Skip to contentSkip to conversation

2.6.0-test5 vs. modem cards

DISCUSSION Norman Diamond 2 participants 3 messages Sep 14, 2003

lore
Share & tools
lore
#0 lore raw

When inserting a modem card, 2.6.0-test5 continues to have problems which earlier 2.6.0-test versions had, which can be solved by booting 2.4.19.

In file 8250_cs.c:
Line 61, identifies itself as "serial_cs.c" instead of "8250_cs.c". Line 119 identifies itself as "serial_cs" instead of "8250_cs". My partial understanding of Linux PCMCIA operations yields a guess that line 119 is part of the cause for failure during execution, whereas line 61 only potentially confuses future maintainers.

Later in the same source file, calls to register_serial() and unregister_serial() compile but fail during execution. Of course in order to make it execute in the first place I have to manually modprobe 8250_cs, because of the reason mentioned above. /var/log/messages gets reports that those symbols are unknown.

#1 lore raw
Quoted 6 lines · Norman Diamond

In file 8250_cs.c:
Line 61, identifies itself as "serial_cs.c" instead of "8250_cs.c". Line 119 identifies itself as "serial_cs" instead of "8250_cs". My partial understanding of Linux PCMCIA operations yields a guess that line 119 is part of the cause for failure during execution, whereas line 61 only potentially confuses future maintainers.

Its going to get renamed back to serial_cs shortly.

Quoted 5 lines

Later in the same source file, calls to register_serial() and unregister_serial() compile but fail during execution. Of course in order to make it execute in the first place I have to manually modprobe 8250_cs, because of the reason mentioned above. /var/log/messages gets reports that those symbols are unknown.

I have no idea how you managed that. The configuration subsystem does not allow you to build 8250_cs.c as a module without building 8250.c in some manner, and 8250.c provides those symbols.

Signature · 5 lines
Russell King ([email protected])	http://www.arm.linux.org.uk/personal/
Linux kernel maintainer of:
  2.6 ARM Linux   - http://www.arm.linux.org.uk/
  2.6 PCMCIA      - http://pcmcia.arm.linux.org.uk/
  2.6 Serial core
#2 lore raw

"Russell King" <[email protected]> replied to me:

Quoted 3 lines
Quoted 1 line

In file 8250_cs.c:

Its going to get renamed back to serial_cs shortly.

OK.

Quoted 9 lines
Quoted 5 lines

Later in the same source file, calls to register_serial() and unregister_serial() compile but fail during execution. Of course in order to make it execute in the first place I have to manually modprobe 8250_cs, because of the reason mentioned above. /var/log/messages gets reports that those symbols are unknown.

I have no idea how you managed that. The configuration subsystem does not allow you to build 8250_cs.c as a module without building 8250.c in some manner, and 8250.c provides those symbols.

I built 8250.c as a module, but might have neglected to modprobe it. Since a manual modprobe of 8250_cs also neglected to automatically modprobe its dependency 8250, there might be breakage in modprobe, or this might just be additional breakage in the naming of 8250 vs. serial.