aboutsummaryrefslogtreecommitdiff
path: root/drivers/serial
diff options
context:
space:
mode:
authorVivek Goyal <vgoyal@in.ibm.com>2007-01-11 01:52:44 +0100
committerAndi Kleen <andi@basil.nowhere.org>2007-01-11 01:52:44 +0100
commite96763d6e52c22a8821466e4d01c0c995b8087be (patch)
tree655f07a67c0d75b738d9721832dbe284b0a4818c /drivers/serial
parent4a5d107a9a79ef3404209a43866554cba451dda4 (diff)
[PATCH] i386: make apic probe function non-init
o struct genapic contains pointer to probe() function which is of type __init. Hence MODPOST generates warning if kernel is compiled with CONFIG_RELOCATABLE=y for i386. WARNING: vmlinux - Section mismatch: reference to .init.text: from .data between 'apic_summit' (at offset 0xc058b504) and 'apic_bigsmp' WARNING: vmlinux - Section mismatch: reference to .init.text: from .data between 'apic_bigsmp' (at offset 0xc058b5a4) and 'cpu.4471' WARNING: vmlinux - Section mismatch: reference to .init.text: from .data between 'apic_es7000' (at offset 0xc058b644) and 'apic_default' WARNING: vmlinux - Section mismatch: reference to .init.text: from .data between 'apic_default' (at offset 0xc058b6e4) and 'interrupt' o One of the possible options is to put special case check in MODPOST to not emit warnings for this case but I think it is not a very good option in terms of maintenance. o Another option is to make probe() function non __init. Anyway this function is really small so not freeing this memory after init is not a big deal. Secondly, from a programming perspective, probably genapic should not provide pointers to functions which have been freed as genapic is non __init and is used even after initialization is complete. Signed-off-by: Vivek Goyal <vgoyal@in.ibm.com> Signed-off-by: Andi Kleen <ak@suse.de> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Cc: Andi Kleen <ak@suse.de> Signed-off-by: Andrew Morton <akpm@osdl.org>
Diffstat (limited to 'drivers/serial')
0 files changed, 0 insertions, 0 deletions