[PATCH RFC 087/111] staging: etnaviv: align command stream size to 64 bit

[Lucas Stach]

Am Donnerstag, den 02.04.2015, 17:20 +0100 schrieb Russell King - ARM
Linux:
> On Thu, Apr 02, 2015 at 05:30:29PM +0200, Lucas Stach wrote:
> > It is legal for the userspace to pass in a command stream of a size
> > aligned to 32 bit, if that is where the last user command ends. The
> > kernel then needs to insert a LINK command at the end of the stream,
> > which needs to be aligned to 64 bit, so the kernel may insert an
> > additional 32bits of padding in the stream. Align the stream size
> > to account for that in the size and command stream validator checks.
> > 
> > Signed-off-by: Lucas Stach <l.stach at pengutronix.de>
> > ---
> >  drivers/staging/etnaviv/etnaviv_gem_submit.c | 5 ++++-
> >  1 file changed, 4 insertions(+), 1 deletion(-)
> > 
> > diff --git a/drivers/staging/etnaviv/etnaviv_gem_submit.c b/drivers/staging/etnaviv/etnaviv_gem_submit.c
> > index 7bd4912ab8ad..965096be5219 100644
> > --- a/drivers/staging/etnaviv/etnaviv_gem_submit.c
> > +++ b/drivers/staging/etnaviv/etnaviv_gem_submit.c
> > @@ -377,8 +377,11 @@ int etnaviv_ioctl_gem_submit(struct drm_device *dev, void *data,
> >  
> >  		/*
> >  		 * We must have space to add a LINK command at the end of
> > -		 * the command buffer.
> > +		 * the command buffer. Align buffer size to the next 64bit
> > +		 * quantity, as that's the point where we need to insert the
> > +		 * next command.
> >  		 */
> > +		submit_cmd.size = ALIGN(submit_cmd.size, 8);
> >  		max_size = etnaviv_obj->base.size - 8;
> 
> I wonder if it's an error if the command size is not a multiple of 8?
> I know that the command stream is always aligned to 8 bytes for the 2D
> cores, but I don't know about the 3D or VG cores.
> 
The start of the commands must always be 64bit aligned, it's the same
for all pipes. The size can be dword aligned if the last command in the
stream is something like SET_STATE with a length of 2. In that case one
needs to insert another padding dword, but as it's the end of the stream
userspace may omit that.

So that more a question of policy: do we want userspace to always
specify the size including the padding and reject streams with a size
not double-dword aligned, or do we just fix it up in the kernel, as it's
equally easily done there. I was a bit on the fence here and decided to
go the "let the kernel fix things" route.

Regards,
Lucas
-- 
Pengutronix e.K.             | Lucas Stach                 |
Industrial Linux Solutions   | http://www.pengutronix.de/  |