linux/drivers, branch v2.6.35.11

revert-drm-radeon-kms-properly-compute-group_size-on-6xx-7xx

2011-02-06T19:04:06Z

Revert drm/radeon/kms: properly compute group_size on 6xx/7xx From: Steve Conklin We discovered a regression for Radeon users in our latest proposed kernel for 2.6.35 (Maverick), and have isolated it to this patch: http://git.kernel.org/?p=linux/kernel/git/longterm/linux-2.6.35.y.git;a=commit;h=b8e9a4a45f8427837f4dba89 +bda4d4e3f3a5c726 We took that patch as part of 2.6.35.10, and one of our testers has reported that our build of that kernel also exhibits the problem. These are mainline kernels built with the Ubuntu configs. http://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.35.10-maverick/ Our bug report is here: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/703553 Upstream bug report: https://bugzilla.kernel.org/show_bug.cgi?id=24802 Signed-off-by: Andi Kleen

Input: i8042 - introduce 'notimeout' blacklist for Dell Vostro V13

2011-02-06T19:03:53Z

i8042 controller present in Dell Vostro V13 errorneously signals spurious timeouts. Introduce i8042.notimeout parameter for ignoring i8042-signalled timeouts and apply this quirk automatically for Dell Vostro V13, based on DMI match. In addition to that, this machine also needs to be added to nomux blacklist. Signed-off-by: Jiri Kosina Signed-off-by: Dmitry Torokhov Signed-off-by: Andi Kleen

gspca - sonixj: Add a flag in the driver_info table

2011-02-06T19:03:53Z

commit c6c14330717f9850b4b4c054b81424b9979cd07d upstream. Signed-off-by: Jean-François Moine Signed-off-by: Mauro Carvalho Chehab Signed-off-by: Greg Kroah-Hartman Signed-off-by: Andi Kleen

gspca - sonixj: Set the flag for some devices

2011-02-06T19:03:52Z

commit b2272a49e7df37732d73988f00468ce31e1ebc92 upstream. The flag PDN_INV indicates that the sensor pin S_PWR_DN has not the same value as other webcams with the same sensor. For now, only two webcams have been so detected: the Microsoft's VX1000 and VX3000. Signed-off-by: Jean-François Moine Signed-off-by: Mauro Carvalho Chehab Signed-off-by: Greg Kroah-Hartman Signed-off-by: Andi Kleen

block: Deprecate QUEUE_FLAG_CLUSTER and use queue_limits instead

2011-02-06T19:03:52Z

commit e692cb668fdd5a712c6ed2a2d6f2a36ee83997b4 upstream. When stacking devices, a request_queue is not always available. This forced us to have a no_cluster flag in the queue_limits that could be used as a carrier until the request_queue had been set up for a metadevice. There were several problems with that approach. First of all it was up to the stacking device to remember to set queue flag after stacking had completed. Also, the queue flag and the queue limits had to be kept in sync at all times. We got that wrong, which could lead to us issuing commands that went beyond the max scatterlist limit set by the driver. The proper fix is to avoid having two flags for tracking the same thing. We deprecate QUEUE_FLAG_CLUSTER and use the queue limit directly in the block layer merging functions. The queue_limit 'no_cluster' is turned into 'cluster' to avoid double negatives and to ease stacking. Clustering defaults to being enabled as before. The queue flag logic is removed from the stacking function, and explicitly setting the cluster flag is no longer necessary in DM and MD. Reported-by: Ed Lin Signed-off-by: Martin K. Petersen Acked-by: Mike Snitzer Signed-off-by: Jens Axboe Signed-off-by: Greg Kroah-Hartman Signed-off-by: Andi Kleen

libata-sff: fix HSM_ST_ERR handling in __ata_sff_port_intr()

2011-02-06T19:03:51Z

commit 687a993339c4f3a63654746230da3aab8bbdbffd upstream. While separating out BMDMA irq handler from SFF, commit c3b28894 (libata-sff: separate out BMDMA irq handler) incorrectly made __ata_sff_port_intr() consider an IRQ to be an idle one if the host state was transitioned to HSM_ST_ERR by ata_bmdma_port_intr(). This makes BMDMA drivers ignore IRQs reporting host bus error which leads to timeouts instead of triggering EH immediately. Fix it by making __ata_sff_port_intr() consider the IRQ to be an idle one iff the state is HSM_ST_IDLE. This is equivalent to adding HSM_ST_ERR to the "break"ing case but less error-prone. Signed-off-by: Tejun Heo Signed-off-by: Andi Kleen Reported-by: Antonio Toma Signed-off-by: Jeff Garzik Signed-off-by: Greg Kroah-Hartman

