linux-apus-devel

2.4.0-test8 timing problems?

[<jo...@en...>]

2.4.0-test8 still doesn't boot out of box on my A1200/603e. After some
investigation I found out it gets stuck in this loop in
arch/ppc/kernel/time.c:

        do {
                old_stamp = stamp;
                old_sec = sec;
                stamp = get_native_tbl();
                if (__USE_RTC() && stamp < old_stamp) old_stamp -= 1000000000;
                elapsed += stamp - old_stamp;
                sec = ppc_md.get_rtc_time();
        } while ( sec == old_sec && elapsed < 2*HZ*tb_ticks_per_jiffy);

My printk tests indicate that it loops in there but never gets out of the
loop, I don't know why. It might also freeze somewhere inside the
loop after some time. This is dmesg output for a normal boot:

-----------------------------------------------------------------------
Total memory = 64MB; using 0kB for hash table (at 00000000)
Linux version 2.4.0-test8 (root@localhost) (gcc version 2.95.1 19990809 (prerelease)) #31 Sun Oct 8 03:47:23 EEST 2000
Amiga hardware found: [A1200] VIDEO BLITTER AUDIO FLOPPY A1200_IDE KEYBOARD MOUSE SERIAL PARALLEL A2000_CLK CHIP_RAM PAULA LISA ALICE_PAL PCMCIA ZORRO
On node 0 totalpages: 16384
zone(0): 16384 pages.
zone(1): 0 pages.
zone(2): 0 pages.
Kernel command line: root=/dev/hda5 debug=mem
video=pm2fb:640x480-75,font:SUN8x16
amiga_enable_irq: Trying to enable auto-vector IRQ 1
amiga_enable_irq: Trying to enable auto-vector IRQ 3
amiga_enable_irq: Trying to enable auto-vector IRQ 4
amiga_enable_irq: Trying to enable auto-vector IRQ 5
amiga_enable_irq: Trying to enable auto-vector IRQ 7
amiga_enable_irq: Trying to enable auto-vector IRQ 2
amiga_enable_irq: Trying to enable auto-vector IRQ 6
APUS: BATs=1, BUS=67MHz, RAM=70ns, PCI bridge=1
time_init: decrementer frequency = 16.666667 MHz
<<<<<<<<<<<<<<<<<<<<
-----------------------------------------------------------------------

It booted ok after modifying the code so that it will never loop more than
1000 times, after which dmesg says:

-----------------------------------------------------------------------
Warning: real time clock seems stuck!
Console: colour dummy device 80x25
Calibrating delay loop... 111.00 BogoMIPS
Memory: 62600k available (860k kernel code, 392k data, 64k init, 0k highmem)
Dentry-cache hash table entries: 8192 (order: 4, 65536 bytes)
Buffer-cache hash table entries: 4096 (order: 2, 16384 bytes)
Page-cache hash table entries: 16384 (order: 4, 65536 bytes)
Inode-cache hash table entries: 4096 (order: 3, 32768 bytes)
POSIX conformance testing by UNIFIX
Zorro: Probing AutoConfig expansion devices: 1 device
Linux NET4.0 for Linux 2.4
Based upon Swansea University Computer Society NET3.039
NET4: Unix domain sockets 1.0/SMP for Linux NET4.0.
NET4: Linux TCP/IP 1.0 for NET4.0
IP Protocols: ICMP, UDP, TCP
IP: routing cache hash table of 512 buckets, 4Kbytes
TCP: Hash tables configured (established 4096 bind 4096)
Starting kswapd v1.7
Console: switching to colour frame buffer device 80x30
fb0: CVisionPPC/BVisionPPC (Permedia2), using 8192K of video memory.
pty: 256 Unix98 ptys configured
Uniform Multi-Platform E-IDE driver Revision: 6.31
ide: Assuming 50MHz system bus speed for PIO modes; override with idebus=xx
ide0: Gayle IDE interface (A1200 style)
hda: FUJITSU MPC3043AT, ATA DISK drive
ide0 at 80da0000 on irq 0x0000000c
hda: 8448300 sectors (4326 MB), CHS=8940/15/63
Partition check:
 hda: RDSK hda1 hda2 hda3 hda4 hda5 hda6 hda7 hda8 hda9 hda10
Amiga mouse installed.
VFS: Mounted root (ext2 filesystem) readonly.
Freeing unused kernel memory: 64k init
Adding Swap: 103000k swap-space (priority -1)
-----------------------------------------------------------------------
The system seemed to work ok after that, except for serial transfer. It
still gets a _lot_ of tty overruns, losing something like 50% of the
characters. This might be some sort of timing problem related to the
first one? An earlier test kernel booted on this machine (I posted to this
list months ago about it) and it didn't have the time.c problem but serial
wasn't any better that time.

