#
262988 |
|
10-Mar-2014 |
dumbbell |
MFC TTM, a memory manager used by video drivers
This is the last step before the merge of the Radeon KMS driver. Other changes to TTM will be merged with radeonkms.
The following revisions were merged in this single commit:
r247835: Import the preliminary port of the TTM.
The early commit is done to facilitate the off-tree work on the porting of the Radeon driver.
Sponsored by: The FreeBSD Foundation Debugged and tested by: dumbbell MFC after: 1 month
r247848: Fix build with gcc, remove redundand declarations.
Reported and tested by: gjb MFC after: 1 month
r247849: Fix build with gcc, do not use unnamed union.
Reported and tested by: gjb MFC after: 1 month
r248060: drm: Fix a call to free(9) with an incorrect malloc type
While here, the call to free(9) is moved to drm_global.c, near the initial malloc(9).
Reviewed by: Konstantin Belousov (kib@)
r248657: drm/ttm: Fix TTM buffer object refcount
This fixes memory leaks in the radeonkms driver.
Reviewed by: Konstantin Belousov (kib@) Tested by: J.R. Oldroyd <jr@opal.com>
r248661: drm/ttm: Explain why we don't need to acquire a ref in ttm_bo_vm_ctor()
r248663: drm/ttm: Fix a typo: s/pTTM]/[TTM]/
r248666: Do not call malloc(M_WAITOK) while bodev->fence_lock mutex is held. The ttm_buffer_object_transfer() does not need the mutex locked at all, except for the call to the driver sync_obj_ref() method.
Reported and tested by: dumbbell MFC after: 2 weeks
r252864: Remove unneeded page lock around vm_page_insert().
Submitted by: alc
r253710: MFgem r251960: re-check the mgt device object for the requested page after the object was relocked.
Tested by: dumbbell Sponsored by: The FreeBSD Foundation MFC after: 1 week
r254822: drm: In drm_mmap_single, try ttm_bo_mmap_single() before drm_gem_mmap_single()
In drivers such as the Radeon driver, the DRIVER_GEM features flag is set but TTM is used to mmap buffer object.
r254858: drm: Add missing bits to drmP.h, required by the Radeon driver
Some of the FreeBSD-specific definitions are moved to drm_os_freebsd.h. But there's still work to do to clean it up and reduce the diff with Linux' drmP.h.
r254860: drm: Update drm_atomic.h, now that projects/atomic64 is in HEAD
Submitted by: jkim@
r254861: drm/ttm: Import Linux commit 63d0a4195560362e2e00a3ad38fc331d34e1da9b
Author: Maarten Lankhorst <maarten.lankhorst@canonical.com> Date: Tue Jan 15 14:56:37 2013 +0100
drm/ttm: remove lru_lock around ttm_bo_reserve
There should no longer be assumptions that reserve will always succeed with the lru lock held, so we can safely break the whole atomic reserve/lru thing. As a bonus this fixes most lockdep annotations for reservations.
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> Reviewed-by: Jerome Glisse <jglisse@redhat.com>
r254862: drm/ttm: Import Linux commit 7a1863084c9d90ce4b67d645bf9b0f1612e68f62
Author: Maarten Lankhorst <maarten.lankhorst@canonical.com> Date: Tue Jan 15 14:56:48 2013 +0100
drm/ttm: cleanup ttm_eu_reserve_buffers handling
With the lru lock no longer required for protecting reservations we can just do a ttm_bo_reserve_nolru on -EBUSY, and handle all errors in a single path.
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> Reviewed-by: Jerome Glisse <jglisse@redhat.com>
r254863: drm/ttm: Import Linux commit 5e45d7dfd74100d622f9cdc70bfd1f9fae1671de
Author: Maarten Lankhorst <maarten.lankhorst@canonical.com> Date: Tue Jan 15 14:57:05 2013 +0100
drm/ttm: add ttm_bo_reserve_slowpath
Instead of dropping everything, waiting for the bo to be unreserved and trying over, a better strategy would be to do a blocking wait.
This can be mapped a lot better to a mutex_lock-like call.
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Approved by: kib@
r254864: drm/ttm: Import Linux commit f2d476a110bc24fde008698ae9018c99e803e25c
Author: Maarten Lankhorst <maarten.lankhorst@canonical.com> Date: Tue Jan 15 14:57:10 2013 +0100
drm/ttm: use ttm_bo_reserve_slowpath_nolru in ttm_eu_reserve_buffers, v2
This requires re-use of the seqno, which increases fairness slightly. Instead of spinning with a new seqno every time we keep the current one, but still drop all other reservations we hold. Only when we succeed, we try to get back our other reservations again.
This should increase fairness slightly as well.
Changes since v1: - Increase val_seq before calling ttm_bo_reserve_slowpath_nolru and retrying to take all entries to prevent a race.
