On 7/29/19 8:33 AM, Andrii Anisov wrote:
Hello Julien,
On 26.07.19 17:02, Julien Grall wrote:
Hi,
It looks like the vmap solution suggested by Andrew & I is a dead end.
I still think we need to do something in the vmap regardless the
alignment decision to avoid unwanted surprised (i.e the
Hello Julien,
On 26.07.19 17:02, Julien Grall wrote:
Hi,
It looks like the vmap solution suggested by Andrew & I is a dead end. I still
think we need to do something in the vmap regardless the alignment decision to
avoid unwanted surprised (i.e the Page-table not in sync with the vmap state).
Hi,
It looks like the vmap solution suggested by Andrew & I is a dead end. I still
think we need to do something in the vmap regardless the alignment decision to
avoid unwanted surprised (i.e the Page-table not in sync with the vmap state).
We potentially want to add some ASSERT_UNREACHABLE()
On 18.07.19 17:20, Julien Grall wrote:
As Andrew pointed out, this also makes the interface more consistent with how
{,un}map_domain_page() currently works.
OK, got it.
BTW, in the x86 code they do apply PAGE_MASK to va before passing it to
`vunmap()`. I would cleanup all those places as w
On 7/18/19 3:08 PM, Andrii Anisov wrote:
Hello Julien,
Hi,
On 18.07.19 16:44, Julien Grall wrote:
My suggestion about vunmap() still stands though.
+1 for the vunmap solution.
It was my initial idea.
But, the issue is introduced by change 9cc0618. In particular, by the
check in `xen_pt
Hello Julien,
On 18.07.19 16:44, Julien Grall wrote:
My suggestion about vunmap() still stands though.
+1 for the vunmap solution.
It was my initial idea.
But, the issue is introduced by change 9cc0618. In particular, by the check in
`xen_pt_update()`
if ( !IS_ALIGNED(virt, PAGE_SIZE)
Hi,
On 7/18/19 2:43 PM, Andrew Cooper wrote:
On 18/07/2019 14:38, Andrew Cooper wrote:
On 18/07/2019 14:22, Andrii Anisov wrote:
From: Andrii Anisov
After changes introduced by 9cc0618 we are able to vmap/vunmap
page aligned addresses only.
So if we add a page address remainder to the mapped
On 18/07/2019 14:38, Andrew Cooper wrote:
> On 18/07/2019 14:22, Andrii Anisov wrote:
>> From: Andrii Anisov
>>
>> After changes introduced by 9cc0618 we are able to vmap/vunmap
>> page aligned addresses only.
>> So if we add a page address remainder to the mapped virtual address,
>> we have to ma
On 18/07/2019 14:22, Andrii Anisov wrote:
> From: Andrii Anisov
>
> After changes introduced by 9cc0618 we are able to vmap/vunmap
> page aligned addresses only.
> So if we add a page address remainder to the mapped virtual address,
> we have to mask it out before unmapping.
>
> Signed-off-by: And
From: Andrii Anisov
After changes introduced by 9cc0618 we are able to vmap/vunmap
page aligned addresses only.
So if we add a page address remainder to the mapped virtual address,
we have to mask it out before unmapping.
Signed-off-by: Andrii Anisov
---
xen/arch/arm/cpuerrata.c | 2 +-
1 file
10 matches
Mail list logo
