/write on mailbox files, but that will result in extra wakeup in the worst case, not in incorrect behaviour. Signed-off-by:
-
Arnd Bergmann authored
This patch reduces lock complexity of SPU scheduler, particularly for involuntary preemptive switches. As a result the new code does a better job of mapping the highest priority tasks to SPUs. Lock complexity is reduced by using the system default workqueue to perform involuntary saves. In this way we avoid nasty lock ordering problems that the previous code had. A "minimum timeslice" for SPU contexts is also introduced. The intent here is to avoid thrashing. While the new scheduler does a better job at prioritization it still does nothing for fairness. From: Mark Nutter <mnutter@us.ibm.com> Signed-off-by:
-
Arnd Bergmann authored
This patch makes it easier to preempt an SPU context by having the scheduler hold ctx->state_sema for much shorter periods of time. As part of this restructuring, the control logic for the "run" operation is moved from arch/ppc64/kernel/spu_base.c to fs/spufs/file.c. Of course the base retains "bottom half" handlers for class{0,1} irqs. The new run loop will re-acquire an SPU if preempted. From: Mark Nutter <mnutter@us.ibm.com> Signed-off-by: -
Arnd Bergmann authored
spufs is rather noisy when debugging is enabled, this turns off the messages for production use. Signed-off-by:
-
Arnd Bergmann authored
try_module_get returns true when NULL arguments, so we first need to check if there is a module loaded before getting the reference count. Signed-off-by:
-
Arnd Bergmann authored
With the new rules for reserved pages, the spufs now needs working page reference counting. I should probably look into converting to vm_insert_page, but for now this patch makes spufs work again. Signed-off-by:
-
Arnd Bergmann authored
This changes all exported symbols of spufs to EXPORT_SYMBOL_GPL. The spu_ibox_read/spu_wbox_write symbols are not exported any more when the scheduler patch is applied. Signed-off-by:
-
Michael Ellerman authored
Milton has proposed that we should support a "linux,usable-memory" property on memory nodes which describes, in preference to "reg", the regions of memory Linux should use. This facility is required for kdump to inform the second kernel which memory it should use. Signed-off-by:
Haren Myneni <haren@us.ibm.com> Signed-off-by:
Michael Ellerman <michael@ellerman.id.au> Signed-off-by:
-
Michael Ellerman authored
Signed-off-by:
-
Michael Ellerman authored
This patch adds code to parse and setup the crash kernel resource in the first kernel. PPC64 ignores the @x part, we always run at 32 MB. Signed-off-by:
-
Michael Ellerman authored
Implementing the machine_crash_shutdown which will be called by crash_kexec (called in case of a panic, sysrq etc.). Disable the interrupts, shootdown cpus using debugger IPI and collect regs for all CPUs. elfcorehdr= specifies the location of elf core header stored by the crashed kernel. This command line option will be passed by the kexec-tools to capture kernel. savemaxmem= specifies the actual memory size that the first kernel has and this value will be used for dumping in the capture kernel. This command line option will be passed by the kexec-tools to capture kernel. Signed-off-by:
-
Michael Ellerman authored
There's a few places where we need to fix things up for the kernel to work if it's linked at 32MB: - platforms/powermac/smp.c To start secondary cpus on pmac we patch the reset vector, which is fine. Except if we're above 32MB we don't have enough bits for an absolute branch, it needs to relative. - kernel/head_64.s - A few branches in the cpu hold code need to load the full target address and do a bctr. - after_prom_start needs to load PHYSICAL_START as the dest address, not 0. - The exception prolog needs to load the low word of the target adddress, not just the low halfword. - Fixup handling of the initial stab address. - kernel/setup_64.c smp_release_cpus() needs to write 1 to the spinloop flag near 0, not 32 MB. Signed-off-by: -
Michael Ellerman authored
Regardless of where the kernel's linked we always get interrupts at low addresses. This patch creates a trampoline in the first 3 pages of memory, where interrupts land, and patches those addresses to jump into the real kernel code at PHYSICAL_START. We also need to reserve the trampoline code and a bit more in prom.c Signed-off-by:
-
