aboutsummaryrefslogtreecommitdiff
path: root/drivers/gpu
AgeCommit message (Collapse)AuthorFilesLines
2013-08-30drm/radeon: gcc fixes for sumo dpmAlex Deucher1-2/+6
Newer versions of gcc seem to wander off into the weeds when dealing with variable sizes arrays in structs. Rather than indexing the arrays, use pointer arithmetic. See bugs: https://bugs.freedesktop.org/show_bug.cgi?id=66932 https://bugs.freedesktop.org/show_bug.cgi?id=66972 https://bugs.freedesktop.org/show_bug.cgi?id=66945 Signed-off-by: Alex Deucher <[email protected]>
2013-08-30drm/radeonn: gcc fixes for rv7xx/eg/btc dpmAlex Deucher1-2/+3
Newer versions of gcc seem to wander off into the weeds when dealing with variable sizes arrays in structs. Rather than indexing the arrays, use pointer arithmetic. See bugs: https://bugs.freedesktop.org/show_bug.cgi?id=66932 https://bugs.freedesktop.org/show_bug.cgi?id=66972 https://bugs.freedesktop.org/show_bug.cgi?id=66945 Signed-off-by: Alex Deucher <[email protected]>
2013-08-30drm/radeon: gcc fixes for rv6xx dpmAlex Deucher1-2/+3
Newer versions of gcc seem to wander off into the weeds when dealing with variable sizes arrays in structs. Rather than indexing the arrays, use pointer arithmetic. See bugs: https://bugs.freedesktop.org/show_bug.cgi?id=66932 https://bugs.freedesktop.org/show_bug.cgi?id=66972 https://bugs.freedesktop.org/show_bug.cgi?id=66945 Signed-off-by: Alex Deucher <[email protected]>
2013-08-30drm/radeon: gcc fixes for radeon_atombios.cAlex Deucher1-14/+28
Newer versions of gcc seem to wander off into the weeds when dealing with variable sizes arrays in structs. Rather than indexing the arrays, use pointer arithmetic. See bugs: https://bugs.freedesktop.org/show_bug.cgi?id=66932 https://bugs.freedesktop.org/show_bug.cgi?id=66972 https://bugs.freedesktop.org/show_bug.cgi?id=66945 Signed-off-by: Alex Deucher <[email protected]>
2013-08-30drm/radeon: enable UVD interrupts on CIKChristian König1-0/+4
The same as on evergreen. Signed-off-by: Christian König <[email protected]> Reported-by: FrankR Huang <[email protected]> Signed-off-by: Alex Deucher <[email protected]> Cc: [email protected]
2013-08-30drm/radeon: fix init ordering for r600+Alex Deucher6-24/+30
The vram scratch buffer needs to be initialized before the mc is programmed otherwise we program 0 as the GPU address of the default GPU fault page. In most cases we put vram at zero anyway and reserve a page for the legacy vga buffer so in practice this shouldn't cause any problems, but better to make it correct. Was changed in: 6fab3febf6d949b0a12b1e4e73db38e4a177a79e Reported-by: FrankR Huang <[email protected]> Signed-off-by: Alex Deucher <[email protected]> Cc: [email protected]
2013-08-30drm/radeon/dpm: only need to reprogram uvd if uvd pg is enabledAlex Deucher1-6/+9
Avoid needless uvd reprogramming if uvd powergating is disabled. Signed-off-by: Alex Deucher <[email protected]> Reviewed-by: Christian König <[email protected]>
2013-08-30drm/radeon: check the return value of uvd_v1_0_start in uvd_v1_0_initAlex Deucher1-1/+3
No need to try the ring tests if starting the UVD block failed. Signed-off-by: Alex Deucher <[email protected]> Reviewed-by: Christian König <[email protected]>
2013-08-30drm/radeon: split out radeon_uvd_resume from uvd_v4_2_resumeAlex Deucher2-10/+8
For powergating, we just need to re-init the registers, there is no need to restore the uvd BOs. This just adds needless work when powergating uvd for playback while the system is on. We only need to restore the uvd BOs on an actual resume from suspend or when the driver loads. This fixes multi-stream UVD playback on KB systems. Signed-off-by: Alex Deucher <[email protected]> Reviewed-by: Christian König <[email protected]>
2013-08-30radeon kms: fix uninitialised hotplug work usage in r100_irq_process()Sergey Senozhatsky1-4/+6