  Jouko

Re: 2.4.0-test8 timing problems?

[Robert R. <ro...@pl...>]

On Sun, Oct 08, 2000 at 06:30:21PM +0300, Jouko Pynnönen wrote:
> 
> 2.4.0-test8 still doesn't boot out of box on my A1200/603e. After some
> investigation I found out it gets stuck in this loop in
> arch/ppc/kernel/time.c:
> 
>         do {
>                 old_stamp = stamp;
>                 old_sec = sec;
>                 stamp = get_native_tbl();
>                 if (__USE_RTC() && stamp < old_stamp) old_stamp -= 1000000000;
>                 elapsed += stamp - old_stamp;
>                 sec = ppc_md.get_rtc_time();
>         } while ( sec == old_sec && elapsed < 2*HZ*tb_ticks_per_jiffy);
 After a quick glance at sources of 2.4.0-test5 i guess this was the cause of
all the problems with my Amiga booting with anything higher that test5 (and i
blamed Michel and pm2fb ;o))) no wonder i couldn't find anything there =o))

 Anyway i'm updating sources to test8 right now.

 Frank Petzold reported on IRC that lp is broken in test7 (as well as scsi
emulation). In self made test5 i couldn't even load modules (modprobe keeps
saying that device is busy; admitedly it;s realy high possible that i screwed
something with modules as i was adding them later)
 Did anyone actually tried to print anything under test7 or higher?

 Jouko big thanks for finding this =o)))

-- 
 Robert Ramiega  | ro...@pl...  IRC: _Jedi_ | Don't underestimate 
 UIN: 13201047   | http://www.plukwa.net/         | the power of Source

Re: 2.4.0-test8 timing problems?

[Roman Z. <zi...@fh...>]

Hi,

On Mon, 9 Oct 2000, Robert Ramiega wrote:

>  After a quick glance at sources of 2.4.0-test5 i guess this was the cause of
> all the problems with my Amiga booting with anything higher that test5 (and i
> blamed Michel and pm2fb ;o))) no wonder i couldn't find anything there =o))

We solved that problem now (I forgot to include the mailing list in the
reply). The problem was that the hw clock was called too often in a short
period (this part was btw introduced in test7).

>  Frank Petzold reported on IRC that lp is broken in test7 (as well as scsi
> emulation). In self made test5 i couldn't even load modules (modprobe keeps
> saying that device is busy; admitedly it;s realy high possible that i screwed
> something with modules as i was adding them later)

Hmm, according to the linux-kernel list we're not the only people that
have a problem with scsi emulation... :-)

bye, Roman

Re: 2.4.0-test8 timing problems?

[Robert R. <ro...@pl...>]

On Mon, Oct 09, 2000 at 07:32:19PM +0200, Roman Zippel wrote:
> 
> We solved that problem now (I forgot to include the mailing list in the
> reply). The problem was that the hw clock was called too often in a short
> period (this part was btw introduced in test7).
 test6 is missing (at least tag for it; is there a way to list all tags in
CVS ?)
 Anyway that patch to time.c does what it was supposed to (ie i'm runing
test8 right now). So far i had mutt quit on me once with "Bus error" message.

 Michel is there some way of forcing xf4 to use permedia directly instead of
FBdev (it's tad slow) or we have to wait till PCI stuff will get there ?
> 
> Hmm, according to the linux-kernel list we're not the only people that
> have a problem with scsi emulation... :-)
 So at least we're not stranded here =o))
 As for printing text seems to print ok but there is additional page that has
only letter E in upper left corner (the same goes when printing postscript
files) Frank, is it the same symptom as You have on test7 ?

-- 
 Robert Ramiega  | ro...@pl...  IRC: _Jedi_ | Don't underestimate 
 UIN: 13201047   | http://www.plukwa.net/         | the power of Source

Re: 2.4.0-test8 timing problems?

[Michel <da...@re...>]

Robert Ramiega wrote:
> 
> On Mon, Oct 09, 2000 at 07:32:19PM +0200, Roman Zippel wrote:
> >
> > We solved that problem now (I forgot to include the mailing list in the
> > reply). The problem was that the hw clock was called too often in a short
> > period (this part was btw introduced in test7).
>  test6 is missing (at least tag for it; is there a way to list all tags in
> CVS ?)

cvs stat -v file

It's well possible that test6 is missing, I guess Roman just imported from
bitkeeper when he had time and didn't do it with every version. After all,
importing always creates conflicts to deal with.


