Age | Commit message (Collapse) | Author |
|
And now unused kernel option CONFIG_NVMAP_HANDLE_MARSHAL
Change-Id: Ic5a1af334354faeb72de204c8698c78a0aa69935
Signed-off-by: Dan Willemsen <dwillemsen@nvidia.com>
Reviewed-on: http://git-master/r/377221
Reviewed-by: Krishna Reddy <vdumpa@nvidia.com>
|
|
Change-Id: I189c08ecd66b762f2fa978f8bf2f0db6b8bebce7
Signed-off-by: Dan Willemsen <dwillemsen@nvidia.com>
Reviewed-on: http://git-master/r/377220
Reviewed-by: Krishna Reddy <vdumpa@nvidia.com>
Tested-by: Krishna Reddy <vdumpa@nvidia.com>
|
|
Add and Update nvmap trace events to print stats info for use case
analysis.
Fix dma_buf2 events to get dev name at right time.
Change-Id: Ib79de255714d2d28317b4a4ab1f73dbce051623f
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/375107
|
|
Add support for nvmap stats to allow analyzing use cases.
Change-Id: I9b27ecfb8e2d88899bf226fea99bfacfe21c7418
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/373718
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Alex Waterman <alexw@nvidia.com>
|
|
Coverity id : 25526
Change-Id: I99c72db80445f2ddecdda563df048883ef1df31c
Signed-off-by: Deepak Nibade <dnibade@nvidia.com>
Reviewed-on: http://git-master/r/364416
GVS: Gerrit_Virtual_Submit
Reviewed-by: Sachin Nikam <snikam@nvidia.com>
|
|
Add check for invalid nvmap handle
bug 1434818
Change-Id: I8ad4c5f8f40416609bcc819789e5048bbabe638b
Signed-off-by: Vandana Salve <vsalve@nvidia.com>
Reviewed-on: http://git-master/r/360717
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Krishna Reddy <vdumpa@nvidia.com>
|
|
On arm64, cache flushes are bypassed incorrectly during allocation.
This patch re-enables cache flush during allocation and fixes incorrect
args passed to __flush_dcache_page on arm64.
Reduce CONFIG_ARM64 macro checks to keep most of arm and arm64 code common.
Bug 1431946
Change-Id: Ie1c74636db9f9f06b06b8d47f6179cab57c8c177
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/352008
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Hayden Du <haydend@nvidia.com>
Reviewed-by: Alexander Van Brunt <avanbrunt@nvidia.com>
Tested-by: Hayden Du <haydend@nvidia.com>
|
|
Fix marshalling pointers between a 32-bit userspace and a 64-bit
kernel. First, the pointer needs to be marshaled regardless of if it
points to a file or not. Second, shift right to allow the kenrel to
address 34 bits (16 GB) of memory.
Change-Id: Ib347f14bda9ab94c50373f7c879bc17e26aee551
Signed-off-by: Alex Van Brunt <avanbrunt@nvidia.com>
Reviewed-on: http://git-master/r/339722
|
|
fix handle get during pinop when CONFIG_COMPAT is enabled or
pointer size is > 32-bit.
Change-Id: Id63bb3fbdb2a00295144f391a62e29547a095876
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/276388
Tested-by: Bo Yan <byan@nvidia.com>
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Bo Yan <byan@nvidia.com>
|
|
Bug 1248538
Change-Id: I4bbfac2936afcd36114d5d2f0215e8c050da0bd4
Signed-off-by: Peng Du <pdu@nvidia.com>
Reviewed-on: http://git-master/r/196621
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Alexander Van Brunt <avanbrunt@nvidia.com>
|
|
Remove deferred cache flush as pin/unpin would not be
called with dmabuf interface. User space wouldn't
call pin/unpin either going forward.
Change-Id: I4cf1896f5696cb18cf6f63a93efb71b40739e015
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/356126
Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit
|
|
Don't set VM_IO flag as memory allocated by NvMap is not really IO.
Don't set VM_MIXEDMAP flag as NvMap allocated memory is either all
struct page * or pure PFN. For PFN memory, set VM_PFNMAP flag.
Bug 1433646
Change-Id: I626f1f1507547821aa9d5d3f4e2828ba567e953f
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/352819
|
|
Remove support to allow mapping UC/WC as IWB/WB.
