From: Geert U. <ge...@li...> - 2000-08-03 13:06:17
|
On Thu, 3 Aug 2000, Glenn Hisdal wrote: > > On Wed, 2 Aug 2000, Glenn Hisdal wrote: > >> On 02-Aug-2000, you wrote: > >>>> I still can't get the cybervision64 driver to work in 2.4 though (It > >>>> compiles fine, but I can't use it). > >>>> The kernel just uses amifb:vga even if i tell it to use cyberfb... > >> > >>> Looking at drivers/video/fbmem.c, you should try video=cyber:... > >> > >> Yes, that's what I use... (video=cyber:800x600-8) > >> won't work. Guess I'll have to stick with 2.2.10 a litte longer. > > > Please #define CYBERFBDEBUG in drivers/video/cyberfb.c and tell us more. > > done. > Here's the dmesg output : [...] > cyberfb_init: ENTER And nothing more. So it fails in the resource management code. Ah, stupid typo. The CV64 has 4 MB RAM, not 64... I also missed the release in case of an error. Does the patch below help? Just wondering why no one ever noticed this ;-) --- linux-2.4.0-test6-pre1/drivers/video/cyberfb.c.orig Fri Jul 28 21:19:18 2000 +++ linux-2.4.0-test6-pre1/drivers/video/cyberfb.c Thu Aug 3 14:24:20 2000 @@ -1066,7 +1066,7 @@ CyberRegs_phys = CyberMem_phys + 0x00c00000; if (!request_mem_region(CyberRegs_phys, 0x10000, "S3 Trio64")) continue; - if (!request_mem_region(CyberMem_phys, 0x4000000, "RAM")) { + if (!request_mem_region(CyberMem_phys, 0x400000, "RAM")) { release_mem_region(CyberRegs_phys, 0x10000); continue; } @@ -1110,7 +1110,8 @@ if (register_framebuffer(&fb_info) < 0) { DPRINTK("EXIT - register_framebuffer failed\n"); - release_mem_region(board_addr, board_size); + release_mem_region(CyberMem_phys, 0x400000); + release_mem_region(CyberRegs_phys, 0x10000); return -EINVAL; } Gr{oetje,eeting}s, Geert -- Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- ge...@li... 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 |
From: Glenn H. <gh...@c2...> - 2000-08-03 14:22:21
|
Hello, On 03-Aug-2000, Geert wrote: > On Thu, 3 Aug 2000, Glenn Hisdal wrote: >>> On Wed, 2 Aug 2000, Glenn Hisdal wrote: >>>> On 02-Aug-2000, you wrote: >> >>>> I still can't get the cybervision64 driver to work in 2.4 though (It >> >>>> compiles fine, but I can't use it). >> >>>> The kernel just uses amifb:vga even if i tell it to use cyberfb... >> Please #define CYBERFBDEBUG in drivers/video/cyberfb.c and tell us more. >> done. >> Here's the dmesg output : > [...] >> cyberfb_init: ENTER > And nothing more. So it fails in the resource management code. > Ah, stupid typo. The CV64 has 4 MB RAM, not 64... I also missed the > release in case of an error. Does the patch below help? Yepp ! It works now. Thanks :-) - glenn |
From: Michel <dae...@st...> - 2000-08-03 15:45:45
|
Geert Uytterhoeven wrote: > Ah, stupid typo. The CV64 has 4 MB RAM, not 64... I also missed the release > in case of an error. Does the patch below help? > > Just wondering why no one ever noticed this ;-) I guess because noone ever tested it :( Thanks for the fix Geert, I've committed it, as well as a fix for a few unresolved symbols for modules. I'm now uploading a test kernel for the masses... Michel -- There's no place like ~ ______________________________________________________________________________ Earthling Michel Dänzer (MrCooper) \ CS student and free software enthusiast Debian GNU/Linux (powerpc,i386) user \ member of XFree86, Team *AMIGA*, AUGS |
From: Roman Z. <zi...@fh...> - 2000-08-03 16:34:37
|
Hi, > Thanks for the fix Geert, I've committed it, as well as a fix for a few > unresolved symbols for modules. Who needs vmalloc exported? It's an inline function in <linux/vmalloc.h>. Anyway, tomorrow I hopefully can update to test5 and I have a first version of affstools. :) After that I can do some more tests with affs. bye, Roman |
From: Michel <dae...@st...> - 2000-08-04 09:30:12
|
Roman Zippel wrote: > > Thanks for the fix Geert, I've committed it, as well as a fix for a few > > unresolved symbols for modules. > > Who needs vmalloc exported? z2ram AFAIK. > It's an inline function in <linux/vmalloc.h>. That file isn't included in z2ram.c, instead it declares it extern void* vmalloc (unsigned long); Would the right fix be to remove that definition and include vmalloc.h ? > Anyway, tomorrow I hopefully can update to test5 and I have a first > version of affstools. :) After that I can do some more tests with affs. BTW I've made a new support request about the commit notice (asked them to upload the perl scripts to the right place) and they have updated it to 'Done.' - let's hope it really works now... Michel -- Linux is like living in a teepee. No Windows, no Gates, Apache in house. ______________________________________________________________________________ Earthling Michel Dänzer (MrCooper) \ CS student and free software enthusiast Debian GNU/Linux (powerpc,i386) user \ member of XFree86, Team *AMIGA*, AUGS |
From: Roman Z. <zi...@fh...> - 2000-08-04 11:59:40
|
Hi, > That file isn't included in z2ram.c, instead it declares it > > extern void* vmalloc (unsigned long); > > Would the right fix be to remove that definition and include vmalloc.h ? Yep, otherwise z2ram can't be linked into the kernel. > BTW I've made a new support request about the commit notice (asked them to > upload the perl scripts to the right place) and they have updated it to > 'Done.' - let's hope it really works now... It seems so and there is now a mail in the queue from the import, that exceeds the maximum mail size. :-) bye, Roman |
