Commit Graph

43 Commits (303cf6bbf3522cfba19a5950198183b75d195beb)

Author SHA1 Message Date
Michel Dänzer 1ec0df8a25 amdgpu: Rename fd_mutex/list to dev_mutex/list
Seems to better reflect what they're for.

Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
2019-07-01 11:32:05 +02:00
Lucas De Marchi e15e3a65b8 amdgpu: annotate public functions
This was done with:
nm --dynamic --defined-only build/amdgpu/libdrm_amdgpu.so | \
	grep amdgpu_ | \
	cut -d' ' -f3 > /tmp/a.txt

while read sym; do
	read f func line _ <<<$(cscope -d -L -1 $sym)
	if [ ! -z "$f" ]; then
		line=$((line-1))
		sed -i "${line}s/^/drm_public /" $f
	fi
done < /tmp/a.txt

Then the alignment of function arguments were manually fixed all over.
The idea here will be to switch the default visibility to hidden so we
don't export symbols we shouldn't.

Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Acked-by: Eric Engestrom <eric.engestrom@intel.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2018-09-19 22:46:45 -07:00
Christian König 9a38e850a5 amdgpu: use handle table for flink names
Instead of the hash use the handle table.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-and-Tested-by: Junwei Zhang <Jerry.Zhang@amd.com>
2018-08-07 13:27:28 +02:00
Christian König bde850bc32 amdgpu: use handle table for KMS handles
Instead of the hash use the handle table.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-and-Tested-by: Junwei Zhang <Jerry.Zhang@amd.com>
2018-08-07 13:27:22 +02:00
Christian König 7aa1a51133 amdgpu: stop using the hash table for fd_tab
We have so few devices that just walking a linked list is probably
faster.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-and-Tested-by: Junwei Zhang <Jerry.Zhang@amd.com>
2018-08-07 13:26:59 +02:00
Christian König 4fa33e45e6 amdgpu: make sure to set CLOEXEC on duplicated FDs
Otherwise we leak file descriptors into child processes.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-and-Tested-by: Junwei Zhang <Jerry.Zhang@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2018-07-17 16:48:08 +02:00
Jan Vesely c1f2d9b900 amdgpu: Destroy fd_hash table when the last device is removed.
Fixes memory leak on module unload.
Analogous to mesa commit of the same name.
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2018-05-25 12:12:00 -04:00
Jan Vesely 712fa0f3f4 amdgpu: Take a lock before removing devices from fd_tab hash table.
Close the file descriptors under lock as well.
v2: close fds after removing from hash table

Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2018-05-25 12:12:00 -04:00
Michel Dänzer bc9c789073 amdgpu: Deinitialize vamgr_high{,_32}
Fixes memory leaks.

Reviewed-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
2018-04-27 16:42:26 +02:00
Eric Engestrom 0926f0af54 meson,configure: include config.h automatically
This will prevent any more missing `#include "config.h"` bug, at the
cost of having to recompile some files that didn't need to be when
changing build options.

Signed-off-by: Eric Engestrom <eric.engestrom@imgtec.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
2018-03-20 18:19:26 +00:00
Christian König bca585b02f amdgpu: fix "add AMDGPU_VA_RANGE_HIGH"
The range is stored as exclusive, not inclusive. Subtracts one to get
the inclusive interval for the calculation. This fixes crashes when 32bit
addresses are in use.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-and-Tested-by: Michel Dänzer <michel.daenzer@amd.com>
2018-02-28 20:04:32 +01:00
Christian König ff0da7b323 amdgpu: add AMDGPU_VA_RANGE_HIGH
Return high addresses if requested and available.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
2018-02-28 13:27:43 +01:00
Christian König cd8a80493b amdgpu: mostly revert "use the high VA range if possible v2"
This reverts commit 07ea20d5be.

Unfortunately it turned out that this change broke some corner cases in
Mesa.

Revert it for now, but keep the high range in separate VA managers.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
2018-02-28 13:27:38 +01:00
Marek Olšák ad5b702fec amdgpu: add amdgpu_query_sw_info for querying high bits of 32-bit address space
Reviewed-by: Christian König <christian.koenig@amd.com>
2018-02-09 20:30:16 +01:00
Christian König 1cc17744b9 amdgpu: fix high VA mask
That constant needs to be 64bits.

Fixes: amdgpu: use the high VA range if possible v2

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2018-01-23 11:17:43 +01:00
Christian König 07ea20d5be amdgpu: use the high VA range if possible v2
Retire the low range on Vega10 this frees up everything below 0xffff800000000000 for HMM.

v2: keep the 32bit range working.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
2018-01-09 17:15:09 +01:00
Christian König a8e436de5d amdgpu: fix 32bit VA manager max address
The range is exclusive not inclusive.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
2018-01-09 17:15:02 +01:00
Michel Dänzer f05a2b4cb1 amdgpu: Only remember the device's marketing name
There's no point in keeping around the full table of marketing names,
when amdgpu_get_marketing_name only ever returns the device's marketing
name.

