LLC is not SNB/IVB-specific, so we should check for it in a more generic
way.
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Signed-off-by: Eugeni Dodonov <eugeni.dodonov@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
So we have a few places where the drm drivers would like to sleep to
be nice to the system, mainly in the modesetting paths, but we also
have two cases were atomic modesetting must take place, panic writing
and kernel debugger. So provide a central inline to determine if a
sleep or delay should be used and use this in the intel and radeon drivers.
v2: drop intel_drv.h MSLEEP macro, nobody uses it.
Based on patch from Michel Dänzer <michel.daenzer@amd.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=43941
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Dave Airlie <airlied@redhat.com>
ttm tt rework modified the way we allocate and populate the
ttm_tt structure, the AGP side was missing some bit to properly
work. Fix those and fix radeon and nouveau AGP support.
Tested on radeon only so far.
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Use semaphores to sync buffers across rings in the CS
ioctl. Add a reloc flag to allow userspace to skip
sync for buffers.
agd5f: port to latest CS ioctl changes.
v2: add ring lock/unlock to make sure changes hit the ring.
Signed-off-by: Christian König <deathsimple@vodafone.de>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
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: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
In cases where the scanout hw is sufficiently similar between "overlay"
and traditional crtc layers, it might be convenient to allow the driver
to create internal drm_plane helper objects used by the drm_crtc
implementation, rather than duplicate code between the plane and crtc.
A private plane is not exposed to userspace.
Signed-off-by: Rob Clark <rob@ti.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
These registers are automatically incremented by the hardware during
transform feedback to track where the next streamed vertex output
should go. Unlike the previous generation, which had a packet for
setting the corresponding registers to a defined value, gen7 only has
MI_LOAD_REGISTER_IMM to do so. That's a secure packet (since it loads
an arbitrary register), so we need to do it from the kernel, and it
needs to be settable atomically with the batchbuffer execution so that
two clients doing transform feedback don't stomp on each others'
state.
Instead of building a more complicated interface involcing setting the
registers to a specific value, just set them to 0 when asked and
userland can tweak its pointers accordingly.
Signed-off-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Eugeni Dodonov <eugeni.dodonov@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Signed-off-by: Keith Packard <keithp@keithp.com>
Add new ioctls for getting and setting the current destination color
key. This allows for simple overlay display control by matching a color
key value in the primary plane before blending the overlay on top.
v2: remove unnecessary mutex acquire/release around reg accesses
v3: add support for full color key management
v4: fix copy & paste bug in snb_get_colorkey
don't bother checking min/max values against docs as the docs are likely
wrong (how could we handle 10bpc surface formats?)
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
This patch is hdmi display support for exynos drm driver.
There is already v4l2 based exynos hdmi driver in drivers/media/video/s5p-tv
and some low level code is already in s5p-tv and even headers for register
define are almost same. but in this patch, we decide not to consider separated
common code with s5p-tv.
Exynos HDMI is composed of 5 blocks, mixer, vp, hdmi, hdmiphy and ddc.
1. mixer. The piece of hardware responsible for mixing and blending multiple
data inputs before passing it to an output device. The mixer is capable of
handling up to three image layers. One is the output of VP. Other two are
images in RGB format. The blending factor, and layers' priority are controlled
by mixer's registers. The output is passed to HDMI.
2. vp (video processor). It is used for processing of NV12/NV21 data. An image
stored in RAM is accessed by DMA. The output in YCbCr444 format is send to
mixer.
3. hdmi. The piece of HW responsible for generation of HDMI packets. It takes
pixel data from mixer and transforms it into data frames. The output is send
to HDMIPHY interface.
4. hdmiphy. Physical interface for HDMI. Its duties are sending HDMI packets to
HDMI connector. Basically, it contains a PLL that produces source clock for
mixer, vp and hdmi.
5. ddc (display data channel). It is dedicated i2c channel to exchange display
information as edid with display monitor.
With plane support, exynos hdmi driver fully supports two mixer layes and vp
layer. Also vp layer supports multi buffer plane pixel formats having non
contigus memory spaces.
