Skip to content

R
raspberrypi-kernel

openeuler / raspberrypi-kernel

通知 13
Star 1
Fork 0
R
raspberrypi-kernel
切换分支/标签
  1. 06 11月, 2013 1 次提交
  3. 02 11月, 2013 1 次提交
    • M
      drm/radeon: fixup locking inversion between, mmap_sem and reservations · 28a326c5
      由 Maarten Lankhorst 提交于 
      op 08-10-13 18:58, Thomas Hellstrom schreef:
> On 10/08/2013 06:47 PM, Jerome Glisse wrote:
>> On Tue, Oct 08, 2013 at 06:29:35PM +0200, Thomas Hellstrom wrote:
>>> On 10/08/2013 04:55 PM, Jerome Glisse wrote:
>>>> On Tue, Oct 08, 2013 at 04:45:18PM +0200, Christian König wrote:
>>>>> Am 08.10.2013 16:33, schrieb Jerome Glisse:
>>>>>> On Tue, Oct 08, 2013 at 04:14:40PM +0200, Maarten Lankhorst wrote:
>>>>>>> Allocate and copy all kernel memory before doing reservations. This prevents a locking
>>>>>>> inversion between mmap_sem and reservation_class, and allows us to drop the trylocking
>>>>>>> in ttm_bo_vm_fault without upsetting lockdep.
>>>>>>>
>>>>>>> Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>
>>>>>> I would say NAK. Current code only allocate temporary page in AGP case.
>>>>>> So AGP case is userspace -> temp page -> cs checker -> radeon ib.
>>>>>>
>>>>>> Non AGP is directly memcpy to radeon IB.
>>>>>>
>>>>>> Your patch allocate memory memcpy userspace to it and it will then be
>>>>>> memcpy to IB. Which means you introduce an extra memcpy in the process
>>>>>> not something we want.
>>>>> Totally agree. Additional to that there is no good reason to provide
>>>>> anything else than anonymous system memory to the CS ioctl, so the
>>>>> dependency between the mmap_sem and reservations are not really
>>>>> clear to me.
>>>>>
>>>>> Christian.
>>>> I think is that in other code path you take mmap_sem first then reserve
>>>> bo. But here we reserve bo and then we take mmap_sem because of copy
>>> >from user.
>>>> Cheers,
>>>> Jerome
>>>>
>>> Actually the log message is a little confusing. I think the mmap_sem
>>> locking inversion problem is orthogonal to what's being fixed here.

> >>> This patch fixes the possible recursive bo::reserve caused by
> >>> malicious user-space handing a pointer to ttm memory so that the ttm
> >>> fault handler is called when bos are already reserved. That may
> >>> cause a (possibly interruptible) livelock.

>>> Once that is fixed, we are free to choose the mmap_sem ->
>>> bo::reserve locking order. Currently it's bo::reserve->mmap_sem(),
>>> but the hack required in the ttm fault handler is admittedly a bit
>>> ugly.  The plan is to change the locking order to
>>> mmap_sem->bo::reserve

> >>> I'm not sure if it applies to this particular case, but it should be
> >>> possible to make sure that copy_from_user_inatomic() will always
> >>> succeed, by making sure the pages are present using
> >>> get_user_pages(), and release the pages after
> >>> copy_from_user_inatomic() is done. That way there's no need for a
> >>> double memcpy slowpath, but if the copied data is very fragmented I
> >>> guess the resulting code may look ugly. The get_user_pages()
> >>> function will return an error if it hits TTM pages.

>>> /Thomas
>> get_user_pages + copy_from_user_inatomic is overkill. We should just
>> do get_user_pages which fails with ttm memory and then use copy_highpage
>> helper.
>>
>> Cheers,
>> Jerome
> Yeah, it may well be that that's the preferred solution.
>
> /Thomas
>
I still disagree, and shuffled radeon_ib_get around to be called sooner.

How does the patch below look?
8<-------
Allocate and copy all kernel memory before doing reservations. This prevents a locking
inversion between mmap_sem and reservation_class, and allows us to drop the trylocking
in ttm_bo_vm_fault without upsetting lockdep.

