Problem with the serial port in Fedora Core 6

Gene Heskett gene.heskett at verizon.net
Sat Nov 11 15:45:54 UTC 2006


On Saturday 11 November 2006 10:00, Craig White wrote:
>On Sat, 2006-11-11 at 01:46 -0500, Gene Heskett wrote:
>> >> Now, I snipped the USB stuff back out.  And, note that the 8250 is
>> >> set for 32 of them, with 4 being the default.  My working kernel is
>> >> 2 & 2 there. I think I need a beer, coffee isn't working anymore :)
>> >>
>> >> Can we file a bug?
>> >
>> >----
>> >perhaps you should try to absorb the information from Mikkel's post
>> >first - the one that you seemingly have ignored, not the one you
>> > already responded to.
>>
>> I have read it, several times.  The symptoms he reports are identical
>> to mine including the responses of setserial.
>
>----
>not that post - that is what I was saying...this post
>
>https://www.redhat.com/archives/fedora-list/2006-November/msg02696.html
>----
I haven't looked at that one yet Craig.

>>   And, you are missing my
>> point, which is that I can swap the two drives, putting the drive with
>> FC2 on it back in the hda position, boot 2.6.19-rc5 from it, and it
>> all Just Works(TM).  I'd image the FP, if booted to an older release,
>> would fine his serial ports are just fine too.
>
>----
>not at all missing your point. It's a theory assuming that the kernel is
>the issue.

Ok, but then how does one explain that the bootlog doesn't contain ANY 
references to the ttyS* being found?  The only serial related stuff in my 
dmesg is:
PNP: PS/2 Controller [PNP0303:PS2K] at 0x60,0x64 irq 1
PNP: PS/2 controller doesn't have AUX irq; using default 12
serio: i8042 AUX port at 0x60,0x64 irq 12
serio: i8042 KBD port at 0x60,0x64 irq 1
mice: PS/2 mouse device common for all mice

And thats it, nary a character about tty* discovery.  So the kernel DIDN'T 
find them, and you can beat around the bush all you want, if the kernel 
didn't find them at boot time, they are NOT going to be available for 
use.

>His message is that udev controls this but you can spend all the time
>you want trying to compile a kernel to fix your lack of understanding of
>udev if you choose. His suggestion is that the ports are there but
>perhaps not where you expect them (i.e. ttyS0 and ttyS1 might not be the
>correct handles).
>----
>
>> In other words Craig, the hardware is fine, that particular software
>> is a 10 day old carcass.
>
>----
>in the hands of someone who wants to build a mountain out of a molehill
>to solve a problem that this someone that has a theory that this is a
>kernel problem but no actual comprehension of udev. On FC-2, udev barely
>existed (I don't recall exactly but I do recall that FC-2 was the first
>release of 2.6 kernel for Fedora so I am sure that if udev existed in
>FC-2, it was not implemented anything like it's form in FC-6).

I never used it with FC2.  And udev can generate all the ttyS* devices it 
wants to, and does make 4 of them, but if the kernel didn't find them 
when it booted, they aren't going to be usable regardless of what udev 
does.  Its a kernel bug in the kernel supplied with FC6.

2.6.18-1.2798.fc6xen

Now can we agree that?  Or are you going to revert to the denegrating 
attitude that I have NDI what I'm talking about that you exhibited the 
last time we tangled?
>----
>
>> Now, I've just spent the better part of 3 hours playing 10,000 monkeys
>> in the FC6/boot/grub/grub.conf, trying make either a direct boot from
>> the drive containing the FC2 install, or to chainload it after setting
>> a root (hd1,0) between the title line & the kernel lines, and specing
>> root=/dev/hdb7 as a kernel argument.  That would be a whole lot
>> simpler.
>>
>> But I cannot get any of my entries that point to /dev/hdb to even show
>> up in the boot menu on a reboot.  I have added /dev/hdb to the
>> devices.list also, and I have edited the /etc/fstab of the FC2 disk to
>> reflect that its now /dev/hdb instead of /dev/hda.
>>
>> Running an 'mbchk grub.conf' reports theres "no multiboot header", and
>> while the word 'multiboot' is used in the info file, this "multiboot
>> header" isn't defined that I can find.  Thats what I love about linux
>> docs, always 6 months out of date, like when did they change the
>> 'initrd' line to be a 'module' line?  Or are there subtle differences
>> but the usage happens to appear identical?
>>
>> Anyway, here is the grub.conf as it exists now, whats wrong with it?
>> ----------------
>> # grub.conf generated by anaconda
>> #
>> # Note that you do not have to rerun grub after making changes to this
>> file
>> # NOTICE:  You have a /boot partition.  This means that
>> #          all kernel and initrd paths are relative to /boot/, eg.
>> #          root (hd0,0)
>> #          kernel /vmlinuz-version ro root=/dev/VolGroup00/LogVol00
>> #          initrd /initrd-version.img
>> #boot=/dev/hda
>> default=0
>> timeout=7
>> splashimage=(hd0,0)/grub/splash.xpm.gz
>> # hiddenmenu
>> # 0
>> title Fedora Core (2.6.18-1.2798.fc6xen)
>> 	root (hd0,0)
>> 	kernel /xen.gz-2.6.18-1.2798.fc6
>> 	module /vmlinuz-2.6.18-1.2798.fc6xen ro root=/dev/VolGroup00/LogVol00
>> 	module /initrd-2.6.18-1.2798.fc6xen.img
>>
>> # 1
>> title Fedora Core 6 (2.6.19-rc5)
>> 	root(hd0,0)
>> 	kernel /vmlinuz-2.6.19-rc5 ro root=/dev/VolGroup00/LogVol00 rhgb
>> quiet module /initrd-2.6.19-rc5.img
>>
>> # 2
>> title Fedora Core 2 Menu
>> 	rootnoverify (hd1,0)
>> 	chainloader +1
>> # 3
>> title DOS
>>         rootnoverify (hd0,1)
>>         chainloader +1
>>
>> --------------
>> # 1 will unpack, and boot to a missing console message, locked up, so
>> somethings still missing even after I added the LVM stuffs.  It works
>> fine for FC2, but without the LVM stuff that I built into this version
>> after getting a very early crash because it wasn't there.  And believe
>> it or not, the last 'DOS' entry does display in the boot menu.
>
>----
>I can't tell that anything is wrong with your grub.conf other than I've
>never seen the labels that you have added in grub and wonder about them.
>I also wonder if the grub conf that you think is loading is the one that
>is actually loading since when you add a second disk to the equation,
>BIOS doesn't always see disk (hd0) or (hd1) as you think it will.

The bios see's them just fine, all 4 of them.  The FC6 install is on a 
160GB as hd0, the FC2 install is on a 120GB as hd1.  A new lightscribe 
dual layer burner is hd2, and hd3 is a 200GB that amanda yses 180GB of 
for vtapes, FC2's /var is on hd3, and FC2's swap is on hd3. And if you 
are referring to the # 0 etc labels, I've been using them since grub was 
new, what 5 years ago?  There are in fact, 26 such entries in my FC2 
grub.conf, and they all display correctly.  Mmm, it just occured to me 
that maybe its expecting a pair of () around the word 'menu', or maybe a 
kernel version?  But heck, its (the title line) supposedly just a label 
to display to the user AFAIK.

>Craig

-- 
Cheers, Gene




More information about the users mailing list