Acked-by: Slava Abramov <slava.abramov@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2017-12-05 12:50:20 +01:00
Michel Dänzer 85c6b0b00a amdgpu: Clean up amdgpu_parse_asic_ids error handling
* Move error message printing into amdgpu_parse_asic_ids and make it
  return void
* Print only "Invalid format" error message if parse_one_line returns
  -EINVAL
* Use strerror instead of printing the (negative) error code in hex

Acked-by: Slava Abramov <slava.abramov@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2017-12-01 16:52:01 +01:00
Christian König 931f01964a Revert "amdgpu: fix 32bit VA manager max address"
This reverts commit 944f6665de.

Accidentially pushed an imcomplete patch.

Signed-off-by: Christian König <christian.koenig@amd.com>
2017-11-09 10:36:10 +01:00
Christian König 0c2481edf7 Revert "amdgpu: use the high VA range if possible"
This reverts commit 6c0ea4b0c5.

Accidentially pushed an incomplete patch.

Signed-off-by: Christian König <christian.koenig@amd.com>
2017-11-09 10:35:44 +01:00
Christian König 6c0ea4b0c5 amdgpu: use the high VA range if possible
This frees up the low range for HMM.

Signed-off-by: Christian König <christian.koenig@amd.com>
2017-11-08 16:12:56 +01:00
Christian König 944f6665de amdgpu: fix 32bit VA manager max address
The range is exclusive not inclusive.

Signed-off-by: Christian König <christian.koenig@amd.com>
2017-11-08 16:12:29 +01:00
Christian König a844925c8a amdgpu: sanitize 64bit VA manager
Adding the extra reservation of the 32bit space to the 64bit manager is
complete nonsense and just a waste of memory and CPU cycles.

Signed-off-by: Christian König <christian.koenig@amd.com>
2017-11-08 16:11:40 +01:00
Marek Olšák 1dd84e01a9 amdgpu: print error messages when amdgpu_device_initialize is failing
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2017-09-28 22:48:26 +02:00
Xiaojie Yuan 7e6bf88cac amdgpu: move asic id table to a separate file
v2: fix an off by one error and leading white spaces
v3: use thread safe strtok_r(); initialize len before calling getline();
    change printf() to drmMsg(); add initial amdgpu.ids
v4: integrate some recent internal changes, including format changes
v5: fix line number for empty/commented lines; realloc to save memory;
    indentation changes
v6: remove a line error
v7: [Michel Dänzer]
* Move amdgpu.ids to new data directory
* Remove placeholder entries from amdgpu.ids
* Set libdrmdatadir variable in configure.ac instead of Makefile.am
  [Emil Velikov]
* Use isblank() instead of open-coding it [Emil Velikov]
* Don't leak asic_id_table memory if realloc fails [Emil Velikov]
* Check and bump table_max_size at the beginning of the while loop [Emil
  Velikov]
* Initialize table_max_size to the number of entries in data/amdgpu.ids
v8: [Michel Dänzer]
* Make sure amdgpu_asic_id.c gets rebuilt when amdgpu.ids changes

Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
Signed-off-by: Samuel Li <Samuel.Li@amd.com>
2017-06-13 10:21:39 +09:00
Alex Xie fe7cb34eda amdgpu: vamgr can be a struct instead of a pointer
vamgr is an integral part of amdgpu_device. We don't need to calloc and free it.
This can save CPU time, reduce heap fragmentation.

Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
Signed-off-by: Alex Xie <AlexBin.Xie@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
[Grazvydas Ignotas: rebase, correct a typo in commit message]
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-02-02 15:22:53 -05:00
Alex Xie 067e9a1d47 amdgpu: vamgr_32 can be a struct instead of a pointer
vamgr_32 is an integral part of amdgpu_device. We don't need to calloc and free it.
This can save CPU time, reduce heap fragmentation.

Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
Signed-off-by: Alex Xie <AlexBin.Xie@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
[Grazvydas Ignotas: rebase, correct a typo in commit message]
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-02-02 15:22:45 -05:00
Alex Xie 7a03cdf6a7 amdgpu: Free/uninit vamgr_32 in theoretically correct order
vamgr_32 is a region inside general VAM range. It is better to free and
deinitialize it before general VAM range.

Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
Signed-off-by: Alex Xie <AlexBin.Xie@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-02-02 15:22:07 -05:00
Junwei Zhang 670f1e4fda amdgpu: add the function to get the marketing name (v4)
This function is used to look up the marking name
for a specific board.

v2: agd: Squash in subsequent updates to the table.
v3: [Michel Dänzer]
* Make amdgpu_asic_id_table static, so it's not exported from
  libdrm_amdgpu.so.1
* Add amdgpu_get_marketing_name to amdgpu-symbols-check
* Fix indentation of second line of if statement
* Squash in another change removing redundant entries
* Change spelling of "RADEON" -> "Radeon"
* Remove "(TM)" from a minority of entries
v4: [Michel Dänzer]
* Use const char* instead of fixed size array for marketing_name (Emil
  Velikov)