From: Michel <dae...@st...> - 2000-08-04 12:58:48
|
Roman Zippel wrote: > > That file isn't included in z2ram.c, instead it declares it > > > > extern void* vmalloc (unsigned long); > > > > Would the right fix be to remove that definition and include vmalloc.h ? > > Yep, otherwise z2ram can't be linked into the kernel. Does my fix look good? > > BTW I've made a new support request about the commit notice (asked them to > > upload the perl scripts to the right place) and they have updated it to > > 'Done.' - let's hope it really works now... > > It seems so and there is now a mail in the queue from the import, that > exceeds the maximum mail size. :-) I've approved it and removed the size limit. Please don't abuse it, people :) BTW Roman: do you take care of removed files & dirs when upgrading? I noticed there are still drivers/sound/dmasound.[ch] files in our tree, even though they are removed in 2.4 in favour of a dedicated dmasound directory. Michel -- Here I am! Now what are your other two wishes? ______________________________________________________________________________ Earthling Michel Dänzer (MrCooper) \ CS student and free software enthusiast Debian GNU/Linux (powerpc,i386) user \ member of XFree86, Team *AMIGA*, AUGS |
From: Roman Z. <zi...@fh...> - 2000-08-04 14:59:24
|
Hi, > Does my fix look good? Yes. :) > BTW Roman: do you take care of removed files & dirs when upgrading? cvs import takes care of that. > I noticed > there are still drivers/sound/dmasound.[ch] files in our tree, even though > they are removed in 2.4 in favour of a dedicated dmasound directory. There are still in the bitkeeper tree, it always takes some time until old dirs/files are removed. bye, Roman |
From: <fh...@at...> - 2000-08-02 01:53:41
|
In <Pin...@en...>, on 08/02/00 at 08:08 AM, No Spam <no...@no...> said: >On Tue, 1 Aug 2000, Gordon Huby wrote: >> signal.c >> signal.c: In function `do_sigaction': >> signal.c:921: fixed or forbidden register 68 (0) was spilled for class CR0_REGS. >> This may be due to a compiler bug or to impossible asm statements or clauses. One problem I've noticed with the built it myself compiler (I don't know if the problem I will describe comes from there though) is that the temporary assembler files sometime seem to not get written fully or are corrupted by the compiler before the assembler attempts to assemble them. At least that's the only explanation I can come up with. I get an error something like invalid assembler instruction or parse error. Restarting the compile *most* of the time seems to eliminate the error, but not all the time. Fred |
From: Michel <da...@re...> - 2000-08-02 08:30:00
|
fh...@at... wrote: > One problem I've noticed with the built it myself compiler (I don't know > if the problem I will describe comes from there though) is that the > temporary assembler files sometime seem to not get written fully or are > corrupted by the compiler before the assembler attempts to assemble them. > At least that's the only explanation I can come up with. I get an error > something like invalid assembler instruction or parse error. Restarting > the compile *most* of the time seems to eliminate the error, but not all > the time. Have you tried the -pipe option, which eliminates the temporary files? Michel -- It is easier to fix Unix than to live with NT. ______________________________________________________________________________ Earthling Michel Dänzer (MrCooper) \ CS student and free software enthusiast Debian GNU/Linux (powerpc,i386) user \ member of XFree86, Team *AMIGA*, AUGS |
From: Michel <dae...@st...> - 2000-08-01 21:43:39
|
Gordon Huby wrote: > signal.c: In function `do_sigaction': > signal.c:921: fixed or forbidden register 68 (0) was spilled for class > CR0_REGS. > This may be due to a compiler bug or to impossible asm > statements or clauses. Ah, Ken had the same problem, also using egcs 1.1.2 . I could build it using gcc 2.95.2 . And the kernel even booted! :) Michel -- Secret hacker rule #11: hackers read manuals. ______________________________________________________________________________ Earthling Michel Dänzer (MrCooper) \ CS student and free software enthusiast Debian GNU/Linux (powerpc,i386) user \ member of XFree86, Team *AMIGA*, AUGS |
From: Ken T. <ke...@we...> - 2000-08-01 22:15:23
|
On Tue, 1 Aug 2000, Michel [iso-8859-1] Dänzer wrote: > Ah, Ken had the same problem, also using egcs 1.1.2 . I could build it using > gcc 2.95.2 . And the kernel even booted! :) Is your gcc a 'build it at home in an evening' or from a distribution ? Ken. |
From: Michel <da...@re...> - 2000-08-02 08:35:26
|
Ken Tyler wrote: > > Ah, Ken had the same problem, also using egcs 1.1.2 . I could build it > > using gcc 2.95.2 . And the kernel even booted! :) > > Is your gcc a 'build it at home in an evening' or from a distribution ? The latest from Debian. The same for binutils. Michel -- So many Christians, so few lions. ______________________________________________________________________________ Earthling Michel Dänzer (MrCooper) \ CS student and free software enthusiast Debian GNU/Linux (powerpc,i386) user \ member of XFree86, Team *AMIGA*, AUGS |