summaryrefslogtreecommitdiff
path: root/linux-core
AgeCommit message (Collapse)Author
2008-03-16[via] Allow a little larger stride for SG DMA DownloadFromScreen.Thomas Hellstrom
2008-03-16ati: fix rs690 igp gart by allocating the page table in 32-bit memoryDave Airlie
2008-03-16drm: this u32 should be a dma_addr_tDave Airlie
doesn't fix anything but just making it consistent
2008-03-14fix build problemsAlan Hourihane
2008-03-14drm: Fix race that can lockup the kernelMike Isely
The i915_vblank_swap() function schedules an automatic buffer swap upon receipt of the vertical sync interrupt. Such an operation is lengthy so it can't be allowed to happen in normal interrupt context, thus the DRM implements this by scheduling the work in a kernel softirq-scheduled tasklet. In order for the buffer swap to work safely, the DRM's central lock must be taken, via a call to drm_lock_take() located in drivers/char/drm/drm_irq.c within the function drm_locked_tasklet_func(). The lock-taking logic uses a non-interrupt-blocking spinlock to implement the manipulations needed to take the lock. This semantic would be safe if all attempts to use the spinlock only happen from process context. However this buffer swap happens from softirq context which is really a form of interrupt context. Thus we have an unsafe situation, in that drm_locked_tasklet_func() can block on a spinlock already taken by a thread in process context which will never get scheduled again because of the blocked softirq tasklet. This wedges the kernel hard. To trigger this bug, run a dual-head cloned mode configuration which uses the i915 drm, then execute an opengl application which synchronizes buffer swaps against the vertical sync interrupt. In my testing, a lockup always results after running anywhere from 5 minutes to an hour and a half. I believe dual-head is needed to really trigger the problem because then the vertical sync interrupt handling is no longer predictable (due to being interrupt-sourced from two different heads running at different speeds). This raises the probability of the tasklet trying to run while the userspace DRI is doing things to the GPU (and manipulating the DRM lock). The fix is to change the relevant spinlock semantics to be the interrupt-blocking form. After this change I am no longer able to trigger the lockup; the longest test run so far was 20 hours (test stopped after that point). Note: I have examined the places where this spinlock is being employed; all are reasonably short bounded sequences and should be suitable for interrupts being blocked without impacting overall kernel interrupt response latency. Signed-off-by: Mike Isely <isely@pobox.com>
2008-03-12Merge branch 'intel-post-reloc'Thomas Hellstrom
Conflicts: linux-core/drm_compat.c linux-core/drm_compat.h linux-core/drm_ttm.c shared-core/i915_dma.c Bump driver minor to 13 due to introduction of new relocation type.
2008-03-12Bug # 14712Thomas Hellstrom
Disable page saving for GPU read-only TTMs.
2008-03-12Dont allow !sysadmin clients to alter the memory type ofThomas Hellstrom
NO_EVICT buffers.
2008-03-12Fix for debug memory routinesAlan Hourihane
2008-03-12Add error messageAlan Hourihane
2008-03-12Add an emergency pinnable memory quota for root-only processes.Thomas Hellstrom
2008-03-12Fix kernel crash when we hit OOM conditions.Thomas Hellstrom
(Alan Hourihane)
2008-03-12Avoid duplicate calls to drm_ttm_bind in some cases.Thomas Hellstrom
2008-03-12Make sure other TTM memory types than TT is really unbound when evicted.Thomas Hellstrom
2008-03-10remove unneeded load callDave Airlie
2008-03-07flush_agp_mappings commitDave Airlie
2008-03-06drm/bo: allow non-suser priv to add kernel BOs.Dave Airlie
modprobe can be run with dropped capabilities we still want the kernel bos to work.
2008-03-05drm: Fix for non-coherent DMA PowerPCBenjamin Herrenschmidt
This patch fixes bits of the DRM so to make the radeon DRI work on non-cache coherent PCI DMA variants of the PowerPC processors. It moves the few places that needs change to wrappers to that other architectures with similar issues can easily add their own changes to those wrappers, at least until we have more useful generic kernel API. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2008-03-06ttm: make sure userspace can't destroy kernel create memory managersDave Airlie
this adds something to say the kernel initialised the memory region not the userspace. and blocks userspace from deallocating kernel areas
2008-03-06drm/ttm: add ioctl to get back memory managed area sizedDave Airlie
taken from modesetting branch but could be useful outside it.
2008-03-06drm: reorganise minor number handling using code from modesetting branchDave Airlie
Rip out the whole head thing and replace it with an idr and drm_minor structure.
2008-03-05i915: Evict if relocatee buffer is CACHED_MAPPED beforeXiang, Haihao
writting relocations, otherwise the GPU probably sees some inconsistent data. Fix fd.o bug#14656
2008-03-04Clarify through the names what drm_ttm_alloc_pages() and friend actually did.Eric Anholt
These are all about the page directory (pointers to pages) rather than the actual pages backing the allocation.
2008-02-29FIX_KMAP_BEGIN requires CONFIG_HIMEM (see include/asm-i386.h/fixmap.h)Patrice Mandin
2008-02-29Working revision.Thomas Hellstrom
2008-02-29Fix compilation breakage on x86-64.Thomas Hellstrom
2008-02-29More post-ioctl work.Thomas Hellstrom
2008-02-29agp: export the correct symbolDave Airlie
2008-02-28Add a compat kmap_atomic_prot_pfn to do quick kernel map / unmaps ofThomas Hellstrom
PCI- or high memory. This is substantially more efficient than drm_bo_kmap, since the mapping only lives on a single processor. Unmapping is done use kunmap_atomic(). Flushes only a single tlb() entry. Add a support utility int drm_bo_pfn_prot() that returns the pfn and desired page protection for a given bo offset. This is all intended for relocations in bound TTMS or vram. Mapping-accessing-unmapping must be atomic, either using preempt_xx() macros or a spinlock.
