Skip to contentSkip to conversation

2.4-pre5[{-}xyz]: 4 machines, feedback only

DISCUSSION Mario Vanoni 3 participants 4 messages Sep 8, 2002 – Sep 9, 2002

lore
Share & tools
lore
#0 lore raw

gr4: UP PIII550 1GB mem SCSI only, 4 HD's
gr5: UP Celeron996 512MB mem IDE. 2 HD's
va1: dual SMP PIII550 1GB mem SCSI only, 4 HD's
va3: dual SMP PIII1266 2GB mem,
SCSI HD's (4), IDE DVD + burner

2.4.20-pre5: all machines OK >24h

2.4.20-pre5aa1: gr4 network 10Mbits/BNC _dead_,

                gr5, va1 and va3 OK >24h

2.4.20-pre5-ac4: gr4 & va1 OK, >24h
gr5 & va3 _not_ compilable, at ide/idedriver.o stop

All kernels _without_ modules, pure static, and no other patches.

Regards

Mario, _not_ in lmkl

PS gr4 is a production machine,
touchable maybe on weekends !!!

#1 lore raw
Quoted 5 lines · Mario Vanoni

...
2.4.20-pre5-ac4: gr4 & va1 OK, >24h
gr5 & va3 _not_ compilable,
at ide/idedriver.o stop
...

Could you send the error message and the .config that caused the compile error?

Quoted 4 lines

Regards

Mario, _not_ in lmkl
...

TIA
Adrian

Signature · 3 lines
You only think this is a free country. Like the US the UK spends a lot of
time explaining its a free country because its a police state.
								Alan Cox
#2 lore raw
In reply to Mario Vanoni
Quoted 5 lines · Mario Vanoni

...
2.4.20-pre5-ac4: gr4 & va1 OK, >24h
gr5 & va3 _not_ compilable,
at ide/idedriver.o stop
...

Mario sent me in a private mail the .configs. He answered at "make *config" that he didn't read the release notes. That's good because it uncovered the following that is a real bug:

<-- snip -->

...

        --end-group \
        -o vmlinux

drivers/ide/idedriver.o: In function `proc_ide_read_drivers':
drivers/ide/idedriver.o(.text+0x3fe): undefined reference to ide_modules' drivers/ide/idedriver.o: In function proc_ide_read_identify':
drivers/ide/idedriver.o(.text+0x635): undefined reference to `taskfile_lib_get_identify'
drivers/ide/idedriver.o: In function `proc_ide_read_settings':
drivers/ide/idedriver.o(.text+0x73d): undefined reference to `ide_read_setting'
drivers/ide/idedriver.o: In function `proc_ide_write_settings':
drivers/ide/idedriver.o(.text+0x977): undefined reference to ide_find_setting_by_name' drivers/ide/idedriver.o(.text+0x9a2): undefined reference to ide_write_setting'
drivers/ide/idedriver.o: In function `proc_ide_write_driver':
drivers/ide/idedriver.o(.text+0xbfb): undefined reference to `ide_replace_subdriver'
drivers/ide/idedriver.o: In function `create_proc_ide_drives':
drivers/ide/idedriver.o(.text+0xdbb): undefined reference to `generic_subdriver_entries'
drivers/ide/idedriver.o: In function `create_proc_ide_interfaces':
drivers/ide/idedriver.o(.text+0xf44): undefined reference to ide_hwifs' drivers/ide/idedriver.o(.text+0xf49): undefined reference to ide_hwifs' drivers/ide/idedriver.o(.text+0xf4e): undefined reference to ide_hwifs' drivers/ide/idedriver.o: In function destroy_proc_ide_interfaces':
drivers/ide/idedriver.o(.text+0xfa8): undefined reference to ide_hwifs' drivers/ide/idedriver.o(.text+0xfad): undefined reference to ide_hwifs' drivers/ide/idedriver.o(.text+0xfb2): more undefined references to `ide_hwifs' follow
make: *** [vmlinux] Error 1

<-- snip -->

ide-proc.c is compiled into idedriver.o even if no IDE support is compiled into the kernel. That isn't new. The problem that causes these undefined references is that in -ac4 some functions that use functions from other ide files (which aren't compiled when building a kernel without IDE support) are no longer static (because they are now exported to modules).

Quoted 4 lines

Regards

Mario, _not_ in lmkl
...

cu
Adrian

Signature · 3 lines
You only think this is a free country. Like the US the UK spends a lot of
time explaining its a free country because its a police state.
								Alan Cox
#3 lore raw
Quoted 2 lines

ide-proc.c is compiled into idedriver.o even if no IDE support is compiled into the kernel. That isn't new. The problem that causes these undefined

Thts fine

Quoted 3 lines

references is that in -ac4 some functions that use functions from other ide files (which aren't compiled when building a kernel without IDE support) are no longer static (because they are now exported to modules).

idedriver.o shouldnt be getting into a kernel without IDE. Thats the real problem