Re: [uml-devel] [PATCH] UML: Fix some apparent bitrot

[Boaz Harrosh <bharrosh@panasas.com>]

On 06/25/2009 06:59 AM, Stephen Rothwell wrote:
> Hi all,
> 
> I applied the following patch to yesterday's linux-next and the i386 uml
> defconfig build succeeded.

Stephan Hi, 
Does this means that there will be a UML build every linux-next now?

> 
> The last hunk of this patch (arch/um/include/asm/mmu_context.h) needs to
> be notified to Rusty Russell (cc'd) since that only applied after merging
> his tree.
> 
> The rest appears to apply ok to Linus' tree.
> 
> Please split this and send it to the appropriate people and cc me.
> 

I agree, the minimum of fixing the arch definitions should be separated
from the uml-net-drivers (Which conflicts).

[One patch that needs to go to Linus Yesterday]
arch/um/include/asm/dma-mapping.h
arch/um/include/asm/mmu_context.h

[Second patch through Rusty's tree]
arch/um/drivers/slip_kern.c
arch/um/drivers/slirp_kern.c

Paul? we did not receive any feedback from you. I will wait until Sunday,
then send a patch to Andrew Morton, perhaps he can send such a fix to Linus.
Given UML maintainers have disappeared, right on a merge window.

> On Mon, 22 Jun 2009 11:43:52 +0300 Boaz Harrosh <bharrosh@panasas.com> wrote:
>> On 06/20/2009 03:55 AM, Paul Menage wrote:
>>> UML: Fix some apparent bitrot
>>>
>>> - migration of net_device methods into net_device_ops
>>> - dma_sync_single() changes
>>> - cpumask_clear() -> cpumask_clear_cpu()
>>>
>>> Signed-off-by: Paul Menage <menage@google.com>
>>>
>>> --
>>>
>>> Fixes the following compile errors:
>>>
>>> include/linux/dma-mapping.h:113: error: redefinition of 'dma_sync_single'
>>> arch/um/include/asm/dma-mapping.h:84: error: previous definition of 'dma_sync_single' was here
>>> include/linux/dma-mapping.h: In function 'dma_sync_single':
>>> include/linux/dma-mapping.h:117: error: implicit declaration of function 'dma_sync_single_for_cpu'
>>> include/linux/dma-mapping.h: At top level:
>>> include/linux/dma-mapping.h:120: error: redefinition of 'dma_sync_sg'
>>> arch/um/include/asm/dma-mapping.h:91: error: previous definition of 'dma_sync_sg' was here
>>> include/linux/dma-mapping.h: In function 'dma_sync_sg':
>>> include/linux/dma-mapping.h:124: error: implicit declaration of function 'dma_sync_sg_for_cpu'
>>>
>>>
>>> arch/um/drivers/slirp_kern.c: In function 'slirp_init':
>>> arch/um/drivers/slirp_kern.c:35: error: 'struct net_device' has no member named 'init'
>>>
>>>
>>> arch/um/include/asm/mmu_context.h: In function 'switch_mm':
>>> arch/um/include/asm/mmu_context.h:38: warning: passing argument 1 of 'cpumask_clear' makes pointer from integer without
>>> a cast
>>> arch/um/include/asm/mmu_context.h:38: error: too many arguments to function 'cpumask_clear'
>>>
>>> ---
>> Stephan hi.
>>
>> The above breakage is the usual stuff we get every merge window. With people doing
>> cross arch work and neglecting UML.
>>
>> Who is the person or people responsible for the large linux-next compilation rig?
>> Is it possible to also add a UML traget, at least for the x86_64/x86_32 sub-arch?
>>
>> Thanks
>> Boaz
>>
>>>  4 files changed, 3 insertions(+), 5 deletions(-)
>>>
>>> diff --git a/arch/um/drivers/slip_kern.c b/arch/um/drivers/slip_kern.c
>>> index 5ec1756..dd2aadc 100644
>>> --- a/arch/um/drivers/slip_kern.c
>>> +++ b/arch/um/drivers/slip_kern.c
>>> @@ -30,7 +30,6 @@ static void slip_init(struct net_device *dev, void *data)
>>>  
>>>  	slip_proto_init(&spri->slip);
>>>  
>>> -	dev->init = NULL;
>>>  	dev->hard_header_len = 0;
>>>  	dev->header_ops = NULL;
>>>  	dev->addr_len = 0;
>>> diff --git a/arch/um/drivers/slirp_kern.c b/arch/um/drivers/slirp_kern.c
>>> index f15a6e7..e376284 100644
>>> --- a/arch/um/drivers/slirp_kern.c
>>> +++ b/arch/um/drivers/slirp_kern.c
>>> @@ -32,7 +32,6 @@ void slirp_init(struct net_device *dev, void *data)
>>>  
>>>  	slip_proto_init(&spri->slip);
>>>  
>>> -	dev->init = NULL;
>>>  	dev->hard_header_len = 0;
>>>  	dev->header_ops = NULL;
>>>  	dev->addr_len = 0;
>>> diff --git a/arch/um/include/asm/dma-mapping.h b/arch/um/include/asm/dma-mapping.h
>>> index 90fc708..378de4b 100644
>>> --- a/arch/um/include/asm/dma-mapping.h
>>> +++ b/arch/um/include/asm/dma-mapping.h
>>> @@ -79,14 +79,14 @@ dma_unmap_sg(struct device *dev, struct scatterlist *sg, int nhwentries,
>>>  }
>>>  
>>>  static inline void
>>> -dma_sync_single(struct device *dev, dma_addr_t dma_handle, size_t size,
>>> +dma_sync_single_for_cpu(struct device *dev, dma_addr_t dma_handle, size_t size,
>>>  		enum dma_data_direction direction)
>>>  {
>>>  	BUG();
>>>  }
>>>  
>>>  static inline void
>>> -dma_sync_sg(struct device *dev, struct scatterlist *sg, int nelems,
>>> +dma_sync_sg_for_cpu(struct device *dev, struct scatterlist *sg, int nelems,
>>>  	    enum dma_data_direction direction)
>>>  {
>>>  	BUG();
>>> diff --git a/arch/um/include/asm/mmu_context.h b/arch/um/include/asm/mmu_context.h
>>> index edd719e..34d8130 100644
>>> --- a/arch/um/include/asm/mmu_context.h
>>> +++ b/arch/um/include/asm/mmu_context.h
>>> @@ -35,7 +35,7 @@ static inline void switch_mm(struct mm_struct *prev, struct mm_struct *next,
>>>  	unsigned cpu = smp_processor_id();
>>>  
>>>  	if(prev != next){
>>> -		cpumask_clear(cpu, mm_cpumask(prev));
>>> +		cpumask_clear_cpu(cpu, mm_cpumask(prev));
>>>  		cpumask_set_cpu(cpu, mm_cpumask(next));
>>>  		if(next != &init_mm)
>>>  			__switch_mm(&next->context.id);
>>>
>