Skip to contentSkip to conversation

atafb line length

PATCH/RFC Geert Uytterhoeven 2 participants 6 messages Nov 16, 2008 – Dec 17, 2008 Framebuffer layer

lore
Share & tools
lore
#0 lore raw

Make sure fix->line_length is always set, as some applications need it because they don't have fallback code if line_length is zero.

Works on ARAnyM (Falcon emulation), but par->next_line is not set on any other Atari variant?

---
 drivers/video/atafb.c |    9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)
drivers/video/atafb.c +5 -4
@@ -613,7 +613,7 @@ static int tt_encode_fix(struct fb_fix_s
 	fix->xpanstep = 0;
 	fix->ypanstep = 1;
 	fix->ywrapstep = 0;
-	fix->line_length = 0;
+	fix->line_length = par->next_line;
 	fix->accel = FB_ACCEL_ATARIBLITT;
 	return 0;
 }
@@ -917,7 +917,8 @@ static int falcon_encode_fix(struct fb_f
 		fix->visual = FB_VISUAL_TRUECOLOR;
 		fix->xpanstep = 2;
 	}
-	fix->line_length = 0;
+	fix->line_length =
+		(par->hw.falcon.line_width + par->hw.falcon.line_offset) * 2;
 	fix->accel = FB_ACCEL_ATARIBLITT;
 	return 0;
 }
@@ -1844,7 +1845,7 @@ static int stste_encode_fix(struct fb_fi
 		fix->ypanstep = 0;
 	}
 	fix->ywrapstep = 0;
-	fix->line_length = 0;
+	fix->line_length = par->next_line;
 	fix->accel = FB_ACCEL_ATARIBLITT;
 	return 0;
 }
@@ -2161,7 +2162,7 @@ static int ext_encode_fix(struct fb_fix_
 	fix->xpanstep = 0;
 	fix->ypanstep = 0;
 	fix->ywrapstep = 0;
-	fix->line_length = 0;
+	fix->line_length = par->next_line;
 	return 0;
 }
 

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

#1 lore raw

Hi Geert,

Quoted 2 lines

Make sure fix->line_length is always set, as some applications need it because they don't have fallback code if line_length is zero.

I did set line_length to zero for a reason while hacking on the driver, but my recollection of the reasons is quite fuzzy. I'll give it a test run on the real hardware.

As with all other Atari bugs recently introduced, it may equally well be sloppy coding on my part.

Quoted 2 lines

Works on ARAnyM (Falcon emulation), but par->next_line is not set on any other Atari variant?

fix->line_length may not be used by anything other than the VIDEL. I'll check that in the hardware specific code.

Thanks for adopting atafb :-)

Michael

#2 lore raw
In reply to Geert Uytterhoeven

Hi Geert,

Quoted 5 lines

Make sure fix->line_length is always set, as some applications need it because they don't have fallback code if line_length is zero.

Works on ARAnyM (Falcon emulation), but par->next_line is not set on any other Atari variant?

You're right - while testing on the Falcon framebuffer I noticed fix->line_length needs to be set, and kludged around to get it set to something more or less sane, but I could not test the other code paths.

It looks like par->next_line should be set to linelen in all other decode_var variants. The quilt tree here does have fix->line_length = par->next_line already, what is your patch based on?

I had

fix->line_length = (par->hw.falcon.line_width + par->hw.falcon.line_offset) *

		par->hw.falcon.bpp / 8;

originally - I suppose this had to change with your cfb fixes?

Michael

#3 lore raw
Quoted 6 lines · Michael Schmitz
Quoted 5 lines

Make sure fix->line_length is always set, as some applications need it because they don't have fallback code if line_length is zero.

Works on ARAnyM (Falcon emulation), but par->next_line is not set on any other Atari variant?

Quoted 7 lines

You're right - while testing on the Falcon framebuffer I noticed fix->line_length needs to be set, and kludged around to get it set to something more or less sane, but I could not test the other code paths.