> Michel is there some way of forcing xf4 to use permedia directly instead of
> FBdev (it's tad slow) or we have to wait till PCI stuff will get there ?

Either that or till someone fixes the glint driver. Unfortunately, I won't
probably be able to do anything before three weeks from now (vacation in
between :). So other takers are highly welcome.


Michel


-- 
Earthling Michel Dänzer (MrCooper)  \  CS student and free software enthusiast
Debian GNU/Linux (powerpc,i386) user \   member of XFree86 and the DRI project

Re: 2.4.0-test8 timing problems?

[Sven L. <lu...@dp...>]

On Tue, Oct 10, 2000 at 02:39:14PM +0200, Michel Dänzer wrote:
> Robert Ramiega wrote:
> > Michel is there some way of forcing xf4 to use permedia directly instead of
> > FBdev (it's tad slow) or we have to wait till PCI stuff will get there ?
> 
> Either that or till someone fixes the glint driver. Unfortunately, I won't
> probably be able to do anything before three weeks from now (vacation in
> between :). So other takers are highly welcome.

Michel, what exactly is involved here, ...

I am doing glint driver work right now anyway, and could get a look at it.

(but my bvision is not in my apus box right now, i had power supply problems
and it didn't want to boot).

Friendly,

Sven LUTHER

Re: 2.4.0-test8 timing problems?

[Michel <da...@re...>]

Sven LUTHER wrote:
> 
> On Tue, Oct 10, 2000 at 02:39:14PM +0200, Michel Dänzer wrote:
> > Robert Ramiega wrote:
> > > Michel is there some way of forcing xf4 to use permedia directly instead
> > > of FBdev (it's tad slow) or we have to wait till PCI stuff will get
> > > there ?
> >
> > Either that or till someone fixes the glint driver. Unfortunately, I won't
> > probably be able to do anything before three weeks from now (vacation in
> > between :). So other takers are highly welcome.
> 
> Michel, what exactly is involved here, ...
> 
> I am doing glint driver work right now anyway, and could get a look at it.
> 
> (but my bvision is not in my apus box right now, i had power supply problems
> and it didn't want to boot).

That's bad, but anyway. The problem is that the glint driver segfaults with
Option "UseFBDev", or at least without PCI support, don't know exactly. There
are probably logs in the user list archive.

I suspect the problem may be in the fbdevhw layer (thanks to someone else
'fixing' something there ;).


Michel


-- 
Earthling Michel Dänzer (MrCooper)  \  CS student and free software enthusiast
Debian GNU/Linux (powerpc,i386) user \   member of XFree86 and the DRI project

Re: 2.4.0-test8 timing problems?

[Sven L. <lu...@dp...>]

On Tue, Oct 10, 2000 at 03:39:29PM +0200, Michel Dänzer wrote:
> Sven LUTHER wrote:
> > 
> > On Tue, Oct 10, 2000 at 02:39:14PM +0200, Michel Dänzer wrote:
> > > Robert Ramiega wrote:
> > > > Michel is there some way of forcing xf4 to use permedia directly instead
> > > > of FBdev (it's tad slow) or we have to wait till PCI stuff will get
> > > > there ?
> > >
> > > Either that or till someone fixes the glint driver. Unfortunately, I won't
> > > probably be able to do anything before three weeks from now (vacation in
> > > between :). So other takers are highly welcome.
> > 
> > Michel, what exactly is involved here, ...
> > 
> > I am doing glint driver work right now anyway, and could get a look at it.
> > 
> > (but my bvision is not in my apus box right now, i had power supply problems
> > and it didn't want to boot).
> 
> That's bad, but anyway. The problem is that the glint driver segfaults with
> Option "UseFBDev", or at least without PCI support, don't know exactly. There
> are probably logs in the user list archive.
> 
> I suspect the problem may be in the fbdevhw layer (thanks to someone else
> 'fixing' something there ;).

You don't think it is endianess related or maybe ppc specific, isn't it ?

Maybe i could try it out on my work box, using XF4 on top of my G400
framebuffer.

Anyway, this would be difficult if done without a working box in front to test
it. 

Will think about it.

Friendly,

Sven LUTHER

Re: 2.4.0-test8 timing problems?