2008-02-28Initial commit.Thomas Hellstrom
2008-02-28Add a compat kmap_atomic_prot_pfn to do quick kernel map / unmaps ofThomas Hellstrom
PCI- or high memory. This is substantially more efficient than drm_bo_kmap, since the mapping only lives on a single processor. Unmapping is done use kunmap_atomic(). Flushes only a single tlb() entry. Add a support utility int drm_bo_pfn_prot() that returns the pfn and desired page protection for a given bo offset. This is all intended for relocations in bound TTMS or vram. Mapping-accessing-unmapping must be atomic, either using preempt_xx() macros or a spinlock.
2008-02-22i915: put ARX back into index mode before doing restoreJesse Barnes
Fixes resume from hibernate in some configurations.
2008-02-21Don't free irq resources until after we've unregistered the handler.Kristian Høgsberg
2008-02-21linux: Clean up vblank related resources from drm_irq_uninstall().Michel Dänzer
This fixes at least two problems: * The vblank_disable_fn timer callback could get called after the DRM was de-initialized, e.g. on X server shutdown. * Leak of vblank related resources when disabling and re-enabling the IRQ, e.g. on an X server reset.
2008-02-20drm: add support for passing state into the suspend hooks.Dave Airlie
fix i915 driver to use state for hibernate save avoidance. Signed-off-by: Dave Airlie <airlied@redhat.com>
2008-02-16[915]: more registers for S3 (DSPCLK_GATE_D, CACHE_MODE_0, MI_ARB_STATE)Keith Packard
Failing to preserve the MI_ARB_STATE register was causing FIFO underruns on the VGA output on my HP 2510p after resume.
2008-02-07Restore pipeconf regs unconditionallyJesse Barnes
On many chipsets, the checks for DPLL enable or VGA mode will prevent the pipeconf regs from being restored, which could result in a blank display or X failing to come back after resume. So restore them unconditionally along with actually restoring pipe B's palette correctly.
2008-02-07i915: save/restore interrupt stateJesse Barnes
On resume, if the interrupt state isn't restored correctly, we may end up with a flood of unexpected or ill-timed interrupts, which could cause the kernel to disable the interrupt or vblank events to happen at the wrong time. So save/restore them properly.
2008-02-07Fix vblank enable/disable callbacksJesse Barnes
There were two problems with the existing callback code: the vblank enable callback happened multiple times per disable, making drivers more complex than they had to be, and there was a race between the final decrement of the vblank usage counter and the next enable call, which could have resulted in a put->schedule disable->get->enable->disable sequence, which would be bad. So add a new vblank_enabled array to track vblank enable on per-pipe basis, and add a lock to protect it along with the refcount + enable/disable calls to fix the race.
2008-02-05i915: Re-report breadcrumbs on poll to the fence manager,Thomas Hellstrom
since a breadcrumb may actually turn up before a corresponding fence object has been placed on the fence ring.
2008-01-31Add an fence_class_manager::last_queued_sequence member, since aThomas Hellstrom
sequence number may actually turn up before the corresponding fence object has been queued on the ring. Fence drivers can use this member to determine whether a sequence number must be re-reported.
2008-01-30Simplify the fencing code and differentiate between flushes andThomas Hellstrom
waiting types. Add a "command_stream_barrier" method to the bo driver.
2008-01-28Fix hibernate save/restore of VGA attribute regsJesse Barnes
In hibernate, we may end up calling the VGA save regs function twice, so we need to make sure it's idempotent. That means leaving ARX in index mode after the first save operation. Fixes hibernate on 965.
2008-01-24Fixup modeset ioctl number & typedef usageJesse Barnes
Should be 0x08 rather than 0xa0, and shouldn't use typedefs.
2008-01-23i915/flush: get the ret the right way aroundDave Airlie
2008-01-22Merge branch 'master' into vblank-rework, including mach64 supportJesse Barnes
Conflicts: linux-core/drmP.h linux-core/drm_drv.c shared-core/i915_drv.h shared-core/i915_irq.c shared-core/mga_irq.c shared-core/radeon_irq.c shared-core/via_irq.c Mostly trivial conflicts. mach64 support from Mathieu Bérard.
2008-01-17i915: fix invalid opcode exception on cpus without clflushKyle McMartin
i915_flush_ttm was unconditionally executing a clflush instruction to (obviously) flush the cache. Instead, check if the cpu supports clflush, and if not, fall back to calling wbinvd to flush the entire cache. Signed-off-by: Kyle McMartin <kmcmartin@redhat.com>
2008-01-15i915: Add chipset id for Intel Integrated Graphics DeviceZhenyu Wang
This adds new chipset id in drm. Signed-off-by: Zhenyu Wang <zhenyu.z.wang@intel.com>
2008-01-15Define i915_compat.c upper_32_bits for kernels < 2.6.21Thomas Hellstrom
2008-01-15 this is to fix a deadloop in drm hang system issue.Zou Nan hai
(1 << bits) is an undefined value when bits == 32. gcc may generate 1 with this expression which will lead to an infinite retry loop in drm_ht_just_insert_please. Because of the different implement of hash_long, this issue is more frequenly see on 64 bit system