On Thu, Oct 16, 2014 at 7:49 AM, Bjorn Helgaas <bhelgaas@xxxxxxxxxx> wrote:
> [+cc Stephen]
>
> On Wed, Oct 15, 2014 at 5:28 PM, Bjorn Helgaas <bhelgaas@xxxxxxxxxx> wrote:
>> [+cc linux-mips]
>>
>> On Wed, Oct 15, 2014 at 11:06 AM, Bjorn Helgaas <bhelgaas@xxxxxxxxxx> wrote:
>>> When the "weak" attribute is on a declaration in a header file, every
>>> definition where the header is included becomes weak, and the linker
>>> chooses one definition based on link order (see 10629d711ed7 ("PCI: Remove
>>> __weak annotation from pcibios_get_phb_of_node decl")).
>>>
>>> Move the "weak" attribute from the declaration to the default definition so
>>> we always prefer a non-weak definition over the weak one, independent of
>>> link order.
>>>
>>> Signed-off-by: Bjorn Helgaas <bhelgaas@xxxxxxxxxx>
>>> CC: linux-mips@xxxxxxxxxxxxxx
Dropping this permanently since I haven't heard from any MIPS folks.
>>> ---
>>> arch/mips/include/asm/vpe.h | 2 +-
>>> arch/mips/kernel/vpe-mt.c | 2 +-
>>> 2 files changed, 2 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/arch/mips/include/asm/vpe.h b/arch/mips/include/asm/vpe.h
>>> index 7849f3978fea..80e70dbd1f64 100644
>>> --- a/arch/mips/include/asm/vpe.h
>>> +++ b/arch/mips/include/asm/vpe.h
>>> @@ -122,7 +122,7 @@ void release_vpe(struct vpe *v);
>>> void *alloc_progmem(unsigned long len);
>>> void release_progmem(void *ptr);
>>>
>>> -int __weak vpe_run(struct vpe *v);
>>> +int vpe_run(struct vpe *v);
>>> void cleanup_tc(struct tc *tc);
>>>
>>> int __init vpe_module_init(void);
>>> diff --git a/arch/mips/kernel/vpe-mt.c b/arch/mips/kernel/vpe-mt.c
>>> index 2e003b11a098..0e5899a2cd96 100644
>>> --- a/arch/mips/kernel/vpe-mt.c
>>> +++ b/arch/mips/kernel/vpe-mt.c
>>> @@ -23,7 +23,7 @@ static int major;
>>> static int hw_tcs, hw_vpes;
>>>
>>> /* We are prepared so configure and start the VPE... */
>>> -int vpe_run(struct vpe *v)
>>> +int __weak vpe_run(struct vpe *v)
>>> {
>>> unsigned long flags, val, dmt_flag;
>>> struct vpe_notifications *notifier;
>>>
>
> Just FYI, this patch was in linux-next today, but I dropped it
> temporarily because Fengguang's auto-builder found the following issue
> with it:
>
> All error/warnings:
>
> arch/mips/kernel/vpe.c: In function 'vpe_release':
>>> arch/mips/kernel/vpe.c:830:29: error: the address of 'vpe_run' will always evaluate as 'true' [-Werror=address]
> if ((vpe_elfload(v) >= 0) && vpe_run) {
> ^
> cc1: all warnings being treated as errors