[Michel <da...@re...>]

Sven LUTHER wrote:
> 
> On Tue, Oct 10, 2000 at 03:39:29PM +0200, Michel Dänzer wrote:
> > Sven LUTHER wrote:
> > >
> > > On Tue, Oct 10, 2000 at 02:39:14PM +0200, Michel Dänzer wrote:
> > > > Robert Ramiega wrote:
> > > > > Michel is there some way of forcing xf4 to use permedia directly
> > > > > instead of FBdev (it's tad slow) or we have to wait till PCI stuff
> > > > > will get there ?
> > > >
> > > > Either that or till someone fixes the glint driver. Unfortunately, I
> > > > won't probably be able to do anything before three weeks from now
> > > > (vacation in between :). So other takers are highly welcome.
> > >
> > > Michel, what exactly is involved here, ...
> > >
> > > I am doing glint driver work right now anyway, and could get a look at
> > > it.
> > >
> > > (but my bvision is not in my apus box right now, i had power supply
> > > problems and it didn't want to boot).
> >
> > That's bad, but anyway. The problem is that the glint driver segfaults
> > with Option "UseFBDev", or at least without PCI support, don't know
> > exactly. There are probably logs in the user list archive.
> >
> > I suspect the problem may be in the fbdevhw layer (thanks to someone else
> > 'fixing' something there ;).
> 
> You don't think it is endianess related or maybe ppc specific, isn't it ?

Not likely.

> Maybe i could try it out on my work box, using XF4 on top of my G400
> framebuffer.

I don't think the bug would show up there. The glint driver is still the only
one (besides fbdev) which uses fbdevhw in Probe().

> Anyway, this would be difficult if done without a working box in front to
> test it.

I think it's not impossible that the bug would even show up if you tried to
use the glint driver with only amifb.


Michel


-- 
Earthling Michel Dänzer (MrCooper)  \  CS student and free software enthusiast
Debian GNU/Linux (powerpc,i386) user \   member of XFree86 and the DRI project

Re: 2.4.0-test8 timing problems?

[Sven L. <lu...@dp...>]

On Tue, Oct 10, 2000 at 03:58:41PM +0200, Michel Dänzer wrote:
> Sven LUTHER wrote:
> > 
> > On Tue, Oct 10, 2000 at 03:39:29PM +0200, Michel Dänzer wrote:
> > > Sven LUTHER wrote:
> > > >
> > > > On Tue, Oct 10, 2000 at 02:39:14PM +0200, Michel Dänzer wrote:
> > > > > Robert Ramiega wrote:
> > > > > > Michel is there some way of forcing xf4 to use permedia directly
> > > > > > instead of FBdev (it's tad slow) or we have to wait till PCI stuff
> > > > > > will get there ?
> > > > >
> > > > > Either that or till someone fixes the glint driver. Unfortunately, I
> > > > > won't probably be able to do anything before three weeks from now
> > > > > (vacation in between :). So other takers are highly welcome.
> > > >
> > > > Michel, what exactly is involved here, ...
> > > >
> > > > I am doing glint driver work right now anyway, and could get a look at
> > > > it.
> > > >
> > > > (but my bvision is not in my apus box right now, i had power supply
> > > > problems and it didn't want to boot).
> > >
> > > That's bad, but anyway. The problem is that the glint driver segfaults
> > > with Option "UseFBDev", or at least without PCI support, don't know
> > > exactly. There are probably logs in the user list archive.
> > >
> > > I suspect the problem may be in the fbdevhw layer (thanks to someone else
> > > 'fixing' something there ;).
> > 
> > You don't think it is endianess related or maybe ppc specific, isn't it ?
> 
> Not likely.
> 
> > Maybe i could try it out on my work box, using XF4 on top of my G400
> > framebuffer.
> 
> I don't think the bug would show up there. The glint driver is still the only
> one (besides fbdev) which uses fbdevhw in Probe().

What about a permedia2 board then ? Maybe i should write a fbdev for my dual
pm3 board, i wanted to do it anyway. But the current fbdev support in the
glint driver would have to be changed anyway, as the permedia3 is a 128 bit
chip, unlike the permedia2.

> > Anyway, this would be difficult if done without a working box in front to
> > test it.
> 
> I think it's not impossible that the bug would even show up if you tried to
> use the glint driver with only amifb.

huh ?

Anyway, i will be giving it a look this week end.

Friendly,

Sven LUTHER

Re: 2.4.0-test8 timing problems?

