Very Computer

Hello,

Yesterday I compiled the kernel 0.99p11 with the scsi-patches applied,
and booted it from my terminal - so I could not see the results, and
it worked so I did not bother.

Today I looked at my machines first screen and saw the following:

(bootup messages as usual)

(issue as usual)

desaster login: Unable to handle paging request at adress c000004e8
Oops: 0000
EIP:    0010:0015C721
EFLAGS: 00010006
eax: 00000060   ebx: 0000004e4   ecx: 00775000   edx: 000003fd
esi: 0016ff30   edi: 0016ff30    epi: 00440000
ds: 0018  es: 0018  fs: 002b   gs: 002b
Pid: 721, processnr: 20
8b 7b 04 8b 6b 08 8b 4e 20 eb

The machine was up and running, and a second working login was on the
console, though...

Cheers, Michael Will
PS: Where in the zSystem.map do I find a hint about what did this?
    I grepped for 15C721 and 10006, but it did not find any...

Quote:>desaster login: Unable to handle paging request at adress c000004e8
>Oops: 0000
>EIP:    0010:0015C721
>EFLAGS: 00010006
>eax: 00000060   ebx: 0000004e4   ecx: 00775000   edx: 000003fd
>esi: 0016ff30   edi: 0016ff30    epi: 00440000
>ds: 0018  es: 0018  fs: 002b   gs: 002b
>Pid: 721, processnr: 20
>8b 7b 04 8b 6b 08 8b 4e 20 eb
>The machine was up and running, and a second working login was on the
>console, though...

And after a reboot I tried "selection" (I have compiled it into the kernel),
but after entering "selection" I got the following:
Trying to free free memory (00017000): memory probably corrupted
PC = 001140df
VFS: iput: trying to free free inode
VFS: device 0/0, inode 21812, mode=00100711
Segmentation fault

Hey, what?

Cheers, Michael Will
PS: I compiled PLIP into it - and it reported 3 parallel ports with its own
    irqs, but I shure have only one... could this pose problems?

Quote:>>desaster login: Unable to handle paging request at adress c000004e8
>>Oops: 0000
>>EIP:    0010:0015C721
>>EFLAGS: 00010006
>>eax: 00000060   ebx: 0000004e4   ecx: 00775000   edx: 000003fd
>>esi: 0016ff30   edi: 0016ff30    epi: 00440000
>>ds: 0018  es: 0018  fs: 002b   gs: 002b
>>Pid: 721, processnr: 20
>>8b 7b 04 8b 6b 08 8b 4e 20 eb

>>The machine was up and running, and a second working login was on the
>>console, though...

>And after a reboot I tried "selection" (I have compiled it into the kernel),
>but after entering "selection" I got the following:
>Trying to free free memory (00017000): memory probably corrupted
>PC = 001140df
>VFS: iput: trying to free free inode
>VFS: device 0/0, inode 21812, mode=00100711
>Segmentation fault

>Hey, what?

Hmm.  I don't know exactly which SCSI patches you used, but if they are
the ones I'm thinking of by Eric, I'm afraid they seem to have at least
some bugs in the fs/exec.c part (sorry, Eric).  You might want to check
out the ALPHA-pl12 on nic.funet.fi: it contains the scsi patches but not
the fs/exec.c patches.  It also contains the updated FPU-emulation code
and some other things: be sure to read the README carefully (yes, I have
actually updated the README in the tar-file).  The README also contains
instructions on how to parse those pesky kernel error messages.

                Linus