This feature is no longer in use not compatible with ION.
Bug 1407581
Change-Id: I29e2d3bab99f9b3c7c9d769f7ddb24d6bf852126
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/331296
|
|
Check for invalid handle passed to nvmap to
avoid crash in nvmap
bug 1415394
Change-Id: Iacb386c3cd7b4d84f5d7e04881d58402b853a7b5
Reviewed-on: http://git-master/r/338640
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Bharat Nihalani <bnihalani@nvidia.com>
Tested-by: Bharat Nihalani <bnihalani@nvidia.com>
Reviewed-by: Krishna Reddy <vdumpa@nvidia.com>
|
|
Allocate fd's from 1024 to overcome the __FD_SETSIZE limitation
issue for select/pselect() syscalls.
Change-Id: I91df849ad5840b6037b5a46d087aab52d8527eeb
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/331978
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Robert Morell <rmorell@nvidia.com>
GVS: Gerrit_Virtual_Submit
|
|
fd to id conversion should be independent of
NVMAP_HANDLE_MARSHAL config option.
Bug 1400295
Change-Id: I3dfd9478808becf5124c20df4943d0b0e05046ab
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/329778
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Alex Waterman <alexw@nvidia.com>
GVS: Gerrit_Virtual_Submit
|
|
isolate fd creation from ref, which would allow
ref code cleanup once handles are represented with fd.
Bug 1400295
Change-Id: I0c5ea3783c191983e2c23595ace001f60cb4b600
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/328469
Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit
|
|
Don't marshal/unmarshal id when COMPAT is
not enabled or USE_FD_FOR_HANDLE is enabled.
This is necessary for supporting use of Fd's for
handles.
Change-Id: I59dd3fdfbc3380064c27a762bcb1146c1b1267bc
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/301398
|
|
This reverts commit 2bb4316f73d285f23ff74f0dcbfa397248915d1d.
Change-Id: I2aa826738c331fa2cdf5ed0b1cadafb53b63f3dc
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: http://git-master/r/329039
GVS: Gerrit_Virtual_Submit
Reviewed-by: Krishna Reddy <vdumpa@nvidia.com>
|
|
This reverts commit 42000f34bb18fc5cc36089003fa32848b8e82ed1.
Change-Id: I3f08b7c4eceb519663d1e4ec833f586660d4e354
Reviewed-on: http://git-master/r/328096
Reviewed-by: Krishna Reddy <vdumpa@nvidia.com>
Tested-by: Krishna Reddy <vdumpa@nvidia.com>
|
|
This commit implements nvmap wrappers to the CPA API that will
only call CPA functions if CONFIG_NVMAP_CPA is enabled.
Bug 1393434
Change-Id: Ia3a020b9cc971accd661df2ad81fe40451c33355
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: http://git-master/r/326266
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Krishna Reddy <vdumpa@nvidia.com>
|
|
hande fd to id failure correct.
Change-Id: I0b3b201579dc0f010f67231555c2c1a294a79039
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/294248
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Jon Mayo <jmayo@nvidia.com>
|
|
convert nvmap_create_client/_dmabuf_export_from_ref/_ref_to_id,
nvmap_pin/_unpin into private API.
remove obsolete API nvmap_ref_to_user_id/_get_addr_from_user_id
as well.
Change-Id: I18025307d9d819d6c19df6dc558a03febd6fef14
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
(cherry picked from commit 0d0c64fc93436c0671672d2a6b55d64b55db8b50)
Signed-off-by: Ajay Nandakumar <anandakumarm@nvidia.com>
|
|
support to return Fd's to user space instead of kernel virtual
address.
Change-Id: I25f4d883b88ccd249889eeacab99d0c71c25ae81
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/271085
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Alex Waterman <alexw@nvidia.com>
(cherry picked from commit c9d5edf8c7a72313c117fd5eb2cb57e727ba3986)
Signed-off-by: Ajay Nandakumar <anandakumarm@nvidia.com>
|
|
remove unused iovmm header file.
Change-Id: Ib5bcaf238db61fe5dddba0cbb518fbc89499fa5e
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/278416
(cherry picked from commit 50dd50c8d031e8f54703b7e40d39afb71dd3a756)
Signed-off-by: Ajay Nandakumar <anandakumarm@nvidia.com>
|
|
marshal/unmarshal id differently to allow using
Fd's to represent NvMap handles.