[Michel <da...@re...>]

Sven LUTHER wrote:

> > > Maybe i could try it out on my work box, using XF4 on top of my G400
> > > framebuffer.
> >
> > I don't think the bug would show up there. The glint driver is still the
> > only one (besides fbdev) which uses fbdevhw in Probe().
> 
> What about a permedia2 board then ? Maybe i should write a fbdev for my dual
> pm3 board, i wanted to do it anyway. But the current fbdev support in the
> glint driver would have to be changed anyway, as the permedia3 is a 128 bit
> chip, unlike the permedia2.

I don't see where that would be a problem wrt fbdev?


> > > Anyway, this would be difficult if done without a working box in front
> > > to test it.
> >
> > I think it's not impossible that the bug would even show up if you tried
> > to use the glint driver with only amifb.
> 
> huh ?

The glint driver doesn't know if pm2fb is there before it finds out ;) Maybe
the problem is there. Then it might also segfault when there is only amifb
(like it does when there is only pm2fb).


> Anyway, i will be giving it a look this week end.

Thanks Sven.


Michel


-- 
Earthling Michel Dänzer (MrCooper)  \  CS student and free software enthusiast
Debian GNU/Linux (powerpc,i386) user \   member of XFree86 and the DRI project

Re: 2.4.0-test8 timing problems?

[Sven L. <lu...@dp...>]

On Tue, Oct 10, 2000 at 04:09:10PM +0200, Michel Dänzer wrote:
> Sven LUTHER wrote:
> 
> > > > Maybe i could try it out on my work box, using XF4 on top of my G400
> > > > framebuffer.
> > >
> > > I don't think the bug would show up there. The glint driver is still the
> > > only one (besides fbdev) which uses fbdevhw in Probe().
> > 
> > What about a permedia2 board then ? Maybe i should write a fbdev for my dual
> > pm3 board, i wanted to do it anyway. But the current fbdev support in the
> > glint driver would have to be changed anyway, as the permedia3 is a 128 bit
> > chip, unlike the permedia2.
> 
> I don't see where that would be a problem wrt fbdev?

Well, in the glint driver, you make assumption that the detected chip is a
pm2. but the pm3 is quite different. I don't remember well, but they would be
some things to change. It is not much, but it needs checking.

> > > > Anyway, this would be difficult if done without a working box in front
> > > > to test it.
> > >
> > > I think it's not impossible that the bug would even show up if you tried
> > > to use the glint driver with only amifb.
> > 
> > huh ?
> 
> The glint driver doesn't know if pm2fb is there before it finds out ;) Maybe
> the problem is there. Then it might also segfault when there is only amifb
> (like it does when there is only pm2fb).
> 
> > Anyway, i will be giving it a look this week end.
> 
> Thanks Sven.

No problem, ...

Sven Luther

Re: 2.4.0-test8 timing problems?

[Robert R. <ro...@pl...>]

On Tue, Oct 10, 2000 at 03:39:29PM +0200, Michel Dänzer wrote:
> That's bad, but anyway. The problem is that the glint driver segfaults with
> Option "UseFBDev", or at least without PCI support, don't know exactly. There
> are probably logs in the user list archive.
 Actually it doesn't matter if UseFBDev is there or not. 
 In the attachement You'll find small tgz with my XF86Config and 4 Xfree.log
(with some combination of BusID and UseFBDev) I hope that helps.

-- 
 Robert Ramiega  | ro...@pl...  IRC: _Jedi_ | Don't underestimate 
 UIN: 13201047   | http://www.plukwa.net/         | the power of Source

Re: 2.4.0-test8 timing problems?

[Michel <da...@re...>]

Robert Ramiega wrote:
> 
> On Tue, Oct 10, 2000 at 03:39:29PM +0200, Michel Dänzer wrote:
> > That's bad, but anyway. The problem is that the glint driver segfaults
> > with Option "UseFBDev", or at least without PCI support, don't know
> > exactly. There are probably logs in the user list archive.
>  Actually it doesn't matter if UseFBDev is there or not.

Yes, of course, I didn't think of that. fbdevhw is used in the Probe()
function regardless of this option if there is no PCI support.


>  In the attachement You'll find small tgz with my XF86Config and 4 Xfree.log
> (with some combination of BusID and UseFBDev) I hope that helps.

Interesting. Looks like it always crashes if you specify the bus ID.
Otherwise, it gets further, but then crashes. Sven, could it be due to some of
your PM3 changes?


Michel


-- 
Earthling Michel Dänzer (MrCooper)  \  CS student and free software enthusiast
Debian GNU/Linux (powerpc,i386) user \   member of XFree86 and the DRI project

Re: 2.4.0-test8 timing problems?