Commit a01c34f72e7cd2624570818f579b5ab464f93de2 (radeon kms: do not flush uninitialized hotplug work) moved work initialisation phase to the last step of radeon_irq_kms_init(). Meelis Roos reported that this causes problems on his machine because drm_irq_install() uses hotplug work on r100. hotplug work flushed in radeon_irq_kms_fini(), with two possible cases: -- radeon_irq_kms_fini() call after successful radeon_irq_kms_init() -- radeon_irq_kms_fini() call after unsuccessful (or not called at all) radeon_irq_kms_init() The latter one causes flush work on uninitialised hotplug work. Move work initialisation before drm_irq_install(), but keep existing agreement to flush hotplug work in radeon_irq_kms_fini() only for `irq.installed' (successful radeon_irq_kms_init()) case. WARNING: CPU: 0 PID: 243 at kernel/workqueue.c:1378 __queue_work+0x132/0x16d() Call Trace: [<c12319b3>] ? dump_stack+0xa/0x13 [<c1022600>] ? warn_slowpath_common+0x75/0x8a [<c1031010>] ? __queue_work+0x132/0x16d [<c1031010>] ? __queue_work+0x132/0x16d [<c102269e>] ? warn_slowpath_null+0x1b/0x1f [<c1031010>] ? __queue_work+0x132/0x16d [<c103107b>] ? queue_work_on+0x30/0x40 [<f8aed3f3>] ? r100_irq_process+0x16d/0x1e6 [radeon] [<f8ae77cf>] ? radeon_driver_irq_preinstall_kms+0xc2/0xc5 [radeon] [<f8974d77>] ? drm_irq_install+0xb2/0x1ac [drm] [<f897604d>] ? drm_vblank_init+0x196/0x1d2 [drm] [<f8ae78d3>] ? radeon_irq_kms_init+0x33/0xc6 [radeon] [<f8aef35a>] ? r100_startup+0x1a3/0x1d6 [radeon] [<f8ad77c8>] ? radeon_ttm_init+0x26e/0x287 [radeon] [<f8aef752>] ? r100_init+0x2b3/0x309 [radeon] [<c118082e>] ? vga_client_register+0x39/0x40 [<f8ac535f>] ? radeon_device_init+0x54b/0x61b [radeon] [<f8ac40fd>] ? cail_mc_write+0x13/0x13 [radeon] [<f8ac6864>] ? radeon_driver_load_kms+0x82/0xda [radeon] [<f8978bbd>] ? drm_get_pci_dev+0x136/0x22d [drm] [<f8ac409b>] ? radeon_pci_probe+0x6c/0x86 [radeon] [<c112acf6>] ? pci_device_probe+0x4c/0x83 [<c11846c7>] ? driver_probe_device+0x80/0x184 [<c112a848>] ? pci_match_id+0x18/0x36 [<c1184837>] ? __driver_attach+0x44/0x5f [<c11833f4>] ? bus_for_each_dev+0x50/0x5a [<c118433e>] ? driver_attach+0x14/0x16 [<c11847f3>] ? __device_attach+0x28/0x28 [<c1184045>] ? bus_add_driver+0xd6/0x1bf [<c1184c22>] ? driver_register+0x78/0xcf [<f8ba8000>] ? 0xf8ba7fff [<c10003bf>] ? do_one_initcall+0x8b/0x121 [<c101e668>] ? change_page_attr_clear+0x2e/0x33 [<f8ba8000>] ? 0xf8ba7fff [<c101e689>] ? set_memory_ro+0x1c/0x20 [<c104de94>] ? set_page_attributes+0x11/0x12 [<c104f6e1>] ? load_module+0x12fa/0x17e8 [<c107483b>] ? map_vm_area+0x22/0x31 [<c104fc36>] ? SyS_init_module+0x67/0x7d [<c1234245>] ? sysenter_do_call+0x12/0x26 Reported-by: Meelis Roos <[email protected]> Tested-by: Meelis Roos <[email protected]> Signed-off-by: Sergey Senozhatsky <[email protected]> Signed-off-by: Alex Deucher <[email protected]> Cc: [email protected]
2013-08-30drm/radeon/audio: set up the sads on DCE3.2 asicsAlex Deucher2-1/+94
This sets up the short audio descriptors properly on DCE3.2 asics for hdmi audio. Signed-off-by: Alex Deucher <[email protected]>
2013-08-30drm/radeon: fix handling of variable sized arrays for router objectsAlex Deucher1-3/+6
The table has the following format: typedef struct _ATOM_SRC_DST_TABLE_FOR_ONE_OBJECT //usSrcDstTableOffset pointing to this structure { UCHAR ucNumberOfSrc; USHORT usSrcObjectID[1]; UCHAR ucNumberOfDst; USHORT usDstObjectID[1]; }ATOM_SRC_DST_TABLE_FOR_ONE_OBJECT; usSrcObjectID[] and usDstObjectID[] are variably sized, so we can't access them directly. Use pointers and update the offset appropriately when accessing the Dst members. Signed-off-by: Alex Deucher <[email protected]> Cc: [email protected]
2013-08-30drm/radeon: fix resume on some rs4xx boards (v2)Alex Deucher1-3/+6
Setting MC_MISC_CNTL.GART_INDEX_REG_EN causes hangs on some boards on resume. The systems seem to work fine without touching this bit so leave it as is. v2: read-modify-write the GART_INDEX_REG_EN bit. I suspect the problem is that we are losing the other settings in the register. fixes: https://bugs.freedesktop.org/show_bug.cgi?id=52952 Reported-by: Ondrej Zary <[email protected]> Tested-by: Daniel Tobias <[email protected]> Signed-off-by: Alex Deucher <[email protected]> Cc: [email protected]
2013-08-30drm/radeon: fill in gpu_init for berlin GPU coresAlex Deucher1-1/+40
