summaryrefslogtreecommitdiffstats
path: root/drivers/gpu/drm (follow)
Commit message (Collapse)AuthorAgeFilesLines
* drm/i915: replace DRM_DEBUG with DRM_DEBUG_KMS in intel_lvdsyakui_zhao2009-06-111-2/+5
| | | | | | | | Use the DRM_DEBUG_KMS macro definition to print the debug info for the LVDS. Signed-off-by: Zhao Yakui <yakui.zhao@intel.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
* drm: add separate drm debugging levelsyakui_zhao2009-06-111-1/+16
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Now all the DRM debug info will be reported if the boot option of "drm.debug=1" is added. Sometimes it is inconvenient to get the debug info in KMS mode. We will get too much unrelated info. This will separate several DRM debug levels and the debug level can be used to print the different debug info. And the debug level is controlled by the module parameter of drm.debug In this patch it is divided into four debug levels; drm_core, drm_driver, drm_kms, drm_mode. At the same time we can get the different debug info by changing the debug level. This can be done by adding the module parameter. Of course it can be changed through the /sys/module/drm/parameters/debug after the system is booted. Four debug macro definitions are provided. DRM_DEBUG(fmt, args...) DRM_DEBUG_DRIVER(prefix, fmt, args...) DRM_DEBUG_KMS(prefix, fmt, args...) DRM_DEBUG_MODE(prefix, fmt, args...) When the boot option of "drm.debug=4" is added, it will print the debug info using DRM_DEBUG_KMS macro definition. When the boot option of "drm.debug=6" is added, it will print the debug info using DRM_DEBUG_KMS/DRM_DEBUG_DRIVER. Sometimes we expect to print the value of an array. For example: SDVO command, In such case the following four DRM debug macro definitions are added: DRM_LOG(fmt, args...) DRM_LOG_DRIVER(fmt, args...) DRM_LOG_KMS(fmt, args...) DRM_LOG_MODE(fmt, args...) Signed-off-by: Zhao Yakui <yakui.zhao@intel.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
* radeon: remove _DRM_DRIVER from the preadded sarea mapDave Airlie2009-06-111-1/+1
| | | | | | | This shouldn't be there and is what broke r600 late in the 2.6.30 release cycle with Ben's patch. Signed-off-by: Dave Airlie <airlied@redhat.com>
* drm: don't associate _DRM_DRIVER maps with a masterBen Skeggs2009-06-111-1/+2
| | | | | | | | | | | | A driver will use the _DRM_DRIVER map flag to indicate that it wants to be responsible for removing the map itself, bypassing the DRM's automagic cleanup code. Since the multi-master changes this has been broken, resulting in some drivers having their registers unmapped before it's finished with them. Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
* drm: simplify kcalloc() call to kzalloc().Robert P. J. Day2009-06-111-1/+1
| | | | | | | | | | | Calls to kcalloc() for a single element can be simplified to calls to kzalloc(). Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca> Cc: Dave Airlie <airlied@linux.ie> Cc: Eric Anholt <eric@anholt.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Dave Airlie <airlied@redhat.com>
* intelfb: fix spelling of "CLOCK"Pavel Roskin2009-06-111-1/+1
| | | | | | | | Signed-off-by: Pavel Roskin <proski@gnu.org> Cc: Eric Anholt <eric@anholt.net> Cc: Dave Airlie <airlied@linux.ie> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Dave Airlie <airlied@redhat.com>
* drm/i915: Hook connector to encoder during load detection (fixes tv/vga detect)Keith Packard2009-06-101-0/+2
| | | | | | | | | With the DRM-driven DPMS code, encoders are considered idle unless a connector is hooked to them, so mode setting is skipped. This makes load detection fail as none of the hardware is enabled. Signed-off-by: Keith Packard <keithp@keithp.com> Signed-off-by: Eric Anholt <eric@anholt.net>
* drm/i915: NOMEM->NOSPCChris Wilson2009-06-091-5/+5
| | | | | | | | To differentiate between encountering an out-of-memory error with running out of space in the aperture, use ENOSPC for the later. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Eric Anholt <eric@anholt.net>
* drm/i915: use I915_GEM_GPU_DOMAINSChris Wilson2009-06-091-13/+9
| | | | | Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Eric Anholt <eric@anholt.net>
* drm/i915: no need to hold mutex for object lookupChris Wilson2009-06-091-2/+1
| | | | | Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Eric Anholt <eric@anholt.net>
* drm/i915: Call drm_vblank_post_modeset() on error paths.Chris Wilson2009-06-091-4/+2
| | | | | | | | Ensure that the drm_vblank_pre_modeset() is always balanced by drm_vblank_post_modeset() within intel_crtc_mode_set(). Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Eric Anholt <eric@anholt.net>
* drm/i915: OR in the COMMAND read domain for the batch buffer.Chris Wilson2009-06-091-2/+6
| | | | | | | | | The batch buffer may be shared with another read buffer, so we should not ignore any previously set domains, but just or in the command domain (and check that the buffer is not writable). Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Eric Anholt <eric@anholt.net>
* drm/i915: Sanity check execbuffer arguments before touching state.Chris Wilson2009-06-091-11/+27
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | By sending a broken execbuffer (its length was not suitably aligned) I triggered an operation upon a freed object. The invalid alignment was discovered after updating the write_domain on the object but before the object was placed on the active queue. So during the unwind process following the error, the now freed object attempts to flush its non-existent, but outstanding, GPU writes causing this use-after-free. [drm:i915_dispatch_gem_execbuffer] *ERROR* alignment [drm:i915_gem_execbuffer] *ERROR* dispatch failed -22 WARNING: at lib/kref.c:43 warn_slowpath_null+0x10/0x15() Modules linked in: Pid: 4552, comm: lt-csi-drm Not tainted 2.6.30-rc6 #423 Call Trace: [<c0119ef3>] warn_slowpath_fmt+0x57/0x6d [<c014de24>] ? get_pageblock_migratetype+0x18/0x1e [<c014e8fd>] ? free_hot_page+0xa/0xc [<c014e915>] ? __free_pages+0x16/0x1f [<c0153ebf>] ? shmem_truncate_range+0x63e/0x656 [<c015fb2f>] ? slob_page_alloc+0x146/0x1c8 [<c0119f19>] warn_slowpath_null+0x10/0x15 [<c01f55f2>] kref_get+0x1b/0x21 [<c02605db>] i915_gem_object_move_to_active+0x1f/0x56 [<c0261302>] i915_add_request+0x156/0x19a [<c026136e>] i915_gem_object_flush_gpu_write_domain+0x28/0x3f [<c0261eca>] i915_gem_object_unbind+0x4a/0x124 [<c0261fd7>] i915_gem_free_object+0x33/0x9b [<c0250d6b>] drm_gem_object_free+0x28/0x4a [<c0250d43>] ? drm_gem_object_free+0x0/0x4a [<c01f55ce>] kref_put+0x38/0x41 [<c0250cbf>] drm_gem_object_unreference+0x11/0x13 [<c0250d06>] drm_gem_object_handle_unreference+0x1e/0x21 [<c0250d13>] drm_gem_object_release_handle+0xa/0xe [<c01f3e6b>] idr_for_each+0x5f/0x98 [<c0250d09>] ? drm_gem_object_release_handle+0x0/0xe [<c0250daf>] drm_gem_release+0x22/0x34 [<c025046f>] drm_release+0x1e8/0x3c4 [<c0162d25>] __fput+0xaf/0x146 [<c0162dce>] fput+0x12/0x14 [<c01605ef>] filp_close+0x48/0x52 [<c011b182>] put_files_struct+0x57/0x9b [<c011b1e4>] exit_files+0x1e/0x20 [<c011c6b6>] do_exit+0x16d/0x511 [<c03704ab>] ? __schedule+0x3d4/0x3e5 [<c0103f0d>] ? handle_irq+0xd/0x69 [<c011caa7>] do_group_exit+0x4d/0x73 [<c011cae0>] sys_exit_group+0x13/0x17 [<c010268c>] sysenter_do_call+0x12/0x2b Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Eric Anholt <eric@anholt.net>
* drm/i915: Skip lvds with Aopen i945GTt-VFAMichael Cousin2009-06-091-0/+7
| | | | | Signed-off-by: Michael Cousin <mika.cousin@gmail.com> Signed-off-by: Eric Anholt <eric@anholt.net>
* drm/i915: handle interrupt on new chipsetZhenyu Wang2009-06-094-8/+190
| | | | | | | | | | Update interrupt handling methods for IGDNG with new registers for display and graphics interrupt functions. As we won't use irq-based vblank sync in dri2, so display interrupt on new chip will be used for hotplug only in future. Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com> Signed-off-by: Eric Anholt <eric@anholt.net>
* drm/i915: enable MCHBAR if neededJesse Barnes2009-06-092-0/+147
| | | | | | | | | | | | | | | Using the new PNP resource checking code, this patch allows the i915 driver to allocate MCHBAR space if needed and use the BAR to determine current memory settings. [apw@canonical.com: moved to the new generic PNP resource interface] Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org> Signed-off-by: Andy Whitcroft <apw@canonical.com> Signed-off-by: Eric Anholt <eric@anholt.net> failure to update-index after git-am --reject to hand-apply Signed-off-by: Eric Anholt <eric@anholt.net>