Signed-off-by: Junwei Zhang <Jerry.Zhang@amd.com>
Reviewed-by: Flora Cui <Flora.Cui@amd.com>
Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2016-11-07 17:44:27 +09:00
monk.liu f17017b8e2 amdgpu: fix missing deinit on vamgr_32
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: monk.liu <monk.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2015-09-02 09:52:12 -04:00
Jammy Zhou 56d8dd6a9c amdgpu: make vamgr per device v2
Each device can have its own vamgr, so make it per device now.
This can fix the failure with multiple GPUs used in one single
process.

v2: rebase

Signed-off-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
2015-08-17 16:29:26 -04:00
Jammy Zhou ffa305d0fc amdgpu: add flag to support 32bit VA address v4
The AMDGPU_VA_RANGE_32_BIT flag is added to request VA range in the
32bit address space for amdgpu_va_range_alloc.

The 32bit address space is reserved at initialization time, and managed
with a separate VAMGR as part of the global VAMGR. And if no enough VA
space available in range above 4GB, this reserved range can be used as
fallback.

v2: add comment for AMDGPU_VA_RANGE_32_BIT, and add vamgr to va_range
v3: rebase to Emil's drm_private series
v4: fix one warning

Signed-off-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
2015-08-17 16:26:26 -04:00
Emil Velikov bddf4df4a1 amdgpu: hide the final internal functions from global namespace
Thus the only symbols that we export are the ones officially provided by
the API.

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
2015-08-13 17:43:36 +01:00
Emil Velikov a30da8e9b3 amdgpu: cosmetic chances in license boilerplate
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
2015-08-13 17:43:27 +01:00
Emil Velikov f4c2bfd63e amdgpu: add a bunch of missing config.h includes
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
2015-08-13 17:43:24 +01:00
Mario Kleiner 7d418f906c libdrm/amdgpu: Use private fd for amdgpu_device and winsys hash table to fix ZaphodHeads. (v2)
The amdgpu_device for a device node needs its own dup'ed fd, instead
of using the original fd passed in for a screen, to make multi-x-screen
ZaphodHeads configurations work on amdgpu.

The original fd's lifetime differs from that of the amdgpu_device, and from the
one stored in the hash. The hash key is the fd, and in order to compare hash
entries we fstat them, so the fd must be around for as long as the amdgpu_device
is.

This patch for libdrm/amdgpu is a translation of the radeon-winsys ZaphodHeads
fix for mesa's radeon-winsys, from mesa commit 28dda47ae4d974e3e032d60e8e0965c8c068c6d8

"winsys/radeon: Use dup fd as key in drm-winsys hash table to fix ZaphodHeads."

Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com>
Acked-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>

v2: Check for valid fd's being >= 0, because fd == 0 is in theory
    a valid, although unlikely, fd and fd == -1 would denote an
    invalid fd. Thanks to William Lewis for pointing this out.

Reported-by: William Lewis <minutemaidpark@hotmail.com>
Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com>
2015-08-05 13:47:52 -04:00
Marek Olšák 7d7f25c938 amdgpu: fix valgrind warnings
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-08-05 13:47:51 -04:00
Ken Wang 322d02d025 amdgpu: make vamgr global
This is the first sub-patch of va interface task, the va task is
about adding more va management interfaces for UMD, by design, the
vamgr should be per-process rather than per-device.

Signed-off-by: Ken Wang <Qingqing.Wang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
2015-08-05 13:47:50 -04:00
Christian König d5c0b2a172 amdgpu: remove bo_vas hash table v2
Not used any more.

v2: Keep accidental removed lines.

Signed-off-by: Christian König <christian.koenig@amd.com>
2015-08-05 13:47:50 -04:00
Marek Olšák 2d588797f1 amdgpu: fix a valgrind warning
==10307== Conditional jump or move depends on uninitialised value(s)
==10307==    at 0xB9F4CFA: amdgpu_device_initialize (amdgpu_device.c:193)

Reviewed-by: Jammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-08-05 13:47:49 -04:00
Christian König c68d58aa14 amdgpu: compare the primary device names instead
Instead of taking a look at the device major/minor we
just compare the primary device name to figure out if
two fds are pointing to the same device.

Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Christian König <christian.koenig@amd.com>
2015-08-05 13:47:49 -04:00
Alex Deucher 0936139536 drm: add libdrm_amdgpu (v7)
This is the new ioctl wrapper used by the new admgpu driver.
It's primarily used by xf86-video-amdgpu and mesa.

v2: fix amdgpu_drm.h install
v3: Integrate some of the sugestions from Emil:
    clean up Makefile.am, configure.ac
    capitalize header guards
    fix _FILE_OFFSET_BITS with config.h
    use drm_mmap/drm_munmap
    Remove unused ARRAY_SIZE macro
    use shared list implementation
    use shared math implementation
    use drmGetNodeTypeFromFd helper
v4: remove unused tiling defines
v5: include amdgpu.h in Makefile.am
v6: update amdgpu_drm.h
v7: libdrm.h -> libdrm_macros.h

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2015-08-05 13:47:48 -04:00