In exynos drm driver, common drm_hdmi driver to interface with drm framework
has opertion pointers for mixer and hdmi. this drm_hdmi driver is registered as
sub driver of exynos_drm. hdmi has hdmiphy and ddc i2c clients and controls
them. mixer controls all overlay layers in both mixer and vp.
Vblank interrupts for hdmi are handled by mixer internally because drm
framework cannot support multiple irq id. And pipe number is used to check
which display device irq happens.
History
v2: this version
- drm plane feature support to handle overlay layers.
- multi buffer plane pixel format support for vp layer.
- vp layer support
RFCv1: original
- at https://lkml.org/lkml/2011/11/4/164
Signed-off-by: Seung-Woo Kim <sw0312.kim@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
Signed-off-by: Joonyoung Shim <jy0922.shim@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Multi buffer plane pixel format has seperated memory spaces for each
plane. For example, NV12M has Y plane and CbCr plane and these are in
non contiguous memory region. Compared with NV12, NV12M's memory shape
is like following.
NV12 : ______(Y)(CbCr)_______
NV12M : __(Y)_ ..... _(CbCr)__
Signed-off-by: Seung-Woo Kim <sw0312.kim@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
In contrast to kms drivers, sis/via _always_ associated a buffer with
a drm fd. So by the time we reach lastclose, all open drm fds are gone
and with them their associated objects.
So when sis/via call drm_sman_cleanup in their lastclose funcs, that
will free 0 objects.
The owner tracking now serves no purpose at all, hence rip it ou. We
can't kill the corresponding fields in struct drm_memblock_item yet
because we hijack these in the new driver private owner tracking. But
now that drm_sman.c doesn't touch ->owner_list anymore, we need to
kill the list_move hack and properly add the item to the file_priv
list.
Also leave the list_del(&obj->owner_list) in drm_sman_free for the
moment, it will move to the drivers when sman disappears completely.
v2: Remove the redundant INIT_LIST_HEAD as noted by Chris Wilson
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
By attach a driver private struct to each open drm fd.
Because we steal the owner_list from drm_sman until things settle,
use list_move instead of list_add.
This requires to export a drm_sman function temporarily before
drm_sman will die for real completely.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
The exynos fimd supports 5 window overlays. Only one window overlay of
fimd is used by the crtc, so we need plane feature to use the rest
window overlays.
This creates one ioctl exynos specific - DRM_EXYNOS_PLANE_SET_ZPOS, it
is the ioctl to decide for user to assign which window overlay.
Signed-off-by: Joonyoung Shim <jy0922.shim@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
This function returns the number of planes used by a specific pixel
format.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Rob Clark <rob.clark@linaro.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Merge in the upstream tree to bring in the mainline fixes.
Conflicts:
drivers/gpu/drm/exynos/exynos_drm_fbdev.c
drivers/gpu/drm/nouveau/nouveau_sgdma.c
Otherwise each driver would need to keep the information inside
their own framebuffer object structure. Also add offsets[]. BOs
on the other hand are driver specific, so those can be kept in
driver specific structures.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
drm_fourcc.h can be included from user space so use the appropriate types.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
The code happened to compile because the flag wasn't actually used yet.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
* git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi-rc-fixes-2.6: (22 commits)
[SCSI] fcoe: fix fcoe in a DCB environment by adding DCB notifiers to set skb priority
[SCSI] bnx2i: Fixed kernel panic caused by unprotected task->sc->request deref
[SCSI] qla4xxx: check for failed conn setup
[SCSI] qla4xxx: a small loop fix
[SCSI] qla4xxx: fix flash/ddb support
[SCSI] zfcp: return early from slave_destroy if slave_alloc returned early
[SCSI] fcoe: Fix preempt count leak in fcoe_filter_frames()
[SCSI] qla2xxx: Update version number to 8.03.07.12-k.
[SCSI] qla2xxx: Submit all chained IOCBs for passthrough commands on request queue 0.
[SCSI] qla2xxx: Correct fc_host port_state display.
