Re: [PATCH] x86: use the correct function type for native_set_fixmap

2019-10-08 Thread Sami Tolvanen
On Tue, Sep 24, 2019 at 12:51 PM Sami Tolvanen wrote: > Someone else probably knows better, but yes, we could also fix this by > changing set_fixmap to accept enum fixed_addresses as the first > parameter, and changing the type of xen_set_fixmap instead. This approach is actually more

Re: [PATCH] x86: use the correct function type for native_set_fixmap

2019-09-24 Thread Sami Tolvanen
On Mon, Sep 23, 2019 at 4:28 PM Kees Cook wrote: > Is it correct that pv_mmu_ops can't be changed since this is an external > API? Someone else probably knows better, but yes, we could also fix this by changing set_fixmap to accept enum fixed_addresses as the first parameter, and changing the

Re: [PATCH] x86: use the correct function type for native_set_fixmap

2019-09-23 Thread Kees Cook
On Fri, Sep 13, 2019 at 02:14:02PM -0700, Sami Tolvanen wrote: > We call native_set_fixmap indirectly through the function pointer > struct pv_mmu_ops::set_fixmap, which expects the first parameter to be > 'unsigned' instead of 'enum fixed_addresses'. This patch changes the > function type for

[PATCH] x86: use the correct function type for native_set_fixmap

2019-09-13 Thread Sami Tolvanen
We call native_set_fixmap indirectly through the function pointer struct pv_mmu_ops::set_fixmap, which expects the first parameter to be 'unsigned' instead of 'enum fixed_addresses'. This patch changes the function type for native_set_fixmap to match the pointer, which fixes indirect call