Changes since v1:
- Kill extra memcpy for !AGP case.
Signed-off-by: NMaarten Lankhorst <maarten.lankhorst@canonical.com>
Reviewed-by: NJerome Glisse <jglisse@redhat.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>
      28a326c5
  5. 12 2月, 2013 1 次提交
  7. 01 2月, 2013 7 次提交
  9. 11 1月, 2013 2 次提交
  11. 05 1月, 2013 1 次提交
  13. 20 12月, 2012 1 次提交
  15. 14 12月, 2012 2 次提交
  17. 13 12月, 2012 1 次提交
  19. 03 10月, 2012 1 次提交
  21. 27 9月, 2012 1 次提交
  23. 21 9月, 2012 1 次提交
  25. 30 8月, 2012 3 次提交
  27. 20 8月, 2012 2 次提交
  29. 13 8月, 2012 2 次提交
  31. 09 8月, 2012 1 次提交
  33. 16 6月, 2012 1 次提交
  35. 10 5月, 2012 1 次提交
  37. 26 3月, 2012 1 次提交
  39. 20 3月, 2012 3 次提交
  41. 29 2月, 2012 1 次提交
  43. 13 2月, 2012 2 次提交
    • J
      drm/radeon: add support for evergreen/ni tiling informations v11 · 285484e2
      由 Jerome Glisse 提交于 
      evergreen and northern island gpu needs more informations for 2D tiling
than previous r6xx/r7xx. Add field to tiling ioctl to allow userspace
to provide those.

The v8 cs checking change to track color view on r6xx/r7xx doesn't
affect old userspace as old userspace always emited 0 for this register.

v2 fix r6xx/r7xx 2D tiling computation
v3 fix r6xx/r7xx height align for untiled surface & add support for
   tile split on evergreen and newer
v4 improve tiling debugging output
v5 fix tile split code for evergreen and newer
v6 set proper tile split for crtc register
v7 fix tile split limit value
v8 add COLOR_VIEW checking to r6xx/r7xx checker, add evergreen cs
   checking, update safe reg for r600, evergreen and cayman.
   Evergreen checking need some work around for stencil alignment
   issues
v9 fix tile split value range, fix compressed texture handling and
   mipmap calculation, allow evergreen check to be silencious in
   front of current broken userspace (depth/stencil alignment issue)
v10 fix eg 3d texture and compressed texture, fix r600 depth array,
    fix r600 color view computation, add support for evergreen stencil
    split
v11 more verbose debugging in some case
Signed-off-by: NJerome Glisse <jglisse@redhat.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>
      285484e2
    • M
      drm/radeon/kms: add support for streamout v7 · dd220a00
      由 Marek Olšák 提交于 
      v2: agd5f: add strmout CS checking, copy_dw register checking

v3: agd5f: don't use cs_check_reg() for copy_dw checking as it
will incorrectly patch the command stream for certain regs.

v4: agd5f: add warning if safe reg check fails for copy_dw

v5: agd5f: add stricter checking for 6xx/7xx

v6: agd5f: add range checking for copy_dw on eg+,
add sx_surface_sync to safe reg list for 7xx.

v7: agd5f: add stricter checking for eg+
Signed-off-by: NMarek Olšák <maraeo@gmail.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>
      dd220a00
  45. 06 1月, 2012 1 次提交
    • J
      drm/radeon: GPU virtual memory support v22 · 721604a1
      由 Jerome Glisse 提交于 
      Virtual address space are per drm client (opener of /dev/drm).
Client are in charge of virtual address space, they need to
map bo into it by calling DRM_RADEON_GEM_VA ioctl.

First 16M of virtual address space is reserved by the kernel.

Once using 2 level page table we should be able to have a small
vram memory footprint for each pt (there would be one pt for all
gart, one for all vram and then one first level for each virtual
address space).