It looks like par->next_line should be set to linelen in all other decode_var variants. The quilt tree here does have fix->line_length = par->next_line already, what is your patch based on?

It was based on mainline. Yes, a part of my patch had been in my quilt series for a while, but I wanted to post the full thing for review.

Quoted 6 lines

I had

fix->line_length = (par->hw.falcon.line_width + par->hw.falcon.line_offset) *

		par->hw.falcon.bpp / 8;

originally - I suppose this had to change with your cfb fixes?

That should be OK. Strange, it's not what I got from you, cfr. atafb in mainline.

Cfb16 needed
1. fix->line_length to be valid (it was zero)
2. info->pseudo_palette

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

#4 lore raw

Hi,

Quoted 6 lines
Quoted 3 lines

It looks like par->next_line should be set to linelen in all other decode_var variants. The quilt tree here does have fix->line_length = par->next_line already, what is your patch based on?

It was based on mainline. Yes, a part of my patch had been in my quilt series for a while, but I wanted to post the full thing for review.

OK, I see.

Quoted 9 lines
Quoted 6 lines

I had

fix->line_length = (par->hw.falcon.line_width + par->hw.falcon.line_offset) *

		par->hw.falcon.bpp / 8;

originally - I suppose this had to change with your cfb fixes?

That should be OK. Strange, it's not what I got from you, cfr. atafb in mainline.

I concede this may be something I changed later on. atafb had been worked on my Roman Zippel as well after my initial patches so I'll double check this isn't obsolete (I was looking at my current working tree though).

Quoted 3 lines

Cfb16 needed
1. fix->line_length to be valid (it was zero)
2. info->pseudo_palette

OK, so the above fix->line_length may be independent from your fixes.

I'll test that (just not right away, home improvement work has taken a heavy toll on my time lately).

Michael

#5 lore raw
In reply to Geert Uytterhoeven
Quoted 5 lines · Geert Uytterhoeven

Make sure fix->line_length is always set, as some applications need it because they don't have fallback code if line_length is zero.

Works on ARAnyM (Falcon emulation), but par->next_line is not set on any other Atari variant?

And this should fix the remaining ones. I cannot test it, though.

Without this, atafb is completely broken on anything but Falcon, as all the text console drawing operations need a valid par->next_line.

On (emulated) Falcon, we can just use par->next_line, too.

I plan to clean up the atafb/c2p patch series, and get it into 2.6.29.

drivers/video/atafb.c +5 -2
index 8f70296..477ce55 100644
@@ -691,6 +691,7 @@ static int tt_decode_var(struct fb_var_screeninfo *var, struct atafb_par *par)
 		return -EINVAL;
 	par->yres_virtual = yres_virtual;
 	par->screen_base = screen_base + var->yoffset * linelen;
+	par->next_line = linelen;
 	return 0;
 }
 
@@ -914,8 +915,7 @@ static int falcon_encode_fix(struct fb_fix_screeninfo *fix,
 		fix->visual = FB_VISUAL_TRUECOLOR;
 		fix->xpanstep = 2;
 	}
-	fix->line_length =
-		(par->hw.falcon.line_width + par->hw.falcon.line_offset) * 2;
+	fix->line_length = par->next_line;
 	fix->accel = FB_ACCEL_ATARIBLITT;
 	return 0;
 }
@@ -1900,6 +1900,7 @@ static int stste_decode_var(struct fb_var_screeninfo *var,
 		return -EINVAL;
 	par->yres_virtual = yres_virtual;
 	par->screen_base = screen_base + var->yoffset * linelen;
+	par->next_line = linelen;
 	return 0;
 }
 
@@ -2174,6 +2175,8 @@ static int ext_decode_var(struct fb_var_screeninfo *var, struct atafb_par *par)
 	    var->xoffset > 0 ||
 	    var->yoffset > 0)
 		return -EINVAL;
+
+	par->next_line = external_xres_virtual * external_depth / 8;
 	return 0;
 }
 

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds