linux-toradex.git/drivers/video, branch tegra-10.9.7 Linux kernel for Apalis and Colibri modules video: tegra: nvmap: Clean whole L1 instead of VA cleaning 2010-11-24T23:18:45+00:00 vdumpa vdumpa@nvidia.com 2010-10-29T18:37:42+00:00 05074ac6ce18a61467ad6d5dd5e53d56725c8d58 For large allocations, cleaning each page of the allocation can take a significant amount of time. If an allocation that nvmap needs to clean or invalidate out of the cache is significantly larger than the cache, just flush the entire cache. bug 711478 and bug 744221 Revert "video: tegra: nvmap: perform cache maintenance for rw_handle" This reverts commit d963f09c3ebb2d690d266f8f607b4876acaf2ab1. Reviewed-on: http://git-master/r/10213 (cherry picked from commit 6469f378a596f0572035cd27a17851ea86b763c9) Change-Id: I1e514f505db860eb01d575a98f0f80c8794e8463 Reviewed-on: http://git-master/r/10497 Reviewed-by: Eric Werness <ewerness@nvidia.com> Reviewed-by: Markus Holtmanns <mholtmanns@nvidia.com> Tested-by: Markus Holtmanns <mholtmanns@nvidia.com> Reviewed-by: Janne Hellsten <jhellsten@nvidia.com> Reviewed-by: Varun Colbert <vcolbert@nvidia.com> Tested-by: Varun Colbert <vcolbert@nvidia.com> 
For large allocations, cleaning each page of the allocation can
take a significant amount of time. If an allocation that nvmap needs
to clean or invalidate out of the cache is significantly larger than
the cache, just flush the entire cache.

bug 711478 and bug 744221

Revert "video: tegra: nvmap: perform cache maintenance for rw_handle"

This reverts commit d963f09c3ebb2d690d266f8f607b4876acaf2ab1.

Reviewed-on: http://git-master/r/10213
(cherry picked from commit 6469f378a596f0572035cd27a17851ea86b763c9)

Change-Id: I1e514f505db860eb01d575a98f0f80c8794e8463
Reviewed-on: http://git-master/r/10497
Reviewed-by: Eric Werness <ewerness@nvidia.com>
Reviewed-by: Markus Holtmanns <mholtmanns@nvidia.com>
Tested-by: Markus Holtmanns <mholtmanns@nvidia.com>
Reviewed-by: Janne Hellsten <jhellsten@nvidia.com>
Reviewed-by: Varun Colbert <vcolbert@nvidia.com>
Tested-by: Varun Colbert <vcolbert@nvidia.com>
nvmap: Improved compaction algorithm implemented 2010-11-06T18:05:36+00:00 Kirill Artamonov kartamonov@nvidia.com 2010-10-15T09:42:58+00:00 86350e7a8c6757f4887ddc5b973b406af1eb9d03 Fast compaction algorithm: for every free block from the beginning of CO, try to reallocate block before and after the free block. Once enough free space for requested allocation is freed, return. Full compaction triggers if fast one fails: it works like fast one but also uses block moving algorithm which doesn't require extra CO memory. It compacts all heap, merging as much free space as possible. Also fixed few bugs. bug 740842 bug 746972 (cherry picked from commit 171edd17a191c8a4a4f9a8434dec796089cda141) Change-Id: I40ec2c49a3c3f37fd7149e8123eda59da4caef39 Reviewed-on: http://git-master/r/10149 Reviewed-by: Kirill Artamonov <kartamonov@nvidia.com> Tested-by: Kirill Artamonov <kartamonov@nvidia.com> Reviewed-by: Bharat Nihalani <bnihalani@nvidia.com> 
Fast compaction algorithm: for every free block from the
beginning of CO, try to reallocate block before
and after the free block. Once enough free space for
requested allocation is freed, return.

Full compaction triggers if fast one fails: it works like
fast one but also uses block moving algorithm which
doesn't require extra CO memory. It compacts all heap,
merging as much free space as possible.