[SCSI] qla2xxx: Disable generating pause frames when firmware hang detected for ISP82xx.
[SCSI] qla2xxx: Clear mailbox busy flag during premature mailbox completion for ISP82xx.
[SCSI] qla2xxx: Encapsulate prematurely completing mailbox commands during ISP82xx firmware hang.
[SCSI] qla2xxx: Display IPE error message for ISP82xx.
[SCSI] qla2xxx: Return the correct value for a mailbox command if 82xx is in reset recovery.
[SCSI] qla2xxx: Enable Minidump by default with default capture mask 0x1f.
[SCSI] qla2xxx: Stop unconditional completion of mailbox commands issued in interrupt mode during firmware hang.
[SCSI] qla2xxx: Revert back the request queue mapping to request queue 0.
[SCSI] qla2xxx: Don't call alloc_fw_dump for ISP82XX.
[SCSI] qla2xxx: Check for SCSI status on underruns.
...
* 'drm-intel-fixes' of git://people.freedesktop.org/~keithp/linux:
drm/i915/dp: Dither down to 6bpc if it makes the mode fit
drm/i915: enable semaphores on per-device defaults
drm/i915: don't set unpin_work if vblank_get fails
drm/i915: By default, enable RC6 on IVB and SNB when reasonable
iommu: Export intel_iommu_enabled to signal when iommu is in use
drm/i915/sdvo: Include LVDS panels for the IS_DIGITAL check
drm/i915: prevent division by zero when asking for chipset power
drm/i915: add PCH info to i915_capabilities
drm/i915: set the right SDVO transcoder for CPT
drm/i915: no-lvds quirk for ASUS AT5NM10T-I
drm/i915: Treat pre-gen4 backlight duty cycle value consistently
drm/i915: Hook up Ivybridge eDP
drm/i915: add multi-threaded forcewake support
In i915 driver, we do not enable either rc6 or semaphores on SNB when dmar
is enabled. The new 'intel_iommu_enabled' variable signals when the
iommu code is in operation.
Cc: Ted Phelps <phelps@gnusto.com>
Cc: Peter <pab1612@gmail.com>
Cc: Lukas Hejtmanek <xhejtman@fi.muni.cz>
Cc: Andrew Lutomirski <luto@mit.edu>
CC: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Eugeni Dodonov <eugeni.dodonov@intel.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Use DCB notifiers to set the skb priority to allow packets
to be steered and tagged correctly over DCB enabled drivers
that setup traffic classes.
This allows queue_mapping() routines to be removed in these
drivers that were previously inspecting the ethertype of
every skb to mark FCoE/FIP frames.
Signed-off-by: John Fastabend <john.r.fastabend@intel.com>
Signed-off-by: Robert Love <robert.w.love@intel.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
Exactly like roundup_pow_of_two(1), the rounddown version was buggy for
the case of a compile-time constant '1' argument. Probably because it
originated from the same code, sharing history with the roundup version
from before the bugfix (for that one, see commit 1a06a52ee1: "Fix
roundup_pow_of_two(1)").
However, unlike the roundup version, the fix for rounddown is to just
remove the broken special case entirely. It's simply not needed - the
generic code
1UL << ilog2(n)
does the right thing for the constant '1' argment too. The only reason
roundup needed that special case was because rounding up does so by
subtracting one from the argument (and then adding one to the result)
causing the obvious problems with "ilog2(0)".
But rounddown doesn't do any of that, since ilog2() naturally truncates
(ie "rounds down") to the right rounded down value. And without the
ilog2(0) case, there's no reason for the special case that had the wrong
value.
tl;dr: rounddown_pow_of_two(1) should be 1, not 0.
Acked-by: Dmitry Torokhov <dtor@vmware.com>
Cc: stable@kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/cjb/mmc:
mmc: core: Fix deadlock when the CONFIG_MMC_UNSAFE_RESUME is not defined
mmc: sdhci-s3c: Remove old and misprototyped suspend operations
mmc: tmio: fix clock gating on platforms with a .set_pwr() method
mmc: sh_mmcif: fix clock gating on platforms with a .down_pwr() method
mmc: core: Fix typo at mmc_card_sleep
mmc: core: Fix power_off_notify during suspend
mmc: core: Fix setting power notify state variable for non-eMMC
mmc: core: Add quirk for long data read time
mmc: Add module.h include to sdhci-cns3xxx.c
mmc: mxcmmc: fix falling back to PIO
mmc: omap_hsmmc: DMA unmap only once in case of MMC error
Adds a quirk that sets the data read timeout to a fixed value instead
of relying on the information in the CSD. The timeout value chosen
is 300ms since that has proven enough for the problematic cards found,
but could be increased if other cards require this.
This patch also enables this quirk for certain Micron cards known to
have this problem.
Signed-off-by: Stefan Nilsson XK <stefan.xk.nilsson@stericsson.com>
Signed-off-by: Ulf Hansson <ulf.hansson@stericsson.com>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Cc: <stable@kernel.org>
Signed-off-by: Chris Ball <cjb@laptop.org>
* git://git.kernel.org/pub/scm/linux/kernel/git/cmetcalf/linux-tile:
arch/tile: use new generic {enable,disable}_percpu_irq() routines
drivers/net/ethernet/tile: use skb_frag_page() API
asm-generic/unistd.h: support new process_vm_{readv,write} syscalls
arch/tile: fix double-free bug in homecache_free_pages()
arch/tile: add a few #includes and an EXPORT to catch up with kernel changes.
Use atomic-long operations instead of looping around cmpxchg().
[akpm@linux-foundation.org: massage atomic.h inclusions]
Signed-off-by: Konstantin Khlebnikov <khlebnikov@openvz.org>
Cc: Dave Chinner <david@fromorbit.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* '3.2-rc-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/nab/target-pending: (25 commits)
iscsi-target: Fix hex2bin warn_unused compile message
target: Don't return an error if disabling unsupported features
target/rd: fix or rewrite the copy routine
target/rd: simplify the page/offset computation
target: remove the unused se_dev_list
target/file: walk properly over sg list
target: remove unused struct fields
target: Fix page length in emulated INQUIRY VPD page 86h
target: Handle 0 correctly in transport_get_sectors_6()
target: Don't return an error status for 0-length READ and WRITE
iscsi-target: Use kmemdup rather than duplicating its implementation
iscsi-target: Add missing F_BIT for iscsi_tm_rsp
iscsi-target: Fix residual count hanlding + remove iscsi_cmd->residual_count
target: Reject SCSI data overflow for fabrics using transport_generic_map_mem_to_cmd
target: remove the unused t_task_pt_sgl and t_task_pt_sgl_num se_cmd fields
target: remove the t_tasks_bidi se_cmd field
target: remove the t_tasks_fua se_cmd field
target: remove the se_ordered_node se_cmd field
target: remove the se_obj_ptr and se_orig_obj_ptr se_cmd fields
target: Drop config_item_name usage in fabric TFO->free_wwn()
...
__d_path() API is asking for trouble and in case of apparmor d_namespace_path()
getting just that. The root cause is that when __d_path() misses the root
it had been told to look for, it stores the location of the most remote ancestor
in *root. Without grabbing references. Sure, at the moment of call it had
been pinned down by what we have in *path. And if we raced with umount -l, we
could have very well stopped at vfsmount/dentry that got freed as soon as
prepend_path() dropped vfsmount_lock.
It is safe to compare these pointers with pre-existing (and known to be still
alive) vfsmount and dentry, as long as all we are asking is "is it the same
address?". Dereferencing is not safe and apparmor ended up stepping into
that. d_namespace_path() really wants to examine the place where we stopped,
even if it's not connected to our namespace. As the result, it looked
at ->d_sb->s_magic of a dentry that might've been already freed by that point.
All other callers had been careful enough to avoid that, but it's really
a bad interface - it invites that kind of trouble.
The fix is fairly straightforward, even though it's bigger than I'd like:
* prepend_path() root argument becomes const.
