v3.4-stable-queue build errors

v3.4-stable-queue build errors

[Guenter Roeck]

kernel/printk.c: In function 'register_console':
kernel/printk.c:1501:3: error: size of unnamed array is negative
    BUILD_BUG_ON(sizeof(console_cmdline[i].name) != sizeof(newcon->name));

All architectures / builds are affected.

Caused by 'console: Fix console name size mismatch', which introduces
the BUILD_BUG_ON (but nothing else). Looks like the actual name length
increase in struct console_cmdline (in kernel/printk.c) got lost.

Guenter

Re: v3.4-stable-queue build errors

[Zefan Li]

On 2015/6/13 13:34, Guenter Roeck wrote:
> kernel/printk.c: In function 'register_console':
> kernel/printk.c:1501:3: error: size of unnamed array is negative
>    BUILD_BUG_ON(sizeof(console_cmdline[i].name) != sizeof(newcon->name));
> 
> All architectures / builds are affected.
> 
> Caused by 'console: Fix console name size mismatch', which introduces
> the BUILD_BUG_ON (but nothing else). Looks like the actual name length
> increase in struct console_cmdline (in kernel/printk.c) got lost.
> 

Now should be fixed. Don't know how that happened..

Thanks for testing!

Re: v3.4-stable-queue build errors

[Guenter Roeck]

On 06/14/2015 08:23 PM, Zefan Li wrote:
> On 2015/6/13 13:34, Guenter Roeck wrote:
>> kernel/printk.c: In function 'register_console':
>> kernel/printk.c:1501:3: error: size of unnamed array is negative
>>     BUILD_BUG_ON(sizeof(console_cmdline[i].name) != sizeof(newcon->name));
>>
>> All architectures / builds are affected.
>>
>> Caused by 'console: Fix console name size mismatch', which introduces
>> the BUILD_BUG_ON (but nothing else). Looks like the actual name length
>> increase in struct console_cmdline (in kernel/printk.c) got lost.
>>
>
> Now should be fixed. Don't know how that happened..
>

This problem is gone, but there is another problem.

Building alpha:defconfig ... failed
--------------
Error log:
mm/slub.c: In function 'unfreeze_partials':
mm/slub.c:1920:7: error: 'm' undeclared (first use in this function)
mm/slub.c:1920:7: note: each undeclared identifier is reported only once for each function it appears in
mm/slub.c:1920:12: error: 'M_FREE' undeclared (first use in this function)
make[1]: *** [mm/slub.o] Error 1

This affects most of the builds.

Guenter

Re: v3.4-stable-queue build errors

[Zefan Li]

On 2015/6/15 12:54, Guenter Roeck wrote:
> On 06/14/2015 08:23 PM, Zefan Li wrote:
>> On 2015/6/13 13:34, Guenter Roeck wrote:
>>> kernel/printk.c: In function 'register_console':
>>> kernel/printk.c:1501:3: error: size of unnamed array is negative
>>>     BUILD_BUG_ON(sizeof(console_cmdline[i].name) != sizeof(newcon->name));
>>>
>>> All architectures / builds are affected.
>>>
>>> Caused by 'console: Fix console name size mismatch', which introduces
>>> the BUILD_BUG_ON (but nothing else). Looks like the actual name length
>>> increase in struct console_cmdline (in kernel/printk.c) got lost.
>>>
>>
>> Now should be fixed. Don't know how that happened..
>>
> 
> This problem is gone, but there is another problem.
> 
> Building alpha:defconfig ... failed
> --------------
> Error log:
> mm/slub.c: In function 'unfreeze_partials':
> mm/slub.c:1920:7: error: 'm' undeclared (first use in this function)
> mm/slub.c:1920:7: note: each undeclared identifier is reported only once for each function it appears in
> mm/slub.c:1920:12: error: 'M_FREE' undeclared (first use in this function)
> make[1]: *** [mm/slub.o] Error 1
> 
> This affects most of the builds.
> 

I just backported a slub fix but forgot the last chunk.. My config enables slab
instead of slub, so I didn't catch this.

Now should be fixed. Thanks again.

Re: v3.4-stable-queue build errors

[Guenter Roeck]

On 06/14/2015 11:32 PM, Zefan Li wrote:

>
> Now should be fixed. Thanks again.
>
>
Now we are down to

Building powerpc:defconfig ... failed
--------------
Error log:
arch/powerpc/platforms/powernv/pci-ioda.c: In function 'pnv_ioda_setup_pe_segments':
arch/powerpc/platforms/powernv/pci-ioda.c:395:3: error: passing argument 1 of 'pcibios_resource_to_bus' from incompatible pointer type [-Werror]
    pcibios_resource_to_bus(dev, &region, &io_res);

arch/powerpc/platforms/powernv/pci-ioda.c:425:3: error: passing argument 1 of 'pcibios_resource_to_bus' from incompatible pointer type [-Werror]
    pcibios_resource_to_bus(dev, &region, &m32_res);

which affects a couple of powerpc builds.

Guenter

Re: v3.4-stable-queue build errors

[Zefan Li]

> Now we are down to
> 
> Building powerpc:defconfig ... failed
> --------------
> Error log:
> arch/powerpc/platforms/powernv/pci-ioda.c: In function 'pnv_ioda_setup_pe_segments':
> arch/powerpc/platforms/powernv/pci-ioda.c:395:3: error: passing argument 1 of 'pcibios_resource_to_bus' from incompatible pointer type [-Werror]
>    pcibios_resource_to_bus(dev, &region, &io_res);
> 
> arch/powerpc/platforms/powernv/pci-ioda.c:425:3: error: passing argument 1 of 'pcibios_resource_to_bus' from incompatible pointer type [-Werror]
>    pcibios_resource_to_bus(dev, &region, &m32_res);
> 
> which affects a couple of powerpc builds.
> 

I made some changes in my local tree, but didn't use "quilt add" to add them
to the patch.

Now fixed. Hope this is the last one...