Also fixed few bugs.

bug 740842
bug 746972

(cherry picked from commit 171edd17a191c8a4a4f9a8434dec796089cda141)

Change-Id: I40ec2c49a3c3f37fd7149e8123eda59da4caef39
Reviewed-on: http://git-master/r/10149
Reviewed-by: Kirill Artamonov <kartamonov@nvidia.com>
Tested-by: Kirill Artamonov <kartamonov@nvidia.com>
Reviewed-by: Bharat Nihalani <bnihalani@nvidia.com>
[ARM/Tegra] nvmap fix for crash in spin_lock() call 2010-11-06T18:05:14+00:00 Thomas Roell troell@nvidia.com 2010-10-13T21:23:13+00:00 c970c075df97de1e48fa80461e2e52e7446a666f Gate access to h->carveout.co_heap by (h->alloc && !h->pg_alloc) Bug 743185 (cherry picked from commit daba1dc0a3e050a2739293904cad9e65eb27a559) Change-Id: I5014fb40074e79e4db293767e3ef512ea35401d8 Reviewed-on: http://git-master/r/10148 Reviewed-by: Kirill Artamonov <kartamonov@nvidia.com> Tested-by: Kirill Artamonov <kartamonov@nvidia.com> Reviewed-by: Bharat Nihalani <bnihalani@nvidia.com> 
Gate access to h->carveout.co_heap by (h->alloc && !h->pg_alloc)

Bug 743185

(cherry picked from commit daba1dc0a3e050a2739293904cad9e65eb27a559)

Change-Id: I5014fb40074e79e4db293767e3ef512ea35401d8
Reviewed-on: http://git-master/r/10148
Reviewed-by: Kirill Artamonov <kartamonov@nvidia.com>
Tested-by: Kirill Artamonov <kartamonov@nvidia.com>
Reviewed-by: Bharat Nihalani <bnihalani@nvidia.com>
nvmap: various fixes for bug 743185 2010-10-26T00:57:04+00:00 Venkata (Muni) Anda vanda@nvidia.com 2010-10-09T19:28:21+00:00 17d2d531f3f0b8f0b2dc24fba3ccd02e9dcf7210 Fixed the following: Missing node creation for debug_fs while doing compaction. Missing locks in handle_free Split heap functionality was not creating debugfs node Misc code clean-up. roll up of unmerged changes from other branch Bug 743185 (Cherry picked from commit 396c587981b2d0d1abb1daf481a5fe4bbd0ba70d) Change-Id: I55249e14ea69c3780b24efbfaea9ceef9ef23534 Reviewed-on: http://git-master/r/8236 Reviewed-by: Varun Colbert <vcolbert@nvidia.com> Reviewed-by: Thomas Roell <troell@nvidia.com> Tested-by: Varun Colbert <vcolbert@nvidia.com> 
Fixed the following:
Missing node creation for debug_fs while doing compaction.
Missing locks in handle_free
Split heap functionality was not creating debugfs node

Misc code clean-up.
roll up of unmerged changes from other branch

Bug 743185

(Cherry picked from commit 396c587981b2d0d1abb1daf481a5fe4bbd0ba70d)

Change-Id: I55249e14ea69c3780b24efbfaea9ceef9ef23534
Reviewed-on: http://git-master/r/8236
Reviewed-by: Varun Colbert <vcolbert@nvidia.com>
Reviewed-by: Thomas Roell <troell@nvidia.com>
Tested-by: Varun Colbert <vcolbert@nvidia.com>
[ARM/tegra] clocks: Enabled clock before set rate. 2010-10-20T00:10:34+00:00 Alex Frid afrid@nvidia.com 2010-10-09T06:03:42+00:00 b3c5b3ae3c492dcfc49dc1a7265b801398565d3b Made sure graphics and sdhci clocks are enabled before the respective rate is configured. Change-Id: I2d09ad111321b296a8d8fbb2d4bff02016e12feb Reviewed-on: http://git-master/r/8272 Reviewed-by: Varun Colbert <vcolbert@nvidia.com> Tested-by: Varun Colbert <vcolbert@nvidia.com> 
Made sure graphics and sdhci clocks are enabled before the respective
rate is configured.

