Displaying 4 results from an estimated 4 matches for "nv50_software_context_ctor".
2013 Aug 27
0
[PATCH 5/9] drm/nouveau: Add install/remove semantics for event handlers
...are_chan *chan =
- container_of(event, struct nouveau_software_chan, vblank.event);
+ container_of(event, struct nouveau_software_chan, vblank.event[head]);
struct nv50_software_priv *priv = (void *)nv_object(chan)->engine;
struct nouveau_bar *bar = nouveau_bar(priv);
@@ -161,7 +161,8 @@ nv50_software_context_ctor(struct nouveau_object *parent,
struct nouveau_object **pobject)
{
struct nv50_software_chan *chan;
- int ret;
+ struct nouveau_disp *disp = nouveau_disp(engine);
+ int ret, i;
ret = nouveau_software_context_create(parent, engine, oclass, &chan);
*pobject = nv_object(chan);
@@ -1...
2013 Nov 25
0
[Bug 71824] [NVE6] NULL deref on boot when there is nothing in DCB on 3.13-rc
...oting |when there is nothing in
|with Call Trace |DCB on 3.13-rc
--- Comment #6 from Ilia Mirkin <imirkin at alum.mit.edu> ---
Looks like you have no outputs. I think this will lead to
chan->vblank.nr_event = pdisp->vblank->index_nr;
failing in nv50_software_context_ctor, called from nouveau_accel_init since
pdisp->vblank == NULL. I suspect that ->index_nr is at offset 0x130 which
explains the CR2 that you see.
As a temporary workaround, booting with nouveau.noaccel=1 will avoid the crash
(and also remove your ability to use the card).
--
You are receiving...
2013 Oct 09
0
Null pointer deref in nv50.c after 2fb9d6cf
After 2fb9d6cf (drm/nouveau/core: convert event handler apis to split
create/enable semantics) from the nouveau tree I'm getting a kernel
panic on boot because pdisp is 0 at core/engine/software/nv50.c:179.
The nouveau part of the call trace is:
nv50_software_context_ctor+0x80/0x130
nouveau_object_ctor+0x2d/0xe0
nouveau_object_new+0xfa/0x230
nouveau_channel_new+0x53b/0x6c0
nouveau_drm_load+0x436/0x6c0
...
nouveau_drm_probe+0x202/0x290
...
nouveau_drm_init+0x43/0x1000
Is there anything else you need?
Martin
$ lspci
00:00.0 Host bridge: Intel Corporation 3rd Gen...
2013 Aug 27
11
[PATCH 0/9] drm/nouveau: Cleanup event/handler design
This series was originally motivated by a deadlock, introduced in
commit 1d7c71a3e2f77336df536855b0efd2dc5bdeb41b
'drm/nouveau/disp: port vblank handling to event interface',
due to inverted lock order between nouveau_drm_vblank_enable()
and nouveau_drm_vblank_handler() (the complete
lockdep report is included in the patch 4/5 changelog).
Because this series fixes the vblank event