linux-toradex.git/drivers/gpu/drm/via, branch v3.4.73

drm via: initialize object_idr

2012-06-22T18:36:53+00:00

commit ce020ea53264f1460ae619cfc12f968dbd0b8974 upstream.

The field obejct_idr of struct drm_via_private was introduced with the
commit http://git.kernel.org/?p=linux/kernel/git/torvalds/linux.git;a=commitdiff;h=77ee8f3825054f23b17e9c8f728f061defd86cdc .
In that patch idr_init(&dev->object_name_idr) was called instead of
idr_init(&dev_priv->object_idr) by mistake, leaving the dev_priv->object_idr
uninitialized. To be more exact, the object_idr buffer is filled with zeros
because of kzalloc(), but the dev_priv->object_idr.lock spinlock can cause
system freeze at lib/idr.c:move_to_free_list() when spin_lock_irqsave()
is called on this spinlock.

The patch was tested on Clevo D4J, model D410J laptop, on the following
hardware, without AGP kernel module loaded:

  # lspci -s 01:00.0 -n
  01:00.0 0300: 1106:3108 (rev 01)
  # lspci -s 01:00.0 -v
  01:00.0 VGA compatible controller: VIA Technologies, Inc. K8M800/K8N800/K8N800A [S3 UniChrome Pro] (rev 01) (prog-if 00 [VGA controller])
          Subsystem: CLEVO/KAPOK Computer Device 4702
          Flags: bus master, 66MHz, medium devsel, latency 64, IRQ 16
          Memory at f0000000 (32-bit, prefetchable) [size=64M]
          Memory at d1000000 (32-bit, non-prefetchable) [size=16M]
          Expansion ROM at  [disabled]
          Capabilities: [60] Power Management version 2
          Capabilities: [70] AGP version 3.0

Signed-off-by: Márton Németh 
Reviewed-by: Daniel Vetter 
Signed-off-by: Dave Airlie 
Signed-off-by: Greg Kroah-Hartman

drm: move pci bus master enable into driver.

2012-02-16T18:31:07+00:00

The current enabling of bus mastering in the drm midlayer allows a large
race condition under kexec. When a kexec'ed kernel re-enables bus mastering
for the GPU, previously setup dma blocks may cause writes to random pieces
of memory. On radeon the writeback mechanism can cause these sorts of issues.

This patch doesn't fix the problem, but it moves the bus master enable under
the individual drivers control so they can move enabling it until later in
their load cycle and close the race.

Fix for radeon kms driver will be in a follow-up patch.

Signed-off-by: Dave Airlie

drm/sis|via: don't return stack garbage from free_mem ioctl

2012-01-09T12:11:39+00:00

Fallout from my "kill drm_sman" refactor. Unfortunately gcc seems to
have failed me and not warned about this.

Tested-by: Jakob Bornecrantz  (on via)
Signed-off-by: Daniel Vetter 
Signed-off-by: Dave Airlie

drm/via: use drm_mm instead of drm_sman

2011-12-21T23:33:22+00:00

To make the transition in a piece-wise and bisectable way possible,
I've hijacked the ->owner_list from drm_sman. While transitioning, the
list_add was done by the driver, while the list_del was still done by
the dying sman code.

Now that we are in full control of ->owner_list, do the list_del
ourselves.

v2: Better explain the list_del trickery as suggested by Chris Wilson.

Signed-off-by: Daniel Vetter

drm/sis: track user->memblock mapping with idr

2011-12-21T23:33:21+00:00

Signed-off-by: Daniel Vetter

drm/via: track user->memblock mapping with idr

2011-12-21T23:33:20+00:00

Massive indirection through a hashtable for a simple key->pointer
look-up actually just adds bloat.

v2: Drop the misleading comment noted by Chris Wilson.

Signed-off-by: Daniel Vetter

drm/sman: rip out owner tracking

2011-12-21T23:33:20+00:00

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

drm/via: track obj->drm_fd relations in the driver

2011-12-21T23:33:19+00:00

Exactly like the previous patch for sis.

Signed-off-by: Daniel Vetter

drm: Make the per-driver file_operations struct const

2011-11-11T11:14:47+00:00

From fdf1fdebaa00f81de18c227f32f8074c8b352d50 Mon Sep 17 00:00:00 2001
From: Arjan van de Ven 
Date: Sun, 30 Oct 2011 19:06:07 -0700
Subject: [PATCH] drm: Make the per-driver file_operations struct const

The DRM layer keeps a copy of struct file_operations inside its
big driver struct... which prevents it from being consistent and static.
For consistency (and the general security objective of having such things
static), it's desirable to get this fixed.

This patch splits out the file_operations field to its own struct,
which is then "static const", and just stick a pointer to this into
the driver struct, making it more consistent with how the rest of the
kernel does this.

Signed-off-by: Arjan van de Ven 
Signed-off-by: Dave Airlie

gpu: add module.h to drivers/gpu files as required.

2011-10-31T23:32:03+00:00

So that we don't get build failures once the implicit module.h
presence is removed.

Signed-off-by: Paul Gortmaker