Change-Id: I2d09ad111321b296a8d8fbb2d4bff02016e12feb
Reviewed-on: http://git-master/r/8272
Reviewed-by: Varun Colbert <vcolbert@nvidia.com>
Tested-by: Varun Colbert <vcolbert@nvidia.com>
[ARM/tegra] nvhost: Tuned ACM timeout. 2010-10-16T05:03:35+00:00 Alex Frid afrid@nvidia.com 2010-10-01T06:58:09+00:00 8ec98051794908a84024cc640330a1c5abd375ea Drastic reduction of ACM timeout to 2ms (from 1sec) resulted in power increase for NV omxplayer video-playback (bug 738557). The reason is that short-term scheduling does not provide enough idle time for LP2 state (CPU powered down). On the other hand, 1 sec timeout prevented core voltage scaling during Android music audio-playback (bug 735111). Hence, tuned ACM timeout to 25ms. (cherry picked from commit 32c727ae2c4f91e8d7c9c261921bb62ad14217c9) Change-Id: I96c46c43e9b3089b6a7e01db5f6d122122758c90 Reviewed-on: http://git-master.nvidia.com/r/8470 Tested-by: Aleksandr Frid <afrid@nvidia.com> Reviewed-by: Hoang Pham <hopham@nvidia.com> Reviewed-by: Yu-Huan Hsu <yhsu@nvidia.com> Reviewed-by: Varun Colbert <vcolbert@nvidia.com> Tested-by: Varun Colbert <vcolbert@nvidia.com> 
Drastic reduction of ACM timeout to 2ms (from 1sec) resulted in power
increase for NV omxplayer video-playback (bug 738557). The reason is
that short-term scheduling does not provide enough idle time for LP2
state (CPU powered down). On the other hand, 1 sec timeout prevented
core voltage scaling during Android music audio-playback (bug 735111).
Hence, tuned ACM timeout to 25ms.

(cherry picked from commit 32c727ae2c4f91e8d7c9c261921bb62ad14217c9)

Change-Id: I96c46c43e9b3089b6a7e01db5f6d122122758c90
Reviewed-on: http://git-master.nvidia.com/r/8470
Tested-by: Aleksandr Frid <afrid@nvidia.com>
Reviewed-by: Hoang Pham <hopham@nvidia.com>
Reviewed-by: Yu-Huan Hsu <yhsu@nvidia.com>
Reviewed-by: Varun Colbert <vcolbert@nvidia.com>
Tested-by: Varun Colbert <vcolbert@nvidia.com>
[ARM/Tegra] nvhost : unpin context memory handle before freeing 2010-10-15T22:51:07+00:00 Andrew Wright awright@nvidia.com 2010-10-13T00:23:43+00:00 80e2d4752c5f291d8a63d537f89be14e25a4626c Bug 731923 Reviewed-on: http://git-master/r/7002 (cherry picked from commit b927a389930941297f987cc7967d10d319deba6b) Change-Id: I576de9b2dfd4245f471ca26636a146ead3df021b Reviewed-on: http://git-master.nvidia.com/r/8430 Reviewed-by: Yu-Huan Hsu <yhsu@nvidia.com> Reviewed-by: Venkata (Muni) Anda <vanda@nvidia.com> Tested-by: Andrew Wright <awright@nvidia.com> Reviewed-by: Varun Colbert <vcolbert@nvidia.com> Tested-by: Varun Colbert <vcolbert@nvidia.com> 
Bug 731923
Reviewed-on: http://git-master/r/7002
(cherry picked from commit b927a389930941297f987cc7967d10d319deba6b)