[Sven L. <lu...@dp...>]

On Tue, Oct 10, 2000 at 07:29:10PM +0200, Michel Dänzer wrote:
> Robert Ramiega wrote:
> > 
> > On Tue, Oct 10, 2000 at 03:39:29PM +0200, Michel Dänzer wrote:
> > > That's bad, but anyway. The problem is that the glint driver segfaults
> > > with Option "UseFBDev", or at least without PCI support, don't know
> > > exactly. There are probably logs in the user list archive.
> >  Actually it doesn't matter if UseFBDev is there or not.
> 
> Yes, of course, I didn't think of that. fbdevhw is used in the Probe()
> function regardless of this option if there is no PCI support.
> 
> 
> >  In the attachement You'll find small tgz with my XF86Config and 4 Xfree.log
> > (with some combination of BusID and UseFBDev) I hope that helps.
> 
> Interesting. Looks like it always crashes if you specify the bus ID.
> Otherwise, it gets further, but then crashes. Sven, could it be due to some of
> your PM3 changes?

No this is not possible, all pm3 changes i did test for the the chip to be a
pm3, and i did nothing yet for fbdev use. At least i think so, will be
checking this evening.

Friendly,

Sven LUTHER

Re: 2.4.0-test8 timing problems?

[Michel <dae...@st...>]

Sven LUTHER wrote:
> 
> On Tue, Oct 10, 2000 at 07:29:10PM +0200, Michel Dänzer wrote:
> > Robert Ramiega wrote:
> > >
> > >  In the attachement You'll find small tgz with my XF86Config and 4
> > > Xfree.log (with some combination of BusID and UseFBDev) I hope that
> > > helps.
> >
> > Interesting. Looks like it always crashes if you specify the bus ID.
> > Otherwise, it gets further, but then crashes. Sven, could it be due to
> > some of your PM3 changes?
> 
> No this is not possible, all pm3 changes i did test for the the chip to be a
> pm3, and i did nothing yet for fbdev use.

The problem might be exactly there. If there is no PCI, you can't check the
PCI info. I had originally tried to guard all PCI related stuff by the
FBDevProbed boolean (maybe we should change that to HavePCI or something?),
maybe you accidentally broke the logic with your PM3 changes?

Sorry, I don't have time to check myself right now.


Michel


-- 
Earthling Michel Dänzer (MrCooper)  \  CS student and free software enthusiast
Debian GNU/Linux (powerpc,i386) user \   member of XFree86 and The DRI Project

Re: 2.4.0-test8 timing problems?

[Sven L. <lu...@dp...>]

On Wed, Oct 11, 2000 at 01:55:45PM +0200, Michel Dänzer wrote:
> Sven LUTHER wrote:
> > 
> > On Tue, Oct 10, 2000 at 07:29:10PM +0200, Michel Dänzer wrote:
> > > Robert Ramiega wrote:
> > > >
> > > >  In the attachement You'll find small tgz with my XF86Config and 4
> > > > Xfree.log (with some combination of BusID and UseFBDev) I hope that
> > > > helps.
> > >
> > > Interesting. Looks like it always crashes if you specify the bus ID.
> > > Otherwise, it gets further, but then crashes. Sven, could it be due to
> > > some of your PM3 changes?
> > 
> > No this is not possible, all pm3 changes i did test for the the chip to be a
> > pm3, and i did nothing yet for fbdev use.
> 
> The problem might be exactly there. If there is no PCI, you can't check the
> PCI info. I had originally tried to guard all PCI related stuff by the
> FBDevProbed boolean (maybe we should change that to HavePCI or something?),
> maybe you accidentally broke the logic with your PM3 changes?

No i think not that i did anything like that, the only stuff i did is
duplicate the permedia2v case in switches and rename it permedia3 and do some
modification then.

BTW, maybe the glint driver would need some restructuring, i thinkk. Whatever
happened to the discution about fbdev probing we had some month ago with
(don't remember exactly, but he was one of the big xfree guys.)

Will try to do a pm3 fbdev this weekend.

Friendly,

Sven LUTHER
> 
> Sorry, I don't have time to check myself right now.
> 
> 
> Michel
> 
> 
> -- 
> Earthling Michel Dänzer (MrCooper)  \  CS student and free software enthusiast
> Debian GNU/Linux (powerpc,i386) user \   member of XFree86 and The DRI Project

Re: 2.4.0-test8 timing problems?