* drm/i915: avoid non-atomic sysrq executionJesse Barnes2009-06-051-1/+7
| | | | | | | | | | | The sysrq functions are executed in hardirq context, so we shouldn't be calling sleeping functions from them, like mutex_locks or memory allocations. Fix up the i915 sysrq handler to avoid this. Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org> Signed-off-by: Eric Anholt <eric@anholt.net>
* drm/i915: apply G45 vblank count code to all G4x chips and fix max_frame_countJesse Barnes2009-06-052-3/+4
| | | | | | | | All G4x and newer chips use the new style frame count register, with a full 32 bit frame count. Update the code to reflect this. Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org> Signed-off-by: Eric Anholt <eric@anholt.net>
* drm/i915: add ignore lvds quirk info for AOpen Mini PCJarod Wilson2009-06-051-2/+8
| | | | | | | | | | Fix a FIXME in the intel LVDS bring-up code, adding the appropriate blacklist entry for the AOpen Mini PC, courtesy of a dmidecode dump from Florian Demmer. Signed-off-by: Jarod Wilson <jarod@redhat.com> CC: Florian Demmer <florian@demmer.org> Signed-off-by: Eric Anholt <eric@anholt.net>
* drm/i915: Initialize the SDVO device based on the sdvo info parsed from VBTyakui_zhao2009-06-051-3/+40
| | | | | | | | http://bugs.freedesktop.org/show_bug.cgi?id=20429 Signed-off-by: Zhao Yakui <yakui.zhao@intel.com> [anholt: Massive cleanup of the slave addr function] Signed-off-by: Eric Anholt <eric@anholt.net>
* drm/i915: parse VBT general definition block to get the SDVO device infoyakui_zhao2009-06-052-1/+93
| | | | | | | | | | | | | | | The general definition block contains the child device tables, which include the SDVO device info. For example: device slave address, device dvo port, device type. We will get the info of SDVO device by parsing the general definition blocks. Only when a valid slave address is found, it is regarded as the SDVO device. And the info of DVO port and slave address is recorded. http://bugs.freedesktop.org/show_bug.cgi?id=20429 Signed-off-by: Zhao Yakui <yakui.zhao@intel.com> Signed-off-by: Eric Anholt <eric@anholt.net>
* drm/i915: Add the structure of child_device_config in video BIOS tables.yakui_zhao2009-06-051-8/+93
| | | | | Signed-off-by: Zhao Yakui <yakui.zhao@intel.com> Signed-off-by: Eric Anholt <eric@anholt.net>
* drm/i915: Add Display Port register definesKeith Packard2009-06-051-0/+163
| | | | | | | | | This adds the register definitions for the display port enable register along with those for the GMCH and Link M/N ratios required to drive display port outputs. Signed-off-by: Keith Packard <keithp@keithp.com> Signed-off-by: Eric Anholt <eric@anholt.net>
* drm/i915: Don't trim cursor addresses to 11 bitsKeith Packard2009-06-051-4/+4
| | | | | | | | | We can safely assume that cursor addresses will not extend beyond the addressable screen dimensions; setting the additional bits is harmless in any case. Signed-off-by: Keith Packard <keithp@keithp.com> Signed-off-by: Eric Anholt <eric@anholt.net>
* drm/i915: Set correct TV detection voltage level override valuesMa Ling2009-06-052-3/+6
| | | | | | | | | | | We detect TV connect status by setting DAC voltage level override values as 0.7 voltage for DAC_A/B/C. The corresponding 2-bits shold be 0x2, In order correctly to set last bit as 0, at first we must clean it. It fixed freedesktop.org bug #21204 Signed-off-by: Ma Ling <ling.ma@intel.com> Signed-off-by: Eric Anholt <eric@anholt.net>
* drm/i915: Add LVDS support for IGDNGZhenyu Wang2009-06-052-26/+114
| | | | | Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com> Signed-off-by: Eric Anholt <eric@anholt.net>
* drm/i915: Add HDMI support on IGDNGZhenyu Wang2009-06-052-3/+47
| | | | | Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com> Signed-off-by: Eric Anholt <eric@anholt.net>
* drm/i915: enable kernel modesetting on IGDNGZhenyu Wang2009-06-053-38/+643
| | | | | | | | This adds kernel mode setting on IGDNG with VGA output support. Note that suspend/resume doesn't work yet. Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com> Signed-off-by: Eric Anholt <eric@anholt.net>
* drm/i915: Disable tiling on IGDNG for nowZhenyu Wang2009-06-051-0/+7
| | | | | | | Swizzle bit detection not working right on it. Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com> Signed-off-by: Eric Anholt <eric@anholt.net>