mv_xor: fix race in tasklet function

2011-02-06T19:03:50Z

commit 8333f65ef094e47020cd01452b4637e7daf5a77f upstream. use mv_xor_slot_cleanup() instead of __mv_xor_slot_cleanup() as the former function aquires the spin lock that needed to protect the drivers data. Signed-off-by: Saeed Bishara Signed-off-by: Dan Williams Signed-off-by: Greg Kroah-Hartman Signed-off-by: Andi Kleen

drm/i915/dp: Fix I2C/EDID handling with active DisplayPort to DVI converter

2011-02-06T19:03:50Z

commit 8316f33766a82907c694267ff911e45e256f09f9 upstream. The DisplayPort standard (1.1a) states that: The I2C-over-AUX Reply field is valid only when Native AUX CH Reply field is AUX_ACK (00). When Native AUX CH Reply field is not 00, then, I2C-over-AUX Reply field must be 00 and be ignored. This fixes broken EDID reading when using an active DisplayPort to duallink DVI converter. If the AUX CH replier chooses to defer the transaction, a short read occurs and erroneous data is returned as the i2c reply due to a lack of length checking and failure to check for AUX ACK. As a result, broken EDIDs can look like: 0 1 2 3 4 5 6 7 8 9 a b c d e f 0123456789abcdef 00: bc bc bc ff bc bc bc ff bc bc bc ac bc bc bc 45 ???.???.???????E 10: bc bc bc 10 bc bc bc 34 bc bc bc ee bc bc bc 4c ???????4???????L 20: bc bc bc 50 bc bc bc 00 bc bc bc 40 bc bc bc 00 ???P???.???@???. 30: bc bc bc 01 bc bc bc 01 bc bc bc a0 bc bc bc 40 ???????????????@ 40: bc bc bc 00 bc bc bc 00 bc bc bc 00 bc bc bc 55 ???.???.???.???U 50: bc bc bc 35 bc bc bc 31 bc bc bc 20 bc bc bc fc ???5???1??? ???? 60: bc bc bc 4c bc bc bc 34 bc bc bc 46 bc bc bc 00 ???L???4???F???. 70: bc bc bc 38 bc bc bc 11 bc bc bc 20 bc bc bc 20 ???8??????? ??? 80: bc bc bc ff bc bc bc ff bc bc bc ff bc bc bc ff ???.???.???.???. ... which can lead to: [drm:drm_edid_block_valid] *ERROR* EDID checksum is invalid, remainder [drm:drm_edid_block_valid] *ERROR* Raw EDID: <3>30 30 30 30 30 30 30 32 38 32 30 32 63 63 31 61 000000028202cc1a <3>28 00 02 8c 00 00 00 00 18 00 00 00 00 00 00 00 (............... <3>20 4c 61 73 74 20 62 65 61 63 6f 6e 3a 20 33 32 Last beacon: 32 <3>32 30 6d 73 20 61 67 6f 46 00 05 8c 00 00 00 00 20ms agoF....... <3>36 00 00 00 00 00 00 00 00 0c 57 69 2d 46 69 20 6.........Wi-Fi <3>52 6f 75 74 65 72 01 08 82 84 8b 96 24 30 48 6c Router......$0Hl <3>03 01 01 06 02 00 00 2a 01 00 2f 01 00 32 04 0c .......*../..2.. <3>12 18 60 dd 09 00 10 18 02 00 00 01 00 00 18 00 ..`............. Signed-off-by: David Flynn Signed-off-by: Andi Kleen [ickle: fix up some surrounding checkpatch warnings] Signed-off-by: Chris Wilson Signed-off-by: Greg Kroah-Hartman

drm/radeon/kms: reorder display resume to avoid problems

2011-02-06T19:03:50Z

commit a93f344d3c04e4b84490c65f2a574387c593be40 upstream. On resume, we were attemping to unblank the displays before the timing and plls had be reprogrammed which led to atom timeouts waiting for things that are not yet programmed. Re-program the mode first, then reset the dpms state. This fixes the infamous atombios timeouts on resume. Signed-off-by: Alex Deucher Signed-off-by: Dave Airlie Signed-off-by: Greg Kroah-Hartman Signed-off-by: Andi Kleen

drm/radeon/kms: fix evergreen asic reset

2011-02-06T19:03:50Z

commit 9f0c4f9c2f835eee1bbb93f96bf9483d56f1892b upstream. Only reset the grbm blocks, srbm tends to lock the GPU if not done properly and in most cases is not necessary. Also, no need to call asic init after reset the grbm blocks. Signed-off-by: Alex Deucher Signed-off-by: Andi Kleen Reviewed-by: Jerome Glisse Signed-off-by: Dave Airlie Signed-off-by: Greg Kroah-Hartman