All of lore.kernel.org
 help / color / mirror / Atom feed
From: Ulrich Eckhardt <eckhardt@satorlaser.com>
To: linux-mips@linux-mips.org
Subject: Re: [PATCH] au1100fb.c ported from 2.4 to 2.6
Date: Fri, 4 Feb 2005 14:18:47 +0100	[thread overview]
Message-ID: <200502041418.47589.eckhardt@satorlaser.com> (raw)
In-Reply-To: <1107511003.5266.11.camel@absolute.ascensit.com>

Christian wrote:
> First you have to modify the function au1100fb_setcolreg, add the switch
> case for 4bpp. Look in the Alchemy manual for the format of the palette
> in the 4 bpp case. Then look for my comment "TODO: 8bbp", put a switch
> and for pseudocolor like visual ... I think this is not even necessary.

Thanks, will look.

> More important is that you change the monitor type. 

Yep, finding working parameters without real knowledge of what the registers 
mean was the hardest part, but I have it working now.

> If you want I can 
> send you privately a patch this weekend, but I just cannot assure it
> works, you have to test it and eventually contribute.

I'd appreciate that, thank you.



Currently I'm facing a different problem though, look at this code and the 
bitmap it generates (4bpp):

   unsigned long* fb = my_fb_info->fb_virt_start;
   fb[0]   = 0x0000000f; // *.......
   fb[40]  = 0x000000f0; // .*......
   fb[80]  = 0x000000ff; // **......
   fb[120] = 0xf0000f00; // ..*....*
   fb[160] = 0xff000f0f; // *.*...**
   fb[200] = 0xfff00ff0; // .**..***
   fb[240] = 0xffff0fff; // ***.****

To me, expecting the LSB to be part of the leftmost pixel is not really 
surprising, but functions like cfb_fillrect() (or rather bitfill32() etc) 
expect the order to be the other way around. I'm afraid that other 
framebuffer functions also expect that, but I haven't gotten far enough to 
test those. I only tested cfb_fillrect() and, seeing it fail, started looking 
for answers... any idea?



I stumbled across another .. err .. 'feature': take a look at the appended 
patch, in particular the handling of the mask for the first long, am I 
dreaming or is this in fact handled wrongly in the current FB code? I'm a bit 
confused, because the same mistake is made in pretty much every FB driver I 
looked at.

Thanks

Uli

---

Index: cfbcopyarea.c
===================================================================
RCS file: /home/cvs/linux/drivers/video/cfbcopyarea.c,v
retrieving revision 1.10
diff -u -w -r1.10 cfbcopyarea.c
--- cfbcopyarea.c 12 Oct 2004 01:45:47 -0000 1.10
+++ cfbcopyarea.c 4 Feb 2005 13:11:16 -0000
@@ -70,7 +70,7 @@
   } else {
    // Multiple destination words
    // Leading bits
-   if (first) {
+   if (first != ~0UL) {
     
     FB_WRITEL((FB_READL(src) & first) | 
        (FB_READL(dst) & ~first), dst);
@@ -223,7 +223,7 @@
   } else {
    // Multiple destination words
    // Leading bits
-   if (first) {
+   if (first != ~0UL) {
     FB_WRITEL((FB_READL(src) & first) | (FB_READL(dst) & ~first), dst); 
     dst--;
     src--;
Index: cfbfillrect.c
===================================================================
RCS file: /home/cvs/linux/drivers/video/cfbfillrect.c,v
retrieving revision 1.8
diff -u -w -r1.8 cfbfillrect.c
--- cfbfillrect.c 12 Oct 2004 01:45:47 -0000 1.8
+++ cfbfillrect.c 4 Feb 2005 13:11:16 -0000
@@ -142,7 +142,7 @@
  } else {
   // Multiple destination words
   // Leading bits
-  if (first) {
+  if (first != ~0UL) {
    FB_WRITEL(comp(val, FB_READL(dst), first), dst);
    dst++;
    n -= BITS_PER_LONG-dst_idx;
@@ -166,7 +166,7 @@
   
   // Trailing bits
   if (last)
-   FB_WRITEL(comp(val, FB_READL(dst), first), dst);
+   FB_WRITEL(comp(val, FB_READL(dst), last), dst);
  }
 }
 
@@ -197,7 +197,7 @@
  } else {
   // Multiple destination words
   // Leading bits
-  if (first) {
+  if (first != ~0UL) {
    FB_WRITEL(comp(pat, FB_READL(dst), first), dst);
    dst++;
    pat = pat << left | pat >> right;
@@ -224,7 +224,7 @@
   
   // Trailing bits
   if (last)
-   FB_WRITEL(comp(pat, FB_READL(dst), first), dst);
+   FB_WRITEL(comp(pat, FB_READL(dst), last), dst);
  }
 }
 
@@ -252,7 +252,7 @@
  } else {
   // Multiple destination words
   // Leading bits
-  if (first) {
+  if (first != ~0UL) {
    dat = FB_READL(dst);
    FB_WRITEL(comp(dat ^ val, dat, first), dst);
    dst++;
@@ -287,7 +287,7 @@
   // Trailing bits
   if (last) {
    dat = FB_READL(dst);
-   FB_WRITEL(comp(dat ^ val, dat, first), dst);
+   FB_WRITEL(comp(dat ^ val, dat, last), dst);
   }
  }
 }
@@ -320,7 +320,7 @@
  } else {
   // Multiple destination words
   // Leading bits
-  if (first) {
+  if (first != ~0UL) {
    dat = FB_READL(dst);
    FB_WRITEL(comp(dat ^ pat, dat, first), dst);
    dst++;
@@ -354,7 +354,7 @@
   // Trailing bits
   if (last) {
    dat = FB_READL(dst);
-   FB_WRITEL(comp(dat ^ pat, dat, first), dst);
+   FB_WRITEL(comp(dat ^ pat, dat, last), dst);
   }
  }
 }

  reply	other threads:[~2005-02-04 13:16 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2005-01-12  9:50 [PATCH] au1100fb.c ported from 2.4 to 2.6 Christian
2005-02-02  9:05 ` Ulrich Eckhardt
2005-02-04  9:56   ` Christian
2005-02-04 13:18     ` Ulrich Eckhardt [this message]
2005-02-08 14:14 ` Ulrich Eckhardt

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=200502041418.47589.eckhardt@satorlaser.com \
    --to=eckhardt@satorlaser.com \
    --cc=linux-mips@linux-mips.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.