Change-Id: I86d0391132700b7c493fd20e48d752a82a658039
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/271081
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Alex Waterman <alexw@nvidia.com>
(cherry picked from commit d8038011f1cf34abe28fb539acfead23843364d0)
Signed-off-by: Ajay Nandakumar <anandakumarm@nvidia.com>
|
|
disable deferred cache maint, to avoid unforseen issues, once sgt is
get on handle.
Change-Id: I086fdceb92126a6cad7223ec6ca8b5de9f74ebaf
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/276121
Reviewed-by: Alex Waterman <alexw@nvidia.com>
|
|
Replace nvmap's pin/unpin code with calls to the DMA API. All the
currently supported operations are still supported and have
identical functionality.
Bug 1356091
Change-Id: Ibd04c27e4c46702ca90ab7ff7387420f3127fdb5
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/266236
Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit
|
|
Export nvmap_ref_to_id() so that external modules can use this
API. This is needed for in kernel unit testing of nvmap.
Change-Id: I342a2f8b8d31251ca853a3efd0647e897408e811
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: http://git-master/r/269694
Reviewed-by: Krishna Reddy <vdumpa@nvidia.com>
Tested-by: Krishna Reddy <vdumpa@nvidia.com>
|
|
Rename _nvmap_validate_id_locked() to __nvmap_validate_id_locked().
Change-Id: I7d781d4538015e8d1974bb6104e22f41c8d5cf00
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: http://git-master/r/267558
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Krishna Reddy <vdumpa@nvidia.com>
|
|
rename cache_maint function to __nvmap_cache_maint and
expose it for other nvmap files.
Change-Id: Ieb12eb2c38dc3d226b10f5df71eea50c9d0bfa8e
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/268086
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Jon Mayo <jmayo@nvidia.com>
GVS: Gerrit_Virtual_Submit
|
|
Change-Id: Ida67dd93f145d9f75816816660f968549152c1a6
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/267549
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Alex Waterman <alexw@nvidia.com>
Reviewed-by: Jon Mayo <jmayo@nvidia.com>
|
|
prepare for dmabuf support.
Bug 1290869
Change-Id: I791f341ea20b4e90082a3199844919d176064c68
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/264592
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Alex Waterman <alexw@nvidia.com>
|
|
An access_ok() check does not allow dereferencing usermode pointers
directly, __get_user must be used.
Bug 1352454
Change-Id: I762b1976d35361cb4fba58d762073594e2187493
Signed-off-by: Tuomas Tynkkynen <ttynkkynen@nvidia.com>
Reviewed-on: http://git-master/r/264006
Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit
Reviewed-by: Krishna Reddy <vdumpa@nvidia.com>
|
|
fix incorrect error handling in nvmap_share_dmabuf.
remove unnecessary dependency on DMA_SHARED_BUFFER.
cleanup nvmap dmabuf implementation.
Change-Id: If427fb8c29116b797e7e998e1c68ff0081acc559
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/244907
(cherry picked from commit ebda92705d3f7925e9edfbda7b715f33d2d93fcf)
Reviewed-on: http://git-master/r/251670
Reviewed-by: Mandar Padmawar <mpadmawar@nvidia.com>
Tested-by: Mandar Padmawar <mpadmawar@nvidia.com>
|
|
Introduce a 64-bit safe variant of nvmap_handle_get_param_u32().
Bug 1174439
Change-Id: Ia88d8dea830a2fcb9f8d1e51c8912c2a856c60d7
Signed-off-by: Terje Bergstrom <tbergstrom@nvidia.com>
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/253302
Reviewed-by: Automatic_Commit_Validation_User
|
|
move the content to nvmap_priv.h
Change-Id: I6614e357f64ec19cbb95479697c48c0989a33a5c
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/244556
(cherry picked from commit 866854d9c04c6a13525d4483d25ef71663e0f973)
Reviewed-on: http://git-master/r/249527
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Alex Waterman <alexw@nvidia.com>
|
|
Original merge commit:
commit 001997d7c06bdf9101ef93fce3f266074df97a25
Merge commit '09c210a929e4dae8a4af96f677a576ccbb8a650a' into sifted/04
Change-Id: I1b02b0407348411eb462034322168695ca17ed45
Signed-off-by: Alex Van Brunt <avanbrunt@nvidia.com>
Signed-off-by: Dan Willemsen <dwillemsen@nvidia.com>
|
|
This ioctl extends the existing alloc handle interface
to allow for "kind" specification on Tegra12x or later
SoCs.
