[Intel-gfx] [igvt-g-dev] [bug report] drm/i915/gvt: vGPU display virtualization

[Chris Wilson]

On Fri, Nov 11, 2016 at 04:08:58PM +0800, Zhenyu Wang wrote:
> 
> On 2016.11.10 15:54:51 +0300, Dan Carpenter wrote:
> > Hello Zhi Wang,
> > 
> > The patch 04d348ae3f0a: "drm/i915/gvt: vGPU display virtualization"
> > from Apr 25, 2016, leads to the following static checker warning:
> > 
> > 	drivers/gpu/drm/i915/gvt/edid.c:506 intel_gvt_i2c_handle_aux_ch_write()
> > 	warn: odd binop '0x0 & 0xff'
> > 
> > drivers/gpu/drm/i915/gvt/edid.c
> >    501          /* write the return value in AUX_CH_DATA reg which includes:
> >    502           * ACK of I2C_WRITE
> >    503           * returned byte if it is READ
> >    504           */
> >    505  
> >    506          aux_data_for_write |= (GVT_AUX_I2C_REPLY_ACK & 0xff) << 24;
> > 
> > GVT_AUX_I2C_REPLY_ACK is 0 << 6.  Which is weird.  The whole line is a
> > no-op.
> 
> This is from DP spec for AUX i2c ack reply defined as 0 that we emulated
> to reply. I think it's ok to keep as more explicitly to show reply command.

The line is confusing though: you imply that REPLY_ACK is > 256 and that
you only want the low 8bits stuffed into the high 8bits of the dword.
Sounds very weird for the definition, and you wonder whether it is safe.

	aux_data_for_write |= GVT_AUX_I2C_REPLY_ACK << 24;

is still a no-op but should not affend too many sensibilities.
-Chirs

-- 
Chris Wilson, Intel Open Source Technology Centre