Plan include using the sub allocator for a common vm page table
area and using memcpy to copy vm page table in & out. Or use
a gart object and copy things in & out using dma.

v2: agd5f fixes:
- Add vram base offset for vram pages.  The GPU physical address of a
vram page is FB_OFFSET + page offset.  FB_OFFSET is 0 on discrete
cards and the physical bus address of the stolen memory on
integrated chips.
- VM_CONTEXT1_PROTECTION_FAULT_DEFAULT_ADDR covers all vmid's >= 1

v3: agd5f:
- integrate with the semaphore/multi-ring stuff

v4:
- rebase on top ttm dma & multi-ring stuff
- userspace is now in charge of the address space
- no more specific cs vm ioctl, instead cs ioctl has a new
  chunk

v5:
- properly handle mem == NULL case from move_notify callback
- fix the vm cleanup path

v6:
- fix update of page table to only happen on valid mem placement

v7:
- add tlb flush for each vm context
- add flags to define mapping property (readable, writeable, snooped)
- make ring id implicit from ib->fence->ring, up to each asic callback
  to then do ring specific scheduling if vm ib scheduling function

v8:
- add query for ib limit and kernel reserved virtual space
- rename vm->size to max_pfn (maximum number of page)
- update gem_va ioctl to also allow unmap operation
- bump kernel version to allow userspace to query for vm support

v9:
- rebuild page table only when bind and incrementaly depending
  on bo referenced by cs and that have been moved
- allow virtual address space to grow
- use sa allocator for vram page table
- return invalid when querying vm limit on non cayman GPU
- dump vm fault register on lockup

v10: agd5f:
- Move the vm schedule_ib callback to a standalone function, remove
  the callback and use the existing ib_execute callback for VM IBs.

v11:
- rebase on top of lastest Linus

v12: agd5f:
- remove spurious backslash
- set IB vm_id to 0 in radeon_ib_get()

v13: agd5f:
- fix handling of RADEON_CHUNK_ID_FLAGS

v14:
- fix va destruction
- fix suspend resume
- forbid bo to have several different va in same vm

v15:
- rebase

v16:
- cleanup left over of vm init/fini

v17: agd5f:
- cs checker

v18: agd5f:
- reworks the CS ioctl to better support multiple rings and
VM.  Rather than adding a new chunk id for VM, just re-use the
IB chunk id and add a new flags for VM mode.  Also define additional
dwords for the flags chunk id to define the what ring we want to use
(gfx, compute, uvd, etc.) and the priority.

v19:
- fix cs fini in weird case of no ib
- semi working flush fix for ni
- rebase on top of sa allocator changes

v20: agd5f:
- further CS ioctl cleanups from Christian's comments

v21: agd5f:
- integrate CS checker improvements

v22: agd5f:
- final cleanups for release, only allow VM CS on cayman
Signed-off-by: NJerome Glisse <jglisse@redhat.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>
      721604a1
  47. 20 11月, 2011 1 次提交
    • M
      drm/radeon/kms: add a CS ioctl flag not to rewrite tiling flags in the CS · e70f224c
      由 Marek Olšák 提交于 
      This adds a new optional chunk to the CS ioctl that specifies optional flags
to the CS parser. Why this is useful is explained below. Note that some regs
no longer need the NOP relocation packet if this feature is enabled.
Tested on r300g and r600g with this flag disabled and enabled.

Assume there are two contexts sharing the same mipmapped tiled texture.
One context wants to render into the first mipmap and the other one
wants to render into the last mipmap. As you probably know, the hardware
has a MACRO_SWITCH feature, which turns off macro tiling for small mipmaps,
but that only applies to samplers.
(at least on r300-r500, though later hardware likely behaves the same)

So we want to just re-set the tiling flags before rendering (writing
packets), right? ... No. The contexts run in parallel, so they may
set the tiling flags simultaneously and then fire their command streams
also simultaneously. The last one setting the flags wins, the other one
loses.

Another problem is when one context wants to render into the first and
the last mipmap in one CS. Impossible. It must flush before changing
tiling flags and do the rendering into the smaller mipmaps in another CS.

Yet another problem is that writing copy_blit in userspace would be a mess
involving re-setting tiling flags to please the kernel, and causing races
with other contexts at the same time.

The only way out of this is to send tiling flags with each CS, ideally
with each relocation. But we already do that through the registers.
So let's just use what we have in the registers.
Signed-off-by: NMarek Olšák <maraeo@gmail.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>
      e70f224c
  49. 18 10月, 2011 1 次提交