* __d_path() is never called with NULL/NULL root. It was a kludge
to start with. Instead, we have an explicit function - d_absolute_root().
Same as __d_path(), except that it doesn't get root passed and stops where
it stops. apparmor and tomoyo are using it.
* __d_path() returns NULL on path outside of root. The main
caller is show_mountinfo() and that's precisely what we pass root for - to
skip those outside chroot jail. Those who don't want that can (and do)
use d_path().
* __d_path() root argument becomes const. Everyone agrees, I hope.
* apparmor does *NOT* try to use __d_path() or any of its variants
when it sees that path->mnt is an internal vfsmount. In that case it's
definitely not mounted anywhere and dentry_path() is exactly what we want
there. Handling of sysctl()-triggered weirdness is moved to that place.
* if apparmor is asked to do pathname relative to chroot jail
and __d_path() tells it we it's not in that jail, the sucker just calls
d_absolute_path() instead. That's the other remaining caller of __d_path(),
BTW.
* seq_path_root() does _NOT_ return -ENAMETOOLONG (it's stupid anyway -
the normal seq_file logics will take care of growing the buffer and redoing
the call of ->show() just fine). However, if it gets path not reachable
from root, it returns SEQ_SKIP. The only caller adjusted (i.e. stopped
ignoring the return value as it used to do).
Reviewed-by: John Johansen <john.johansen@canonical.com>
ACKed-by: John Johansen <john.johansen@canonical.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Cc: stable@vger.kernel.org
* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
ftrace: Fix hash record accounting bug
perf: Fix parsing of __print_flags() in TP_printk()
jump_label: jump_label_inc may return before the code is patched
ftrace: Remove force undef config value left for testing
tracing: Restore system filter behavior
tracing: fix event_subsystem ref counting
Provide helper function to compute the kernel memory size needed
for each buffer object. Move all the accounting inside ttm, simplifying
driver and avoiding code duplication accross them.
v2 fix accounting of ghost object, one would have thought that i
would have run into the issue since a longtime but it seems
ghost object are rare when you have plenty of vram ;)
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
Move dma data to a superset ttm_dma_tt structure which herit
from ttm_tt. This allow driver that don't use dma functionalities
to not have to waste memory for it.
V2 Rebase on top of no memory account changes (where/when is my
delorean when i need it ?)
V3 Make sure page list is initialized empty
V4 typo/syntax fixes
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
In TTM world the pages for the graphic drivers are kept in three different
pools: write combined, uncached, and cached (write-back). When the pages
are used by the graphic driver the graphic adapter via its built in MMU
(or AGP) programs these pages in. The programming requires the virtual address
(from the graphic adapter perspective) and the physical address (either System RAM
or the memory on the card) which is obtained using the pci_map_* calls (which does the
virtual to physical - or bus address translation). During the graphic application's
"life" those pages can be shuffled around, swapped out to disk, moved from the
VRAM to System RAM or vice-versa. This all works with the existing TTM pool code
- except when we want to use the software IOTLB (SWIOTLB) code to "map" the physical
addresses to the graphic adapter MMU. We end up programming the bounce buffer's
physical address instead of the TTM pool memory's and get a non-worky driver.
There are two solutions:
1) using the DMA API to allocate pages that are screened by the DMA API, or
2) using the pci_sync_* calls to copy the pages from the bounce-buffer and back.
This patch fixes the issue by allocating pages using the DMA API. The second
is a viable option - but it has performance drawbacks and potential correctness
issues - think of the write cache page being bounced (SWIOTLB->TTM), the
WC is set on the TTM page and the copy from SWIOTLB not making it to the TTM
page until the page has been recycled in the pool (and used by another application).
The bounce buffer does not get activated often - only in cases where we have
a 32-bit capable card and we want to use a page that is allocated above the
4GB limit. The bounce buffer offers the solution of copying the contents
of that 4GB page to an location below 4GB and then back when the operation has been
completed (or vice-versa). This is done by using the 'pci_sync_*' calls.
