Re: Much better framebuffer fixes.

Re: Much better framebuffer fixes.

[Lars Noschinski]

James Simmons <jsimmons@infradead.org> writes:

> Okay. Here are more framebuffer fixes. Please try these fixes and let me 
> know how they work out for you.

Doesn't work for my Radeon 9100, as I get a kernel panic. As of 2.5.66
the Radeon 9100 is not supported, I run a slightly modified version
(diff to 2.5.66 appended). At least the framebuffer seems to work,
till now I had now time for further tests.

Kernel panic (typos possible):
==============================
EFLAGS   00010282
EIP is at fb_set_var+0x23/0xa0
eax: 6b6b6b6b   ebx: dffe3000   ecx: 0000009f   edx: 00000000
esi: dff8fe71   edi: dffe3001   ebp: dff8f464   esp: dff8fe2c
ds: 007b    es: 007b    ss: 0068
Process swapper (pid: 1, threadinfo=dff8e000 task dff8c000)
Stack: dff8fe64 dff8ff04 dffe30ac dffe3000 co45132e dff8fe64 dffe3000 dffe31dc
       00000100 00000000 dff28000 dffe32fc dffe30bc dffe32fe 00000280 000001e0
       00000280 000001e0 00000000 00000000 00000008 00000000 00000000 00000008
Call Trace
     [<c02adbd8>] radeonfb_pci_register+0x68/0x7e0
     [<c022a591>] pci_device_probe+0x41/0x60
     [<c0271237>] bus_match+0x37/0x70
     [<c0271323>] driver_attach+0x43/0x70
     [<c02715d2>] bus_add_driver+0xa2/0xe0
     [<c02719e1>] driver_register+0x41/0x50
     [<c022a6a2>] pci_register_driver+0x42/0x60
     [<c01051a8>] init+0xc8/0x290
     [<c01050e0>] init+0x0/0x290
     [<c010712d>] kernel_thread_helper+0x5/0x18
Code: 83 78 14 00 75 0f 89 df 8d 75 0c fc b9 28 00 00 00 f3 a5 eb
  <0> Kernel panic: Attempted to kill init!
==============================

Patch:
==============================
--- linux-2.5.66/drivers/video/radeonfb.c	2003-04-02 22:10:08.000000000 +0200
+++ linux-2.5.66-fbdev-ln/drivers/video/radeonfb.c	2003-04-02 22:02:42.000000000 +0200
@@ -121,7 +121,8 @@
 	RADEON_ND,
 	RADEON_NE,
 	RADEON_NF,
-	RADEON_NG
+	RADEON_NG,
+	RADEON_QM	/* LN (my Radeon 9100) */
 };
 
 enum radeon_arch {
@@ -168,7 +169,8 @@
 	{ "9700 ND", RADEON_R300 },
 	{ "9700 NE", RADEON_R300 },
 	{ "9700 NF", RADEON_R300 },
-	{ "9700 NG", RADEON_R300 }
+	{ "9700 NG", RADEON_R300 },
+	{ "9100 QM", RADEON_R200 } /* LN (my Radeon 9100) */
 };
 
