seriel console support broken in -mm4 and -mm5

seriel console support broken in -mm4 and -mm5

[Martin J. Bligh]

It works fine in -rc2.

-mm4 prints jibberish (wrong speed / settings?) for serial console, but
the getty stuff comes out file. shutdown just prints more jibberish.

-mm5 prints about half as much jibberish as -mm4 then hangs, seemingly
halfway through boot.

I guess I'll try linus.patch from -mm4 next, unless you have any other
suggestions that'd be more fruitful ...

M.

Re: seriel console support broken in -mm4 and -mm5

[Martin J. Bligh]

> It works fine in -rc2.
> 
> -mm4 prints jibberish (wrong speed / settings?) for serial console, but
> the getty stuff comes out file. shutdown just prints more jibberish.
> 
> -mm5 prints about half as much jibberish as -mm4 then hangs, seemingly
> halfway through boot.
> 
> I guess I'll try linus.patch from -mm4 next, unless you have any other
> suggestions that'd be more fruitful ...

OK, so -mm5 actually does the same as mm4 on my second try, so maybe the
hang is intermittent, or something.

However, linus.patch from -mm4 works fine, so the culprit is one of the
other patches in your tree ... any suggestions for which to shoot first? ;-)

M.

Re: seriel console support broken in -mm4 and -mm5

[Andrew Morton]

"Martin J. Bligh" <mbligh@aracnet.com> wrote:
>
> > It works fine in -rc2.
> > 
> > -mm4 prints jibberish (wrong speed / settings?) for serial console, but
> > the getty stuff comes out file. shutdown just prints more jibberish.
> > 
> > -mm5 prints about half as much jibberish as -mm4 then hangs, seemingly
> > halfway through boot.
> > 
> > I guess I'll try linus.patch from -mm4 next, unless you have any other
> > suggestions that'd be more fruitful ...
> 
> OK, so -mm5 actually does the same as mm4 on my second try, so maybe the
> hang is intermittent, or something.
> 
> However, linus.patch from -mm4 works fine, so the culprit is one of the
> other patches in your tree ... any suggestions for which to shoot first? ;-)
> 

Not really.  Are you using early printk?

The only patches I have which touch drivers/serial/ are
8250-resource-management-fix, linus, pmac_zilog-oops-fix, kgdb-ga.

Lots of patches against drivers/char/*, but I don't see how any of those
can futz the serial settings.  I'm assuming you're using 8250?

drivers/char/cyclades.c~linus
drivers/char/efirtc.c~linus
drivers/char/istallion.c~bk-driver-core
drivers/char/Kconfig~kconfig-url-fixes
drivers/char/keyboard.c~kgdb-ga
drivers/char/mem.c~get_user_pages-handle-VM_IO
drivers/char/mem.c~linus
drivers/char/n_tty.c~move-job-control-stuff-tosignal_struct
drivers/char/random.c~urandom-scalability-fix
drivers/char/rocket.c~move-job-control-stuff-tosignal_struct
drivers/char/rtc.c~linus
drivers/char/sn_serial.c~linus
drivers/char/sonypi.h~linus
drivers/char/stallion.c~bk-driver-core
drivers/char/sx.c~move-job-control-stuff-tosignal_struct
drivers/char/sysrq.c~kgdb-ga
drivers/char/tipar.c~bk-driver-core
drivers/char/tpqic02.c~bk-driver-core
drivers/char/tty_io.c~move-job-control-stuff-tosignal_struct
drivers/char/tty_io.c~remove-down_tty_sem
drivers/char/tty_io.c~tty-locking-again
drivers/char/vc_screen.c~bk-driver-core
drivers/char/viocons.c~linus
drivers/char/viotape.c~linus
drivers/char/vt.c~bk-driver-core
drivers/char/vt.c~con_open-speedup
drivers/char/vt.c~linus
drivers/char/vt.c~move-job-control-stuff-tosignal_struct
drivers/char/vt.c~vt-cleanup
drivers/char/vt_ioctl.c~move-job-control-stuff-tosignal_struct

Re: seriel console support broken in -mm4 and -mm5

[Martin J. Bligh]

>> > It works fine in -rc2.
>> > 
>> > -mm4 prints jibberish (wrong speed / settings?) for serial console, but
>> > the getty stuff comes out file. shutdown just prints more jibberish.
>> > 
>> > -mm5 prints about half as much jibberish as -mm4 then hangs, seemingly
>> > halfway through boot.
>> > 
>> > I guess I'll try linus.patch from -mm4 next, unless you have any other
>> > suggestions that'd be more fruitful ...
>> 
>> OK, so -mm5 actually does the same as mm4 on my second try, so maybe the
>> hang is intermittent, or something.
>> 
>> However, linus.patch from -mm4 works fine, so the culprit is one of the
>> other patches in your tree ... any suggestions for which to shoot first? ;-)
>> 
> 
> Not really.  Are you using early printk?
> 
> The only patches I have which touch drivers/serial/ are
> 8250-resource-management-fix, linus, pmac_zilog-oops-fix, kgdb-ga.
> 
> Lots of patches against drivers/char/*, but I don't see how any of those
> can futz the serial settings.  I'm assuming you're using 8250?

Sometimes my own idleness saves my arse. I finally got around to debugging
it, and it's gone in rc3-mm2 ;-) Yes, was 8250 ...

Thanks,

M