This fills in the GPU specific details for berlin GPU cores so that the driver will work with them. Signed-off-by: Alex Deucher <[email protected]> Cc: [email protected]
2013-08-30drm/radeon/dpm: ungate blocks in dpm disable for kb/kvAlex Deucher1-0/+6
These blocks need to be ungated for the other parts of the driver properly initialize them (e.g., after a gpu reset, etc.). Signed-off-by: Alex Deucher <[email protected]>
2013-08-30drm/radeon/dpm: track uvd gated state for ciAlex Deucher2-0/+14
Track the current uvd gated state on CI to avoid unnecessary state changes when uvd is active. Signed-off-by: Alex Deucher <[email protected]>
2013-08-30drm/radeon: enable uvd dpm on CIAlex Deucher1-0/+2
UVD dpm dynamically adjusts the uvd clocks on demand. Signed-off-by: Alex Deucher <[email protected]>
2013-08-30drm/radeon: disable the GRPH block when we disable the crtcAlex Deucher1-0/+6
Since we aren't using it when the crtc is disabled, turn it off to save power. The GRPH block is the part of the display controller that controls the primary graphics plane (size, address, etc.). Signed-off-by: Alex Deucher <[email protected]>
2013-08-30drm/radeon: atombios hw i2c fixesAlex Deucher1-4/+8
These fixes make writes work properly. Previously only reads worked. Note that this feature is off by default. Signed-off-by: Alex Deucher <[email protected]>
2013-08-30drm/radeon: fix LCD record parsingAlex Deucher1-1/+3
If the LCD table contains an EDID record, properly account for the edid size when walking through the records. This should fix error messages about unknown LCD records. Signed-off-by: Alex Deucher <[email protected]> Cc: [email protected]
2013-08-30drm/radeon: check firmware overrides for mclk/sclk ssAlex Deucher1-0/+9
Check the overrides in the firmware info table before enabling spread spectrum on the engine or memory clocks. Some boards may have valid spread spectrum tables, but shouldn't necessarily have it enabled. Signed-off-by: Alex Deucher <[email protected]>
2013-08-30drm/radeon: update line buffer allocation for dce8Alex Deucher2-7/+26
We need to allocate line buffer to each display when setting up the watermarks. Failure to do so can lead to a blank screen. This fixes blank screen problems on dce8 asics. Based on an initial fix from: Jay Cornwall <[email protected]> Signed-off-by: Alex Deucher <[email protected]> Cc: [email protected]
2013-08-30drm/radeon: update line buffer allocation for dce6Alex Deucher2-4/+23
We need to allocate line buffer to each display when setting up the watermarks. Failure to do so can lead to a blank screen. This fixes blank screen problems on dce6 asics. Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=64850 Based on an initial fix from: Jay Cornwall <[email protected]> Signed-off-by: Alex Deucher <[email protected]> Cc: [email protected]
2013-08-30drm/radeon: update line buffer allocation for dce4.1/5Alex Deucher2-4/+25
We need to allocate line buffer to each display when setting up the watermarks. Failure to do so can lead to a blank screen. This fixes blank screen problems on dce4.1/5 asics. Based on an initial fix from: Jay Cornwall <[email protected]> Signed-off-by: Alex Deucher <[email protected]> Cc: [email protected]
2013-08-30drm/radeon/si: Add support for CP DMA to CS checker for compute v2Tom Stellard2-45/+64
Also add a new RADEON_INFO query to check that CP DMA packets are supported on the compute ring. CP DMA has been supported since the 3.8 kernel, but due to an oversight we forgot to teach the CS checker that the CP DMA packet was legal for the compute ring on Southern Islands GPUs. This patch fixes a bug where the radeon driver will incorrectly reject a legal CP DMA packet from user space. I would like to have the patch backported to stable so that we don't have to require Mesa users to use a bleeding edge kernel in order to take advantage of this feature which is already present in the stable kernels (3.8 and newer). v2: - Don't bump kms version, so this patch can be backported to stable kernels. Cc: [email protected] Signed-off-by: Tom Stellard <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2013-08-30drm/radeon: enable mgcg on CIKAlex Deucher1-3/+3
