On 13/11/15 16:16, Andrew Cooper wrote:
> On 09/11/15 12:00, Ian Campbell wrote:
>> diff --git a/tools/libs/foreignmemory/linux.c
>> b/tools/libs/foreignmemory/linux.c
>> index 01cd42e..86a5a97 100644
>> --- a/tools/libs/foreignmemory/linux.c
>> +++ b/tools/libs/foreignmemory/linux.c
>> @@ -276,6 +276,12 @@ void *xenforeignmemory_map(xenforeignmemory_handle
>> *fmem,
>> return addr;
>> }
>>
>> +int xenforeignmemory_unmap(xenforeignmemory_handle *fmem,
>> + void *addr, unsigned int num)
>> +{
>> + return munmap(addr, (unsigned long)num << PAGE_SHIFT);
> This should be size_t to match the type expected by munmap()
>
>> +}
>> +
>> /*
>> * Local variables:
>> * mode: C
>> diff --git a/tools/libs/foreignmemory/minios.c
>> b/tools/libs/foreignmemory/minios.c
>> index 981d801..dbb152f 100644
>> --- a/tools/libs/foreignmemory/minios.c
>> +++ b/tools/libs/foreignmemory/minios.c
>> @@ -51,6 +51,12 @@ void *xenforeignmemory_map(xenforeignmemory_handle *fmem,
>> return map_frames_ex(arr, num, 1, 0, 1, dom, err, pt_prot);
>> }
>>
>> +int xenforeignmemory_unmap(xenforeignmemory_handle *fmem,
>> + void *addr, unsigned int num)
>> +{
>> + return munmap(addr, num << PAGE_SHIFT);
> This num (and freebsd/netbsd/solaris) needs upcasting before
> multiplication, to avoid loosing the top 12 bits on 64bit platforms.
Sorry - ignore me. I hadn't looked at the subsequent patch while
reviewing this one.
~Andrew
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel
