Re: [PATCH v2] drm/ast: astdp: fix loop timeout check

[Thomas Zimmermann]

Hi

Am 12.08.24 um 08:54 schrieb Dan Carpenter:
> On Mon, Aug 12, 2024 at 08:48:16AM +0200, Thomas Zimmermann wrote:
> > Hi
> >
> > Am 12.08.24 um 08:42 schrieb Dan Carpenter:
> > > This code has an issue because it loops until "i" is set to UINT_MAX but
> > > the test for failure assumes that "i" is set to zero.  The result is that
> > > it will only print an error message if we succeed on the very last try.
> > > Reformat the loop to count forwards instead of backwards.
> > >
> > > Fixes: 2281475168d2 ("drm/ast: astdp: Perform link training during atomic_enable")
> > > Signed-off-by: Dan Carpenter <dan.carpenter@xxxxxxxxxx>
> > > ---
> > > v2: In version one, I introduced a bug where it would msleep(100) after failure
> > >       and that is a pointless thing to do.  Also change the loop to a for loop.
> > > ---
> > >    drivers/gpu/drm/ast/ast_dp.c | 12 +++++-------
> > >    1 file changed, 5 insertions(+), 7 deletions(-)
> > >
> > > diff --git a/drivers/gpu/drm/ast/ast_dp.c b/drivers/gpu/drm/ast/ast_dp.c
> > > index 5d07678b502c..9bc21dd6a54d 100644
> > > --- a/drivers/gpu/drm/ast/ast_dp.c
> > > +++ b/drivers/gpu/drm/ast/ast_dp.c
> > > @@ -146,18 +146,16 @@ void ast_dp_power_on_off(struct drm_device *dev, bool on)
> > >    void ast_dp_link_training(struct ast_device *ast)
> > >    {
> > >    	struct drm_device *dev = &ast->base;
> > > -	unsigned int i = 10;
> > > +	int i;
> > > -	while (i--) {
> > > +	for (i = 0; i < 10; i++) {
> > >    		u8 vgacrdc = ast_get_index_reg(ast, AST_IO_VGACRI, 0xdc);
> > >    		if (vgacrdc & AST_IO_VGACRDC_LINK_SUCCESS)
> > > -			break;
> > > -		if (i)
> > > -			msleep(100);
> > > +			return;
> > > +		msleep(100);
> > But we don't want to wait during the final iteration of this loop. If you
> > want to use the for loop, it should be something like
> >
> > for (i= 0; i < 10; ++i) {
> >
> >      if (i)
> >        msleep(100)
> >
> >      // now test vgacrdc
> > }
> >
> > Best regards
> > Thomas
> I feel like if we really hit this failure path then we won't care about the
> tenth msleep().  I can resend if you want, but I'd prefer to just leave it.

Please resend. Even if the link training ultimately fails, the rest of 
DRM keeps running. 100 msec is not so short to shrug it off IMHO.

Best regards
Thomas

> regards,
> dan carpenter

--
--
Thomas Zimmermann
Graphics Driver Developer
SUSE Software Solutions Germany GmbH
Frankenstrasse 146, 90461 Nuernberg, Germany
GF: Ivo Totev, Andrew Myers, Andrew McDonald, Boudien Moerman
HRB 36809 (AG Nuernberg)