Now that the CP is no longer reset and cg is properly disabled in when appropriate in the dpm code we can now enable mgcg (medium grained clockgating). Signed-off-by: Alex Deucher <[email protected]>
2013-08-30drm/radeon: handle cg in KB/KV dpm codeAlex Deucher1-0/+26
Clockgating needs to be disabled around certain parts of dpm setup otherwise the smc gets into a bad state and dpm doesn't work properly. Signed-off-by: Alex Deucher <[email protected]>
2013-08-30drm/radeon: handle cg in CI dpm codeAlex Deucher1-0/+36
Clockgating needs to be disabled around certain parts of dpm setup otherwise the smc gets into a bad state and dpm doesn't work properly. Signed-off-by: Alex Deucher <[email protected]>
2013-08-30drm/radeon/cik: properly set up the clearstate buffer for pg (v2)Alex Deucher2-3/+103
The format of the clearstate buffer used for pg (powergating) changed between NI and SI. This formats it properly for what the hardware expects on SI+. v2: fix addresses Signed-off-by: Alex Deucher <[email protected]>
2013-08-30drm/radeon: fixes for gfx clockgating on CIKAlex Deucher2-13/+8
Clockgating requires signalling between the CP and the RLC to work properly. Resetting the CP block in the CP resume code messed up the internal coordination between the blocks. Removing the reset allows gfx clockgating to work properly. However, when gfx clock gating is enabled, there is a strange interaction with dpm which causes the chip to stay in the high performance level all the time, so leave gfx clockgating disabled for now. Signed-off-by: Alex Deucher <[email protected]>
2013-08-30drm/radeon: restructure cg/pg on cik (v2)Alex Deucher3-69/+170
- use new cg/pg flags for finer grained clock and powergating control - restructure the cg/pg code so it can be called from other components such as dpm v2: fix build breakage from rebase Signed-off-by: Alex Deucher <[email protected]>
2013-08-30drm/radeon/si: enable DMA pg by defaultAlex Deucher2-15/+7
Enable DMA powergating by default. The DMA engines will be powergated when not in use. Signed-off-by: Alex Deucher <[email protected]>
2013-08-30drm/radeon/si: properly set up the clearstate buffer for pg (v2)Alex Deucher2-39/+145
The format of the clearstate buffer used for pg (powergating) changed between NI and SI. This formats it properly for what the hardware expects on SI. v2: fix addresses Signed-off-by: Alex Deucher <[email protected]>
2013-08-30drm/radeon: enable mgcg on SIAlex Deucher1-5/+5
Now that the CP is no longer reset and cg is properly disabled in when appropriate in the dpm code we can now enable mgcg (medium grained clockgating). Signed-off-by: Alex Deucher <[email protected]>
2013-08-30drm/radeon: handle cg in SI dpm codeAlex Deucher1-0/+38
Clockgating needs to be disabled around certain parts of dpm setup otherwise the smc gets into a bad state and dpm doesn't work properly. Signed-off-by: Alex Deucher <[email protected]>
2013-08-30drm/radeon: fixes for gfx clockgating on SIAlex Deucher2-18/+9
Clockgating requires signalling between the CP and the RLC to work properly. Resetting the CP block in the CP resume code messed up the internal coordination between the blocks. Removing the reset allows gfx clockgating to work properly. However, when gfx clock gating is enabled, there is a strange interaction with dpm which causes the chip to stay in the high performance level all the time, so leave gfx clockgating disabled for now. Signed-off-by: Alex Deucher <[email protected]>
2013-08-30drm/radeon/si: restructure cg code (v3)Alex Deucher4-34/+192
Resturcture clockgating code so that it can be enabled/disabled from other components such as dpm. v2: make function static v3: add fine grained cg controls Signed-off-by: Alex Deucher <[email protected]>
2013-08-30drm/radeon: use new cg/pg flags for SIAlex Deucher2-29/+128
Allows us finer grained control over clock and powergating on SI. Signed-off-by: Alex Deucher <[email protected]>
2013-08-30drm/radeon: add cg and pg flagsAlex Deucher1-0/+35
This commits adds flags for supported clockgating and powergating features. This allows us to more easily track which features are supported on a particular asic and to enable/disable features for debugging. Signed-off-by: Alex Deucher <[email protected]>
2013-08-30drm/radeon: set speaker allocation for DCE3.2Alex Deucher2-0/+49
This updates the audio driver to the speaker allocation block from the EDID. A similar change was just implemented for DCE4-8. Signed-off-by: Alex Deucher <[email protected]>
2013-08-30drm/radeon: set speaker allocation for DCE4/5 (v2)Alex Deucher2-1/+47
This updates the audio driver to the speaker allocation block from the EDID. A similar change was just implemented for DCE6/8. v2: remove unused variables Signed-off-by: Alex Deucher <[email protected]> Acked-by: Rafał Miłecki <[email protected]>
2013-08-30drm/radeon: set speakers allocation earlierRafał Miłecki2-22/+54
Do it before enabling audio channels (in AFMT_AUDIO_PACKET_CONTROL2 register). Signed-off-by: Rafał Miłecki <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2013-08-30drm/radeon: add audio support for DCE6/8 GPUs (v12)Alex Deucher15-62/+455
Similar to DCE4/5, but supports multiple audio pins which can be assigned per afmt block. v2: rework the driver to handle more than one audio pin. v3: try different dto reg v4: properly program dto v5 (ck): change dto programming order v6: program speaker allocation block v7: rebase v8: rebase on Rafał's changes v9: integrated Rafał's comments, update to latest drm_edid_to_speaker_allocation API v10: add missing line break in error message v11: add back audio enabled messages v12: fix copy paste typo in r600_audio_enable Signed-off-by: Alex Deucher <[email protected]> Signed-off-by: Christian König <[email protected]> Acked-by: Rafał Miłecki <[email protected]>
2013-08-30drm/radeon: use loop for initializing AFMT blocksRafał Miłecki1-30/+23
Signed-off-by: Rafał Miłecki <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2013-08-30drm/edid: add a helper function to extract the speaker allocation data block ↵Alex Deucher1-0/+52
(v3) This adds a helper function to extract the speaker allocation data block from the EDID. This data block describes what speakers are present on the display device. v2: update per Ville Syrjälä's comments v3: fix copy/paste typo in memory allocation Signed-off-by: Alex Deucher <[email protected]> Reviewed-by: Ville Syrjälä <[email protected]> Tested-by: Rafał Miłecki <[email protected]>
2013-08-30drm/radeon: separate DMA codeChristian König13-1899/+2181
Similar to separating the UVD code, just put the DMA functions into separate files. Signed-off-by: Christian König <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2013-08-30drm/radeon: separate UVD code v3Christian König15-555/+826
Our different hardware blocks are actually completely separated, so it doesn't make much sense any more to structure the code by pure chipset generations. Start restructuring the code by separating our the UVD block. v2: updated commit message v3: rebased and restructurized start/stop functions for kv dpm. Signed-off-by: Christian König <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2013-08-30drm/radeon: remove special handling for the DMA ringChristian König11-59/+95
Now that we have callbacks for [rw]ptr handling we can remove the special handling for the DMA rings and use the callbacks instead. Signed-off-by: Christian König <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2013-08-30drm/radeon: rework UVD writeback & [rw]ptr handlingChristian König10-31/+38
The hardware just doesn't support this correctly. Disable it before we accidentally write anywhere we shouldn't. Signed-off-by: Christian König <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2013-08-30drm/radeon: rework ring function handlingChristian König3-787/+313
Give the ring functions a separate structure and let the asic structure point to the ring specific functions. This simplifies the code and allows us to make changes at only one point. No change in functionality. Signed-off-by: Christian König <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
Powered by cgit, Gruvboxed by Blaster4385.