[Michel <dae...@st...>]

Sven LUTHER wrote:

> > > > >  In the attachement You'll find small tgz with my XF86Config and 4
> > > > > Xfree.log (with some combination of BusID and UseFBDev) I hope that
> > > > > helps.
> > > >
> > > > Interesting. Looks like it always crashes if you specify the bus ID.
> > > > Otherwise, it gets further, but then crashes. Sven, could it be due to
> > > > some of your PM3 changes?
> > >
> > > No this is not possible, all pm3 changes i did test for the the chip to
> > > be a pm3, and i did nothing yet for fbdev use.
> >
> > The problem might be exactly there. If there is no PCI, you can't check
> > the PCI info. I had originally tried to guard all PCI related stuff by the
> > FBDevProbed boolean (maybe we should change that to HavePCI or
> > something?), maybe you accidentally broke the logic with your PM3 changes?
> 
> No i think not that i did anything like that, the only stuff i did is
> duplicate the permedia2v case in switches and rename it permedia3 and do
> some modification then.
> 
> BTW, maybe the glint driver would need some restructuring, i thinkk.
> Whatever happened to the discution about fbdev probing we had some month ago
> with (don't remember exactly, but he was one of the big xfree guys.)

I guess you mean Egbert Eich?

Of course, if we'd get PCI working in 2.4 as well it would be best to skip the
fbdev probing code altogether probably.


> Will try to do a pm3 fbdev this weekend.

Good luck, remember that this probably won't help you with this bug as it's
only about PCI not being available.


Michel


-- 
Earthling Michel Dänzer (MrCooper)  \  CS student and free software enthusiast
Debian GNU/Linux (powerpc,i386) user \   member of XFree86 and The DRI Project

Re: 2.4.0-test8 timing problems?

[Sven L. <lu...@dp...>]

On Wed, Oct 11, 2000 at 06:20:58PM +0200, Michel Dänzer wrote:
> Sven LUTHER wrote:
> 
> > > > > >  In the attachement You'll find small tgz with my XF86Config and 4
> > > > > > Xfree.log (with some combination of BusID and UseFBDev) I hope that
> > > > > > helps.
> > > > >
> > > > > Interesting. Looks like it always crashes if you specify the bus ID.
> > > > > Otherwise, it gets further, but then crashes. Sven, could it be due to
> > > > > some of your PM3 changes?
> > > >
> > > > No this is not possible, all pm3 changes i did test for the the chip to
> > > > be a pm3, and i did nothing yet for fbdev use.
> > >
> > > The problem might be exactly there. If there is no PCI, you can't check
> > > the PCI info. I had originally tried to guard all PCI related stuff by the
> > > FBDevProbed boolean (maybe we should change that to HavePCI or
> > > something?), maybe you accidentally broke the logic with your PM3 changes?
> > 
> > No i think not that i did anything like that, the only stuff i did is
> > duplicate the permedia2v case in switches and rename it permedia3 and do
> > some modification then.
> > 
> > BTW, maybe the glint driver would need some restructuring, i thinkk.
> > Whatever happened to the discution about fbdev probing we had some month ago
> > with (don't remember exactly, but he was one of the big xfree guys.)
> 
> I guess you mean Egbert Eich?
> 
> Of course, if we'd get PCI working in 2.4 as well it would be best to skip the
> fbdev probing code altogether probably.

Are there no other hardware without pci using the glint kernel ?

> > Will try to do a pm3 fbdev this weekend.
> 
> Good luck, remember that this probably won't help you with this bug as it's
> only about PCI not being available.

If i compile a kernel without pci support, would it show the bug ?

The truth is, i don't have enough space on my apus harddisk to do Xfree
compiles. Maybe i could try erasing all the disk, and doing a minimal install.

BTW, i will post a true bresenham accel fundtion for glint chips to
xfree-glint today, maybe you could try it out on the pm2. Or do you know of a
reason why the current accels only do 45° inclined lines ?

Friendly,

Sven LUTHER

Re: 2.4.0-test8 timing problems?

[Michel <dae...@st...>]

Sven LUTHER wrote:

> > Of course, if we'd get PCI working in 2.4 as well it would be best to skip
> > the fbdev probing code altogether probably.
> 
> Are there no other hardware without pci using the glint kernel ?

I don't know of any - it's a PCI chip, after all. If m68k people want to use
it, they can easily implement PCI support as well.


> > > Will try to do a pm3 fbdev this weekend.
> >
> > Good luck, remember that this probably won't help you with this bug as
> > it's only about PCI not being available.
> 
> If i compile a kernel without pci support, would it show the bug ?

Hopefully, yes.