Change-Id: Ic2918b1102d672185847711be12bc1f6b57d1f65
Signed-off-by: Ken Adams <kadams@nvidia.com>
Reviewed-on: http://git-master/r/144129
Reviewed-by: Jin Qian <jqian@nvidia.com>
Reviewed-by: Jeff Smith <jsmith@nvidia.com>
|
|
and the include for host1x init (invoked from board file)
Change-Id: I5a7cb2e074f6c7395aec5ede1db31b2bdeae5cb0
Reviewed-by: Scott Williams <scwilliams@nvidia.com>
Reviewed-by: Mark Stadler <mastadler@nvidia.com>
Signed-off-by: Jin Qian <jqian@nvidia.com>
Reviewed-on: http://git-master/r/82938
Signed-off-by: Mark Stadler <mastadler@nvidia.com>
|
|
update code to perform handle marshalling based on config option
Change-Id: If7dd8a20088dfa9d593f799ec9ceab006b4a6880
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/235525
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Alex Waterman <alexw@nvidia.com>
Reviewed-by: Bo Yan <byan@nvidia.com>
|
|
detect invalid pointers passed to nvmap and avoid crash in nvmap.
generate warnings for API's that don't return error, to ease
client code debug.
Change-Id: Ia712dd0a535f3b3ca0aebe43f92b7e737a930255
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/234328
Reviewed-by: Automatic_Commit_Validation_User
|
|
Add nvhost private data field and setter and getter for for it.
Bug 1259839
Change-Id: I618046e3205bc1db548fb8f4fefb4be1571ceeb4
Signed-off-by: Terje Bergstrom <tbergstrom@nvidia.com>
Reviewed-on: http://git-master/r/216493
Reviewed-by: Simone Willett <swillett@nvidia.com>
Tested-by: Simone Willett <swillett@nvidia.com>
|
|
rename private nvmap.h to nvmap_priv.h
remove dependency on nvmap_priv.h from trace nvmap.h
Change-Id: I8dc363ec66a09d84f5e32c22228cca23a44e6c66
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/232251
Reviewed-by: Simone Willett <swillett@nvidia.com>
Tested-by: Simone Willett <swillett@nvidia.com>
|
|
avoid using nvmap_handle pointer in client code.
abstract nvmap_handle_ref definition from clients.
update client code to use correct nvmap API.
Change-Id: Ibcbf22d8a6cc3160baaa40f87bbc5677938bbe92
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/230012
Reviewed-by: Mrutyunjay Sawant <msawant@nvidia.com>
Tested-by: Mrutyunjay Sawant <msawant@nvidia.com>
|
|
sysmem support is no longer necessary in nvmap
Bug 1272328
Change-Id: Ic646e1318ac68ee2ddf26bc5b2233cd5540d2fc2
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/228550
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Kaz Fukuoka <kfukuoka@nvidia.com>
|
|
The pte's setup by nvmap during cache_maint and read/write operations
doesn't need to send IPI as a part of arm errata 798181. Use the new api
which avoids the IPI and overhead associated with it.
Bug 1272231
Change-Id: I9d65763a3948470893d250f16aef0c8173ee4350
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/224599
(cherry picked from commit 0e015b031e6eb941a5e4d2af44f07ba044b31b71)
Reviewed-on: http://git-master/r/226781
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Bo Yan <byan@nvidia.com>
GVS: Gerrit_Virtual_Submit
|
|
fix start address calculation issue for split vma's during cache maint.
fix offset and length validation during map into caller pointer.
Bug 1032750
Change-Id: Ia8da8ec8dc84f7b2f873416602f563ba285c5b25
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/223207
Reviewed-by: Kaz Fukuoka <kfukuoka@nvidia.com>
Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit
|
|
Bug 1228120
Change-Id: I3f8066270e25be195ef7a26d64198be326ac35b3
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/204816
Reviewed-by: Automatic_Commit_Validation_User
|
|
Bug 1228120
Change-Id: I5bdeed41b32f2d7d6f4c5250b87a36cded6b54e3
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/204782
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Jon Mayo <jmayo@nvidia.com>
|