[RH9 2.4.20-9] crash "Invalid EIP"

[Lionel Bouton]

Hi,

I posted the following last week and didn't received any reply, in case 
you missed it and have any idea...

One user reported what (at first) seemed to be an SiS5513 IDE related 
bug during kernel loading with the 2.4.20-9 RH9 kernel.

# Summary :

- The hardware is a SiS5513 based IBM Aptiva (Pentium 166 with F00F bug)
- The symptom is an invalid EIP in sis5513.c's config_art_rwp_pio (from
the stack trace summary reported) when trying to configure hdc or hdd.
The kernel boots if passed hdc=none and hdd=none.
- Reverting to the 2.4.18-27.8.0 RH8 kernel solves the problem.
- The problem is 100% reproducible on this hardware.
- different lspci -vxxx tries show unexpected results, see below

# More matter for thought :

What puzzles me is that the sis5513.c driver behaviour at boot time
didn't change for this hardware between these 2 kernels !
I requested lspci -vxxx and dmesg output from fresh cold boots in the
following configurations :

RH8   : 2.4.18-27.8.0 without extra parameters
RH8_2 : 2.4.18-27.8.0 with "hdc=none hdd=none"
RH9   : 2.4.20-9 with "hdc=none hdd=none"

and found out the following :

1/ Some dmesg differences :

RH9 inits the floppy device earlier (before the ide devices instead of
after them),
RH9 adds "Speakup".

2/ There's a difference in the ISA bridge config registers between the
RH8 and the RH8_2 tests :

--- lspci_rh8.txt       2003-04-23 22:19:07.000000000 +0200
+++ lspci_rh8_hdc=none,hdd=none.txt     2003-04-23 22:19:07.000000000 +0200
@@ -17,21 +17,21 @@
00:01.0 ISA bridge: Silicon Integrated Systems [SiS] 85C503/5513 (rev 01)
        Flags: bus master, medium devsel, latency 0
00: 39 10 08 00 07 00 00 02 01 00 01 06 00 00 80 00
10: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
20: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
30: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
40: 0a 0b 0a 80 80 00 00 00 ff ff 10 0f 11 20 04 01
-50: 11 28 02 01 60 00 66 00 9c 2e 12 00 0c e9 00 00
+50: 11 28 02 01 60 00 62 00 9c 2e 12 00 0c e9 00 00
60: 80 80 44 80 82 80 02 00 02 00 04 00 00 00 00 00
70: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
[...]

I don't know what these registers are used for but I don't see any
reason for these changes. I guess there aren't related to the kernel
parameters as previously (without cold boot guarantee) lspci reported
once the following register values too (RH8 or RH8_2 type boot) :

50: 11 28 02 01 60 00 66 00 9c 2e 12 00 36 06 00 00

3/ lspci couldn't find the IDE I/O ports with the RH9 kernel :

00:01.1 IDE interface: Silicon Integrated Systems [SiS] 5513 [IDE] (rev
08) (prog-if 8a [Mas
ter SecP PriP])
        Flags: bus master, fast devsel, latency 0
        I/O ports at <ignored>
        I/O ports at <ignored>
        I/O ports at <ignored>
        I/O ports at <ignored>
        I/O ports at fe80 [size=16]
00: 39 10 13 55 07 00 00 00 08 8a 01 01 00 00 80 00
10: f1 01 00 00 f5 03 00 00 71 01 00 00 75 03 00 00
20: 81 fe 00 00 00 00 00 00 00 00 00 00 00 00 80 00
30: 00 00 00 00 00 00 00 00 00 00 00 00 00 01 00 00
40: 02 03 02 03 02 03 02 03 00 00 46 df 00 02 00 02
50: 02 03 02 03 02 03 02 03 00 00 46 df 00 02 00 02
60: 02 03 02 03 02 03 02 03 00 00 46 df 00 02 00 02
70: 02 03 02 03 02 03 02 03 00 00 46 df 00 02 00 02
80: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
90: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
a0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
b0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
c0: 02 03 02 03 02 03 02 03 00 00 46 df 00 02 00 02
d0: 02 03 02 03 02 03 02 03 00 00 46 df 00 02 00 02
e0: 02 03 02 03 02 03 02 03 00 00 46 df 00 02 00 02
f0: 02 03 02 03 02 03 02 03 00 00 46 df 00 02 00 02

With the RH8 kernel lspci reports the following :
00:01.1 IDE interface: Silicon Integrated Systems [SiS] 5513 [IDE] (rev
08) (prog-if 8a [Mas
ter SecP PriP])
        Flags: bus master, fast devsel, latency 0
        I/O ports at 01f0
        I/O ports at 03f4
        I/O ports at 0170
        I/O ports at 0374
        I/O ports at fe80 [size=16]
[...]

I don't know what to think of this one either.

Hoping this will make some sense for somebody,

LB