> BTW, i will post a true bresenham accel fundtion for glint chips to
> xfree-glint today, maybe you could try it out on the pm2. Or do you know of
> a reason why the current accels only do 45° inclined lines ?

The only thing I can think of is that the glint chips' line algorithm isn't
compatible to Bresenham.


Michel


-- 
Earthling Michel Dänzer (MrCooper)  \  CS student and free software enthusiast
Debian GNU/Linux (powerpc,i386) user \   member of XFree86 and The DRI Project

Re: 2.4.0-test8 timing problems?

[<jo...@en...>]

On Tue, 10 Oct 2000, Robert Ramiega wrote:

>  Anyway that patch to time.c does what it was supposed to (ie i'm runing
> test8 right now). So far i had mutt quit on me once with "Bus error" message.

That patch fixes it on my machine too, but serial transfer still doesn't
work. Roughly half of the characters get lost, even in 9600 bps. I've
tried fiddling with amiserial.c, adding printk's, mb() calls, even 
udelay() calls, but haven't found what is wrong... anyone ideas? It must
be 603e specific since someone told it works fine on a 604.

  Jouko

Re: 2.4.0-test8 timing problems?

[Michel <da...@re...>]

Jouko Pynnönen wrote:

> > Anyway that patch to time.c does what it was supposed to (ie i'm runing
> > test8 right now). So far i had mutt quit on me once with "Bus error"
> > message.
> 
> That patch fixes it on my machine too, but serial transfer still doesn't
> work. Roughly half of the characters get lost, even in 9600 bps. I've
> tried fiddling with amiserial.c, adding printk's, mb() calls, even
> udelay() calls, but haven't found what is wrong... anyone ideas? It must
> be 603e specific since someone told it works fine on a 604.

The last time I tried 2.4 on my BVision, the serial worked at least as good as
with 2.2 . About 5% of the PPP packets got errors.


Michel


-- 
Earthling Michel Dänzer (MrCooper)  \  CS student and free software enthusiast
Debian GNU/Linux (powerpc,i386) user \   member of XFree86 and the DRI project

Re: 2.4.0-test8 timing problems?

[<jo...@en...>]

On Tue, 10 Oct 2000, Michel Dänzer wrote:

> The last time I tried 2.4 on my BVision, the serial worked at least as good as
> with 2.2 . About 5% of the PPP packets got errors.

Here it can't get anywhere near establishing PPP. It usually fails after
the first modem initialization commands because it usually doesn't get
"OK" but " K" or "O ". With a serial cable pppd timeouts with the
negotiation. I also tried turning off the IDE drive with hdparm but it
didn't have any noticeable effect except that the IDE drive got turned
off...

  Jouko

[OT] Linux on Permedia2?

[Michel <da...@re...>]

Michel Dänzer wrote:

> The last time I tried 2.4 on my BVision, the serial worked at least as good
                                  ^^^^^^^
No, I haven't ported Linux to the Permedia2 yet ;) I meant to say 'Blizzard'.

That's the outcome if one doesn't re-read one's own posts...


Michel


-- 
Earthling Michel Dänzer (MrCooper)  \  CS student and free software enthusiast
Debian GNU/Linux (powerpc,i386) user \   member of XFree86 and the DRI project

Re: 2.4.0-test8 timing problems?

[Robert R. <ro...@pl...>]

  Hi!

On Mon, Oct 09, 2000 at 07:32:19PM +0200, Roman Zippel wrote:
> Hi,
> 
> We solved that problem now (I forgot to include the mailing list in the
> reply). The problem was that the hw clock was called too often in a short
> period (this part was btw introduced in test7).
> 
> >  Frank Petzold reported on IRC that lp is broken in test7 (as well as scsi
> > emulation). In self made test5 i couldn't even load modules (modprobe keeps
> > saying that device is busy; admitedly it;s realy high possible that i screwed
> > something with modules as i was adding them later)
> 
> Hmm, according to the linux-kernel list we're not the only people that
> have a problem with scsi emulation... :-)
 what about printing ? under test8 it's screwed badly. Text gets printed
roughly ok but postscript or anything like that gets messy at random points
(the best what i could achieve is half page ok and than some garbage <<like
semigraphic chars>> than some part of printout then garbage <<it wasn't even
complete>>)

 Can this be related to timing probs in some other (than time.c) part of
kernel?

 I also spotted that under some higher loads (like 7 or so) X tend to crash
with signal 7. What this implies ?

-- 
 Robert Ramiega  | ro...@pl...  IRC: _Jedi_ | Don't underestimate 
 UIN: 13201047   | http://www.plukwa.net/         | the power of Source