 
@@ -213,6 +215,7 @@
 	{ PCI_VENDOR_ID_ATI, PCI_DEVICE_ID_ATI_RADEON_NE, PCI_ANY_ID, PCI_ANY_ID, 0, 0, RADEON_NE},
 	{ PCI_VENDOR_ID_ATI, PCI_DEVICE_ID_ATI_RADEON_NF, PCI_ANY_ID, PCI_ANY_ID, 0, 0, RADEON_NF},
 	{ PCI_VENDOR_ID_ATI, PCI_DEVICE_ID_ATI_RADEON_NG, PCI_ANY_ID, PCI_ANY_ID, 0, 0, RADEON_NG},
+	{ PCI_VENDOR_ID_ATI, PCI_DEVICE_ID_ATI_RADEON_QM, PCI_ANY_ID, PCI_ANY_ID, 0, 0, RADEON_QM}, /* LN (my Radeon 9100) */
 	{ 0, }
 };
 MODULE_DEVICE_TABLE(pci, radeonfb_pci_table);
@@ -881,6 +884,16 @@
 				rinfo->pll.ref_div = 12;
 				rinfo->pll.ref_clk = 2700;
 				break;
+			case PCI_DEVICE_ID_ATI_RADEON_ND:
+			case PCI_DEVICE_ID_ATI_RADEON_NE:
+			case PCI_DEVICE_ID_ATI_RADEON_NF:
+			case PCI_DEVICE_ID_ATI_RADEON_NG:
+				rinfo->pll.ppll_max = 40000;
+				rinfo->pll.ppll_min = 20000;
+				rinfo->pll.xclk = 27000;
+				rinfo->pll.ref_div = 12;
+				rinfo->pll.ref_clk = 2700;
+				break;
 			case PCI_DEVICE_ID_ATI_RADEON_QD:
 			case PCI_DEVICE_ID_ATI_RADEON_QE:
 			case PCI_DEVICE_ID_ATI_RADEON_QF:
--- linux-2.5.66/include/linux/pci_ids.h	2003-04-02 22:10:00.000000000 +0200
+++ linux-2.5.66-fbdev-ln/include/linux/pci_ids.h	2003-04-02 22:02:32.000000000 +0200
@@ -279,6 +279,7 @@
 #define PCI_DEVICE_ID_ATI_RADEON_QO	0x514f
 #define PCI_DEVICE_ID_ATI_RADEON_Ql	0x516c
 #define PCI_DEVICE_ID_ATI_RADEON_BB	0x4242
+#define PCI_DEVICE_ID_ATI_RADEON_QM	0x514d /* LN (my Radeon 9100) */
 /* Radeon RV200 (7500) */
 #define PCI_DEVICE_ID_ATI_RADEON_QW	0x5157
 #define PCI_DEVICE_ID_ATI_RADEON_QX	0x5158

Lars Noschinski

Re: Much better framebuffer fixes.

[Lars Noschinski]

James Simmons <jsimmons@infradead.org> writes:
>> Doesn't work for my Radeon 9100, as I get a kernel panic. As of 2.5.66
>> the Radeon 9100 is not supported, I run a slightly modified version
>> (diff to 2.5.66 appended). At least the framebuffer seems to work,
>> till now I had now time for further tests.
>
> I applied your patch. With your patch plus my it does work perfectly find
> for you?

With both patches, the kernel panics at boot time. With only my patch,
the framebuffer is initialised at 640x480. Console switching and X
work perfectly.

But if I try to change the video mode with fbset 2.1, the screen wents
blank after a console switch. I found no way to change resolution at
boot time ("video=radeon:1024x768-32@85" for example did not work).

Re: Much better framebuffer fixes.

[James Simmons]

> Doesn't work for my Radeon 9100, as I get a kernel panic. As of 2.5.66
> the Radeon 9100 is not supported, I run a slightly modified version
> (diff to 2.5.66 appended). At least the framebuffer seems to work,
> till now I had now time for further tests.

I applied your patch. With your patch plus my it does work perfectly find 
for you?

Much better framebuffer fixes.

[James Simmons]

Okay. Here are more framebuffer fixes. Please try these fixes and let me 
know how they work out for you.

The diff is at 

http://phoenix.infradead.org/~jsimmons/fbdev.diff.gz

 Documentation/devices.txt       |    7 -
 arch/ppc/syslib/prom_init.c     |    2 
 drivers/video/aty/aty128fb.c    |   16 +-
 drivers/video/cfbimgblt.c       |    4 
 drivers/video/console/fbcon.c   |  246 +++++++++++++++++++++-------------------
 drivers/video/controlfb.c       |   18 --
 drivers/video/fbmem.c           |   76 ++++--------
 drivers/video/i810/i810.h       |    6 
 drivers/video/i810/i810_accel.c |  140 +++++++++++-----------
 drivers/video/i810/i810_dvt.c   |    3 
 drivers/video/i810/i810_gtf.c   |    7 -
 drivers/video/i810/i810_main.c  |  135 +++++++++------------
 drivers/video/i810/i810_main.h  |    4 
 drivers/video/logo/logo.c       |   69 +++++------
 drivers/video/platinumfb.c      |   28 +---
 drivers/video/radeonfb.c        |   10 +
 drivers/video/riva/fbdev.c      |    2 
 drivers/video/softcursor.c      |   95 ++++-----------
 drivers/video/tdfxfb.c          |   18 +-
 drivers/video/tgafb.c           |    2 
 drivers/video/vga16fb.c         |    4 
 include/linux/fb.h              |    4 
 include/linux/linux_logo.h      |    2 
 23 files changed, 403 insertions(+), 495 deletions(-)

Re: Much better framebuffer fixes.