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Approved by: kib@
r254865: drm/ttm: Import Linux commit cc4c0c4de3c775be22072ec3251f2e581b63d9a0
Author: Maarten Lankhorst <maarten.lankhorst@canonical.com> Date: Tue Jan 15 14:57:28 2013 +0100
drm/ttm: unexport ttm_bo_wait_unreserved
All legitimate users of this function outside ttm_bo.c are gone, now it's only an implementation detail.
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Approved by: kib@
r254866: drm/ttm: Import Linux commit 630541863b29f88c7ab34e647758344e4cd1eafd
Author: Dave Airlie <airlied@gmail.com> Date: Wed Jan 16 14:25:44 2013 +1000
ttm: don't destroy old mm_node on memcpy failure
When we are using memcpy to move objects around, and we fail to memcpy due to lack of memory to populate or failure to finish the copy, we don't want to destroy the mm_node that has been copied into old_copy.
While working on a new kms driver that uses memcpy, if I overallocated bo's up to the memory limits, and eviction failed, then machine would oops soon after due to having an active bo with an already freed drm_mm embedded in it, freeing it a second time didn't end well.
Reviewed-by: Jerome Glisse <jglisse@redhat.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
Approved by: kib@
r254867: drm/ttm: Import Linux commit 014b34409fb2015f63663b6cafdf557fdf289628
Author: Dave Airlie <airlied@gmail.com> Date: Wed Jan 16 15:58:34 2013 +1000
ttm: on move memory failure don't leave a node dangling
if we have a move notify callback, when moving fails, we call move notify the opposite way around, however this ends up with *mem containing the mm_node from the bo, which means we double free it. This is a follow on to the previous fix.
Reviewed-by: Jerome Glisse <jglisse@redhat.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
Approved by: kib@
r254868: drm/ttm: Import Linux commit ff7c60c580d9722f820d85c9c58ca55ecc1ee7c4
Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Mon Jan 14 15:08:14 2013 +0100
drm/ttm: fix fence locking in ttm_buffer_object_transfer, 2nd try
This fixes up
commit e8e89622ed361c46bf90ba4828e685a8b603f7e5 Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Tue Dec 18 22:25:11 2012 +0100
drm/ttm: fix fence locking in ttm_buffer_object_transfer
which leaves behind a might_sleep in atomic context, since the fence_lock spinlock is held over a kmalloc(GFP_KERNEL) call. The fix is to revert the above commit and only take the lock where we need it, around the call to ->sync_obj_ref.
v2: Fixup things noticed by Maarten Lankhorst: - Brown paper bag locking bug. - No need for kzalloc if we clear the entire thing on the next line. - check for bo->sync_obj (totally unlikely race, but still someone else could have snuck in) and clear fbo->sync_obj if it's cleared already.
Reported-by: Dave Airlie <airlied@gmail.com> Cc: Jerome Glisse <jglisse@redhat.com> Cc: Maarten Lankhorst <maarten.lankhorst@canonical.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Dave Airlie <airlied@redhat.com>
Approved by: kib@
r254870: drm/ttm: Make ttm_bo_wait() call uninterruptible in page fault handler
This fixes a crash where a SIGLALRM, heavily used by X.Org, would interrupt the wait, causing the page fault to fail and the "Xorg" process to receive a SIGSEGV.
Approved by: kib@
r254871: drm/ttm: Fix style errors
r254873: drm/ttm: When removing a range of pages from a pool, remove all of them
Submitted by: Mark Kettenis and Jonathan Gray from OpenBSD Approved by: kib@
r254874: drm/ttm: Improve comment in ttm_bo_vm_ctor() about lack of ref acquisition
Approved by: kib@
r254875: ttm: "to_page->valid = VM_PAGE_BITS_ALL" before vm_page_dirty(to_page)
Approved by; kib@
r254876: drm/ttm: Fix unmap of buffer object
Add a new ttm_bo_release_mmap() function to unmap pages in a vm_object_t. Pages are freed when the buffer object is later released.
This function is called in ttm_bo_unmap_virtual_locked(), replacing Linux' unmap_mapping_range(). In particular this is called when a buffer object is about to be moved, so that its mapping is invalidated.
However, we don't use this function in ttm_bo_vm_dtor(), because the vm_object_t is already marked as OBJ_DEAD and the pages will be unmapped.
Approved by: kib@
r254877: drm/ttm: Fix style in ttm_bo_release_mmap()
r254878: drm/ttm: Fix a reversed condition and add missing locks
This allows to run OpenGL applications on at least two test machines with the Radeon driver.
Approved by: kib@
r254879: drm/ttm: Remove unused VM_ALLOC_DMA32 define
r254880: drm: Use the new drm_atomic.h, following the merge of projects/atomic64
Submitted by: jkim@
r259612: ttm_bo_vm_lookup_rb: actually make use of the red-black tree
Previously the code would just iterate over the whole tree as if it were just a list.
Without this change I would observe X server becoming more and more jerky over time.
MFC after: 5 days
|