* drm/i915: Disable opregion on IGDNG for nowZhenyu Wang2009-06-052-3/+8
| | | | | | | Disable OpRegion support for now until verified on new chipsets. Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com> Signed-off-by: Eric Anholt <eric@anholt.net>
* drm/i915: Add new chipset register definitionsZhenyu Wang2009-06-051-0/+447
| | | | | Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com> Signed-off-by: Eric Anholt <eric@anholt.net>
* drm/i915: Add chipset/feature defines for for new chipsetsZhenyu Wang2009-06-051-5/+13
| | | | | | Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com> [anholt: dropped drm_pciids.h hunk to avoid loading an incomplete driver] Signed-off-by: Eric Anholt <eric@anholt.net>
* drm/i915: Change GEM throttling to be 20ms like the comment says.Eric Anholt2009-06-043-19/+61
| | | | | | | | | | | | | | | | | | | keithp didn't like the original 20ms plan because a cooperative client could be starved by an uncooperative client. There may even have been problems with cooperative clients versus cooperative clients. So keithp changed throttle to just wait for the second to last seqno emitted by that client. It worked well, until we started getting more round-trips to the server due to DRI2 -- the server throttles in BlockHandler, and so if you did more than one round trip after finishing your frame, you'd end up unintentionally syncing to the swap. Fix this by keeping track of the client's requests, so the client can wait when it has an outstanding request over 20ms old. This should have non-starving behavior, good behavior in the presence of restarts, and less waiting. Improves high-settings openarena performance on my GM45 by 50%. Signed-off-by: Eric Anholt <eric@anholt.net> Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
* drm/i915: Save/restore cursor state on suspend/resume.Eric Anholt2009-06-042-0/+28
| | | | | | | | | | This may fix cursor corruption in X on resume, which would persist until the cursor was hidden and then shown again. V2: Also include the cursor control regs. Signed-off-by: Eric Anholt <eric@anholt.net> Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
* drm/i915: Remove a bad BUG_ON in the fence management code.Eric Anholt2009-06-041-3/+0
| | | | | | | | | | This could be triggered by a gtt mapping fault on 965 that decides to remove the fence from another object that happens to be active currently. Since the other object doesn't rely on the fence reg for its execution, we don't wait for it to finish. We'll soon be not waiting on 915 most of the time as well, so just drop the BUG_ON. Signed-off-by: Eric Anholt <eric@anholt.net>
* ACPI, i915: build fix (v2)Len Brown2009-05-301-7/+7
| | | | | | | | | | drivers/built-in.o: In function `intel_opregion_init': (.text+0x9d540): undefined reference to `acpi_video_register' v2: move under DRM_I915 from DRM_I915_KMS Signed-off-by: Len Brown <len.brown@intel.com> Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
* i915: Set object to gtt domain when faulting it back inKristian Høgsberg2009-05-271-0/+7
| | | | | | | | | | | | | | | | | | | | | | When a GEM object is evicted from the GTT we set it to the CPU domain, as it might get swapped in and out or ever mmapped regularly. If the object is mmapped through the GTT it can still get evicted in this way by other objects requiring GTT space. When the GTT mapping is touched again we fault it back into the GTT, but fail to set it back to the GTT domain. This means we fail to flush any cached CPU writes to the pages backing the object which will then happen "eventually", typically after we write to the page through the uncached GTT mapping. [anholt: Note that userland does do a set_domain(GTT, GTT) when starting to access the GTT mapping. That covers getting the existing mapping of the object synchronized if it's bound to the GTT. But set_domain(GTT, GTT) doesn't do anything if the object is currently unbound. This fix covers the transition to being bound for GTT mapping.] Fixes glyph and other pixmap corruption during swapping. fd.o bug #21790 Signed-off-by: Kristian Høgsberg <krh@redhat.com> Signed-off-by: Eric Anholt <eric@anholt.net>
* drm/i915: Apply a big hammer to 865 GEM object CPU cache flushing.Eric Anholt2009-05-271-0/+10