Change-Id: I576de9b2dfd4245f471ca26636a146ead3df021b
Reviewed-on: http://git-master.nvidia.com/r/8430
Reviewed-by: Yu-Huan Hsu <yhsu@nvidia.com>
Reviewed-by: Venkata (Muni) Anda <vanda@nvidia.com>
Tested-by: Andrew Wright <awright@nvidia.com>
Reviewed-by: Varun Colbert <vcolbert@nvidia.com>
Tested-by: Varun Colbert <vcolbert@nvidia.com>
[ARM/Tegra] Fix commit f3824fd58e2654a0d8d3159cb1f776d3ad0169a2 2010-10-13T21:57:08+00:00 Zhijun He zhhe@nvidia.com 2010-10-12T04:12:10+00:00 6afcd42bf2fc9e7dc3cad566865238cd71843420 Check for pgalloc is not required. Bug 731923 Reviewed-on: http://git-master/r/7218 (cherry picked from commit 665dba4dd0617f763815bce96dec4a14f8a0c6f6) Change-Id: I82e701555605f4322a95a66b75ffe3f3bce2a5a6 Reviewed-on: http://git-master.nvidia.com/r/8266 Reviewed-by: Varun Colbert <vcolbert@nvidia.com> Tested-by: Varun Colbert <vcolbert@nvidia.com> 
Check for pgalloc is not required.

Bug 731923

Reviewed-on: http://git-master/r/7218
(cherry picked from commit 665dba4dd0617f763815bce96dec4a14f8a0c6f6)

Change-Id: I82e701555605f4322a95a66b75ffe3f3bce2a5a6
Reviewed-on: http://git-master.nvidia.com/r/8266
Reviewed-by: Varun Colbert <vcolbert@nvidia.com>
Tested-by: Varun Colbert <vcolbert@nvidia.com>
[ARM/Tegra] nvmap: fix a crash in _nvmap_handle_free 2010-10-13T21:56:52+00:00 Zhijun He zhhe@nvidia.com 2010-10-12T04:04:34+00:00 0120a5b05699269c4ce6669ea1aa37b8ed2ed0f9 When the memory for a handle is not allocated, and then if the caller calls _nvmap_handle_free it should not do the cache operations on the memory as there is no memory allocated Bug 731923 Reviewed-on: http://git-master.nvidia.com/r/7166 (cherry picked from commit f3824fd58e2654a0d8d3159cb1f776d3ad0169a2) Change-Id: Icf11ece6d07e122991e0d9b6b4a3dbf8c45ea017 Reviewed-on: http://git-master.nvidia.com/r/8265 Reviewed-by: Varun Colbert <vcolbert@nvidia.com> Tested-by: Varun Colbert <vcolbert@nvidia.com> 
When the memory for a handle is not allocated, and then
if the caller calls _nvmap_handle_free it should
not do the cache operations on the memory as there
is no memory allocated

Bug 731923

Reviewed-on: http://git-master.nvidia.com/r/7166
(cherry picked from commit f3824fd58e2654a0d8d3159cb1f776d3ad0169a2)

Change-Id: Icf11ece6d07e122991e0d9b6b4a3dbf8c45ea017
Reviewed-on: http://git-master.nvidia.com/r/8265
Reviewed-by: Varun Colbert <vcolbert@nvidia.com>
Tested-by: Varun Colbert <vcolbert@nvidia.com>
video: tegra: nvhost: fix race condition in hostintr wait list 2010-10-13T21:56:37+00:00 Zhijun He zhhe@nvidia.com 2010-10-12T03:55:20+00:00 13bf45b3c548319aa411541c760516ffe1b5eaff Reviewed-on: http://git-master/r/6507 (cherry picked from commit 3708cd7a859d1ca4f3230bc8bfb018894bb6277e) Change-Id: I1609b8b8d1f2de51f7eafa2105fb1f695903a252 Reviewed-on: http://git-master.nvidia.com/r/8262 Reviewed-by: Varun Colbert <vcolbert@nvidia.com> Tested-by: Varun Colbert <vcolbert@nvidia.com> 
Reviewed-on: http://git-master/r/6507
(cherry picked from commit 3708cd7a859d1ca4f3230bc8bfb018894bb6277e)

Change-Id: I1609b8b8d1f2de51f7eafa2105fb1f695903a252
Reviewed-on: http://git-master.nvidia.com/r/8262
Reviewed-by: Varun Colbert <vcolbert@nvidia.com>
Tested-by: Varun Colbert <vcolbert@nvidia.com>