[Helge Hafting]

James Simmons wrote:
> Okay. Here are more framebuffer fixes. Please try these fixes and let me 
> know how they work out for you.
> 

The cursor line on radeonfb now is solid, instead of the broken line
in 2.5.66.  It is still impossible to set the resolution.

The lilo approach:
append="video=radeonfb:1280x1024-24@60"
This seems to do nothing.  I get the same low resolution as
plain 2.5.66, which looks bad as it don't match the flat screen resolution.

The stty approach:
stty rows 64 cols 160 gives me an odd error message about
stdin being wrong.  Forcing it with -F /dev/tty
gives me lots of random colored text on the screen and the
scrolling is wrong afterwards.  Resolution don't change though.
I run devfs if that matters.

stty rows 25 cols 80 brought things back to normal, or so it seemed.
I switched back to X and noticed that most programs except ls
wouldn't run, they all segfaulted.  Something was wrong, and
a shutdown ended in a BUG in slab.c line 1557 and a hang.


A new thing with this patch is that I get a lot of junk displayed
to the side of the boot penguin. (Random colored letters).
  The junk disappear after a little while though.

Helge Hafting

Re: Much better framebuffer fixes.

[Thomas Molina]

On Thu, 27 Mar 2003, Helge Hafting wrote:

> James Simmons wrote:
> > Okay. Here are more framebuffer fixes. Please try these fixes and let me 
> > know how they work out for you.
> > 
> 
> The cursor line on radeonfb now is solid, instead of the broken line
> in 2.5.66.  It is still impossible to set the resolution.
> 
> The lilo approach:
> append="video=radeonfb:1280x1024-24@60"
> This seems to do nothing.  I get the same low resolution as
> plain 2.5.66, which looks bad as it don't match the flat screen resolution.

The framebuffer behaviour is back to "normal".  I'm not having the 
resolution problems, but I'm not using the same parameters.  My append 
line simply says vga=792.

Any idea of what will go to Linus and when it might make it into mainline?

Re: Much better framebuffer fixes.

[Mike Houston]

On Thu, 27 Mar 2003 10:58:00 +0100
Helge Hafting <helgehaf@aitel.hist.no> wrote:

> The lilo approach:
> append="video=radeonfb:1280x1024-24@60"
> This seems to do nothing.  I get the same low resolution as
> plain 2.5.66, which looks bad as it don't match the flat screen resolution.
>

I just got a new box a few weeks ago and after building my kernel, struggled for an hour until I did a google search and found some other sod who was doing the same thing as me. It's not "radeonfb:" it's simply "radeon:"

This is the mode I use

append = "video=radeon:1024x768-8@85"

Grogan

Re: Much better framebuffer fixes.

[Michael Dreher]

Donnerstag, 27. März 2003 01:18 wrote James Simmons:
> Okay. Here are more framebuffer fixes. Please try these fixes and let
> me know how they work out for you.

Does not help. I still get a vertically splitted display under X if I
have

CONFIG_FB=y
# CONFIG_FB_CIRRUS is not set
# CONFIG_FB_PM2 is not set
# CONFIG_FB_CYBER2000 is not set
# CONFIG_FB_IMSTT is not set
# CONFIG_FB_VGA16 is not set
# CONFIG_FB_VESA is not set
CONFIG_VIDEO_SELECT=y
# CONFIG_FB_HGA is not set
# CONFIG_FB_RIVA is not set
# CONFIG_FB_MATROX is not set
# CONFIG_FB_RADEON is not set
CONFIG_FB_ATY128=y
# CONFIG_FB_ATY is not set
# CONFIG_FB_SIS is not set
# CONFIG_FB_NEOMAGIC is not set
# CONFIG_FB_3DFX is not set
# CONFIG_FB_VOODOO1 is not set
# CONFIG_FB_TRIDENT is not set
# CONFIG_FB_PM3 is not set
# CONFIG_FB_VIRTUAL is not set