| | | | | | | | | | | | | | On the 865, but not the 855, the clflush we do appears to not actually make it out to the hardware all the time. An easy way to safely reproduce was X -retro, which would show that some of the blits involved in drawing the lovely root weave didn't make it out to the hardware. Those blits are 32 bytes each, and 1-2 would be missing at various points around the screen. Other experimentation (doing more clflush, doing more AGP chipset flush, poking at some more device registers to maybe trigger more flushing) didn't help. krh came up with the wbinvd as a way to successfully get all those blits to appear. Signed-off-by: Eric Anholt <eric@anholt.net>
* drm/i915: Fix tiling pitch handling on 8xx.Eric Anholt2009-05-273-6/+17
| | | | | | | | | | | | | The pitch field is an exponent on pre-965, so we were rejecting buffers on 8xx that we shouldn't have. 915 got lucky in that the largest legal value happened to match (8KB / 512 = 0x10), but 8xx has a smaller tile width. Additionally, we programmed that bad value into the register on 8xx, so the only pitch that would work correctly was 4096 (512-1023 pixels), while others would probably give bad rendering or hangs. Signed-off-by: Eric Anholt <eric@anholt.net> fd.o bug #20473.
* Merge branch 'for-linus' of ↵Linus Torvalds2009-05-269-89/+408
|\ | | | | | | | | | | | | | | | | | | | | | | | | git://git.kernel.org/pub/scm/linux/kernel/git/anholt/drm-intel * 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/anholt/drm-intel: drm/i915: Add support for VGA load detection (pre-945). drm/i915: Use an I2C algo to do the flip to SDVO DDC bus. drm/i915: Determine type before initialising connector drm/i915: Return SDVO LVDS VBT mode if no EDID modes are detected. drm/i915: Fetch SDVO LVDS mode lines from VBT, then reserve them i915: support 8xx desktop cursors drm/i915: allocate large pointer arrays with vmalloc
| * drm/i915: Add support for VGA load detection (pre-945).Ma Ling2009-05-261-2/+147
| | | | | | | | | | | | | | | | | | | | Two approaches for VGA detections: hot plug detection for 945G onwards and load pipe detection for Pre-945G. Load pipe detection will get one free pipe, set border color as red and blue, then check CRT status by swf register. This is a sync-up with the 2D driver. Signed-off-by: Ma Ling <ling.ma@intel.com> Signed-off-by: Eric Anholt <eric@anholt.net>
| * drm/i915: Use an I2C algo to do the flip to SDVO DDC bus.Ma Ling2009-05-221-4/+63
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Previously, we would set the control bus switch before calls were made to request EDID information over DDC. But recently the DDC code started doing multiple I2C transfers to get the EDID extensions as well. This tripped up SDVO, because the control bus switch is only in effect until the next STOP after a START. By doing our own algo, we can wrap each i2c transaction on the DDC I2C bus with the control bus switch it requires. freedesktop.org bug #21042 Signed-off-by: Ma Ling <ling.ma@intel.com> [anholt: Hand application for conflict, fixed error path] Signed-off-by: Eric Anholt <eric@anholt.net>
| * drm/i915: Determine type before initialising connectorJonas Bonn2009-05-221-19/+10
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | drm_connector_init sets both the connector type and the connector type_id on the newly initialised connector. As the connector type_id is coupled to the connector type, the connector type cannot simply be modified on an initialised connector. This patch changes the order of operations on intel_sdvo_init so that the type is determined before the connector is intialised. This fixes a bug whereby the name card0-VGA-1 would be allocted to both a CRT and an SDVO connector since the SDVO connector would be initialised with type 'unknown' and hence have its type_id assigned from the wrong pool. Signed-off-by: Jonas Bonn <jonas@southpole.se> Signed-off-by: Eric Anholt <eric@anholt.net>
| * drm/i915: Return SDVO LVDS VBT mode if no EDID modes are detected.Ma Ling2009-05-221-0/+41
| | | | | | | | | | | | | | | | Some new SDVO LVDS hardware doesn't have DDC available, and this should fix the display on it. Signed-off-by: Ma Ling <ling.ma@intel.com> Signed-off-by: Eric Anholt <eric@anholt.net>
| * drm/i915: Fetch SDVO LVDS mode lines from VBT, then reserve themMa Ling2009-05-224-33/+93
| | | | | | | | | | Signed-off-by: Ma Ling <ling.ma@intel.com> Signed-off-by: Eric Anholt <eric@anholt.net>
| * i915: support 8xx desktop cursorsJesse Barnes2009-05-223-8/+39
| | | | | | | | | | | | | | | | | | | | For some reason we never added 8xx desktop cursor support to the kernel. This patch fixes that. [krh: Also set the size on pre-i915 hw.] Tested-by: Kristian Høgsberg <krh@redhat.com> Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org> Signed-off-by: Eric Anholt <eric@anholt.net>
| * drm/i915: allocate large pointer arrays with vmallocJesse Barnes2009-05-191-23/+15
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | For awhile now, many of the GEM code paths have allocated page or object arrays with the slab allocator. This is nice and fast, but won't work well if memory is fragmented, since the slab allocator works with physically contiguous memory (i.e. order > 2 allocations are likely to fail fairly early after booting and doing some work). This patch works around the issue by falling back to vmalloc for >PAGE_SIZE allocations. This is ugly, but much less work than chaining a bunch of pages together by hand (suprisingly there's not a bunch of generic kernel helpers for this yet afaik). vmalloc space is somewhat precious on 32 bit kernels, but our allocations shouldn't be big enough to cause problems, though they're routinely more than a page. Note that this patch doesn't address the unchecked alloc-based-on-ioctl-args in GEM; that needs to be fixed in a separate patch. Also, I've deliberately ignored the DRM's "area" junk. I don't think anyone actually uses it anymore and I'm hoping it gets ripped out soon. [Updated: removed size arg to new free function. We could unify the free functions as well once the DRM mem tracking is ripped out.] fd.o bug #20152 (part 1/3) Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org> Signed-off-by: Eric Anholt <eric@anholt.net>
* | drm: Copy back ioctl data to userspace regardless of return code.Michel Dänzer2009-05-201-1/+1
| | | | | | | | | | | | | | | | | | | | | | Fixes a regression from commit 9d5b3ffc42f7820e8ee07705496955e4c2c38dd9 ('drm: fixup some of the ioctl function exit paths'): The vblank ioctl needs to update the userspace parameters when interrupted by a signal, which was prevented by the return code check. This could cause the X server to hang in drmWaitVBlank(). Signed-off-by: Michel Dänzer <daenzer@vmware.com> Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
* | drm: Round size of SHM maps to PAGE_SIZEBenjamin Herrenschmidt2009-05-201-0/+8
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Currently, userspace can fail to obtain the SAREA mapping (among other reasons) if it passes SAREA_MAX to drmAddMap without aligning it to the page size. This breaks for example on PowerPC with 64K pages and radeon despite the kernel radeon actually doing the right rouding in the first place. The way SAREA_MAX is defined with a bunch of ifdef's and duplicated between libdrm and the X server is gross, ultimately it should be retrieved by userspace from the kernel, but in the meantime, we have plenty of existing userspace built with bad values that need to work. This patch works around broken userspace by rounding the requested size in drm_addmap_core() of any SHM map to the page size. Since the backing memory for SHM maps is also allocated within addmap_core, there is no danger of adjacent memory being exposed due to the increased map size. The only side effect is that drivers that previously tried to create or access SHM maps using a size < PAGE_SIZE and failed (getting -EINVAL), will now succeed at the cost of a little bit more memory used if that happens to be when the map is created. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
| |
| \
| \
| \
*---. \ Merge branches 'release', 'bugzilla-13032', 'bugzilla-13041+', ↵Len Brown2009-05-161-0/+6
|\ \ \ \ | |_|_|/ |/| | | | | | | 'bugzilla-13121', 'bugzilla-13165', 'bugzilla-13243', 'bugzilla-13259', 'resume-sci-en-regression', 'thermal-regression', 'tsc-regression' and 'asus-2.6.30' into release
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-16 13:15:01 +0100