Note: If you look carefully enough in the existing TTM page pool code you will
notice the GFP_DMA32 flag is used - which should guarantee that the provided page
is under 4GB. It certainly is the case, except this gets ignored in two cases:
- If user specifies 'swiotlb=force' which bounces _every_ page.
- If user is using a Xen's PV Linux guest (which uses the SWIOTLB and the
underlaying PFN's aren't necessarily under 4GB).
To not have this extra copying done the other option is to allocate the pages
using the DMA API so that there is not need to map the page and perform the
expensive 'pci_sync_*' calls.
This DMA API capable TTM pool requires for this the 'struct device' to
properly call the DMA API. It also has to track the virtual and bus address of
the page being handed out in case it ends up being swapped out or de-allocated -
to make sure it is de-allocated using the proper's 'struct device'.
Implementation wise the code keeps two lists: one that is attached to the
'struct device' (via the dev->dma_pools list) and a global one to be used when
the 'struct device' is unavailable (think shrinker code). The global list can
iterate over all of the 'struct device' and its associated dma_pool. The list
in dev->dma_pools can only iterate the device's dma_pool.
/[struct device_pool]\
/---------------------------------------------------| dev |
/ +-------| dma_pool |
/-----+------\ / \--------------------/
|struct device| /-->[struct dma_pool for WC]</ /[struct device_pool]\
| dma_pools +----+ /-| dev |
| ... | \--->[struct dma_pool for uncached]<-/--| dma_pool |
\-----+------/ / \--------------------/
\----------------------------------------------/
[Two pools associated with the device (WC and UC), and the parallel list
containing the 'struct dev' and 'struct dma_pool' entries]
The maximum amount of dma pools a device can have is six: write-combined,
uncached, and cached; then there are the DMA32 variants which are:
write-combined dma32, uncached dma32, and cached dma32.
Currently this code only gets activated when any variant of the SWIOTLB IOMMU
code is running (Intel without VT-d, AMD without GART, IBM Calgary and Xen PV
with PCI devices).
Tested-by: Michel Dänzer <michel@daenzer.net>
[v1: Using swiotlb_nr_tbl instead of swiotlb_enabled]
[v2: Major overhaul - added 'inuse_list' to seperate used from inuse and reorder
the order of lists to get better performance.]
[v3: Added comments/and some logic based on review, Added Jerome tag]
[v4: rebase on top of ttm_tt & ttm_backend merge]
[v5: rebase on top of ttm memory accounting overhaul]
[v6: New rebase on top of more memory accouting changes]
[v7: well rebase on top of no memory accounting changes]
[v8: make sure pages list is initialized empty]
[v9: calll ttm_mem_global_free_page in unpopulate for accurate accountg]
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Acked-by: Thomas Hellstrom <thellstrom@vmware.com>
Move the page allocation and freeing to driver callback and
provide ttm code helper function for those.
Most intrusive change, is the fact that we now only fully
populate an object this simplify some of code designed around
the page fault design.
V2 Rebase on top of memory accounting overhaul
V3 New rebase on top of more memory accouting changes
V4 Rebase on top of no memory account changes (where/when is my
delorean when i need it ?)
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
ttm_backend will only exist with a ttm_tt, and ttm_tt
will only be of interest when bound to a backend. Merge them
to avoid code and data duplication.
V2 Rebase on top of memory accounting overhaul
V3 Rebase on top of more memory accounting changes
V4 Rebase on top of no memory account changes (where/when is my
delorean when i need it ?)
V5 make sure ttm is unbound before destroying, change commit
message on suggestion from Tormod Volden
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
Use the ttm_tt pages array for pages allocations, move the list
unwinding into the page allocation functions.
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
This field is not use by any of the driver just drop it.
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
Split btw highmem and lowmem page was rendered useless by the
pool code. Remove it. Note further cleanup would change the
ttm page allocation helper to actualy take an array instead
of relying on list this could drasticly reduce the number of
function call in the common case of allocation whole buffer.
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
This was never use in none of the driver, properly using userspace
page for bo would need more code (vma interaction mostly). Removing
this dead code in preparation of ttm_tt & backend merge.
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
