Hi,
On Wed, Apr 02, 2014 at 11:18:27AM -0400, Jerome Glisse wrote:
> This would imply either to scan all mmu_notifier currently register or to
> have a global flags for the mm to know if there is one mmu_notifier without
> change_pte. Moreover this would means that kvm would remain "broken" if
On Wed, Apr 02, 2014 at 03:52:45PM +0300, Haggai Eran wrote:
> On 03/30/2014 11:33 PM, Jerome Glisse wrote:
> >On Wed, Jan 22, 2014 at 04:01:15PM +0200, Haggai Eran wrote:
> >>I'm worried about the following scenario:
> >>
> >>Given a read-only page, suppose one host thread (thread 1) writes to
>
On 03/30/2014 11:33 PM, Jerome Glisse wrote:
On Wed, Jan 22, 2014 at 04:01:15PM +0200, Haggai Eran wrote:
I'm worried about the following scenario:
Given a read-only page, suppose one host thread (thread 1) writes to
that page, and performs COW, but before it calls the
On 03/30/2014 11:33 PM, Jerome Glisse wrote:
On Wed, Jan 22, 2014 at 04:01:15PM +0200, Haggai Eran wrote:
I'm worried about the following scenario:
Given a read-only page, suppose one host thread (thread 1) writes to
that page, and performs COW, but before it calls the
On Wed, Apr 02, 2014 at 03:52:45PM +0300, Haggai Eran wrote:
On 03/30/2014 11:33 PM, Jerome Glisse wrote:
On Wed, Jan 22, 2014 at 04:01:15PM +0200, Haggai Eran wrote:
I'm worried about the following scenario:
Given a read-only page, suppose one host thread (thread 1) writes to
that page,
Hi,
On Wed, Apr 02, 2014 at 11:18:27AM -0400, Jerome Glisse wrote:
This would imply either to scan all mmu_notifier currently register or to
have a global flags for the mm to know if there is one mmu_notifier without
change_pte. Moreover this would means that kvm would remain broken if one
of
On Wed, Jan 22, 2014 at 04:01:15PM +0200, Haggai Eran wrote:
> On 22/01/2014 15:10, Andrea Arcangeli wrote:
> > On Wed, Jan 15, 2014 at 11:40:34AM +0200, Mike Rapoport wrote:
> >> Commit 6bdb913f0a70a4dfb7f066fb15e2d6f960701d00 (mm: wrap calls to
> >> set_pte_at_notify with invalidate_range_start
On Wed, Jan 22, 2014 at 04:01:15PM +0200, Haggai Eran wrote:
On 22/01/2014 15:10, Andrea Arcangeli wrote:
On Wed, Jan 15, 2014 at 11:40:34AM +0200, Mike Rapoport wrote:
Commit 6bdb913f0a70a4dfb7f066fb15e2d6f960701d00 (mm: wrap calls to
set_pte_at_notify with invalidate_range_start and
On Wed, Jan 22, 2014 at 01:54:59PM -0800, Andrew Morton wrote:
> The changelog fails to describe the end-user visible effects of the
> bug, so I (and others) will be unable to decide which kernel versions
> need patching
>
> Given that the bug has been around for 1.5 years I assume the priority
>
On Wed, 15 Jan 2014 11:40:34 +0200 Mike Rapoport
wrote:
> Commit 6bdb913f0a70a4dfb7f066fb15e2d6f960701d00 (mm: wrap calls to
> set_pte_at_notify with invalidate_range_start and invalidate_range_end)
> breaks semantics of set_pte_at_notify. When calls to set_pte_at_notify
> are wrapped with
On 22/01/2014 15:10, Andrea Arcangeli wrote:
> On Wed, Jan 15, 2014 at 11:40:34AM +0200, Mike Rapoport wrote:
>> Commit 6bdb913f0a70a4dfb7f066fb15e2d6f960701d00 (mm: wrap calls to
>> set_pte_at_notify with invalidate_range_start and invalidate_range_end)
>> breaks semantics of set_pte_at_notify.
On Wed, Jan 15, 2014 at 11:40:34AM +0200, Mike Rapoport wrote:
> Commit 6bdb913f0a70a4dfb7f066fb15e2d6f960701d00 (mm: wrap calls to
> set_pte_at_notify with invalidate_range_start and invalidate_range_end)
> breaks semantics of set_pte_at_notify. When calls to set_pte_at_notify
> are wrapped with
On Wed, Jan 15, 2014 at 11:40:34AM +0200, Mike Rapoport wrote:
Commit 6bdb913f0a70a4dfb7f066fb15e2d6f960701d00 (mm: wrap calls to
set_pte_at_notify with invalidate_range_start and invalidate_range_end)
breaks semantics of set_pte_at_notify. When calls to set_pte_at_notify
are wrapped with
On 22/01/2014 15:10, Andrea Arcangeli wrote:
On Wed, Jan 15, 2014 at 11:40:34AM +0200, Mike Rapoport wrote:
Commit 6bdb913f0a70a4dfb7f066fb15e2d6f960701d00 (mm: wrap calls to
set_pte_at_notify with invalidate_range_start and invalidate_range_end)
breaks semantics of set_pte_at_notify. When
On Wed, 15 Jan 2014 11:40:34 +0200 Mike Rapoport
mike.rapop...@ravellosystems.com wrote:
Commit 6bdb913f0a70a4dfb7f066fb15e2d6f960701d00 (mm: wrap calls to
set_pte_at_notify with invalidate_range_start and invalidate_range_end)
breaks semantics of set_pte_at_notify. When calls to
On Wed, Jan 22, 2014 at 01:54:59PM -0800, Andrew Morton wrote:
The changelog fails to describe the end-user visible effects of the
bug, so I (and others) will be unable to decide which kernel versions
need patching
Given that the bug has been around for 1.5 years I assume the priority
is
Commit 6bdb913f0a70a4dfb7f066fb15e2d6f960701d00 (mm: wrap calls to
set_pte_at_notify with invalidate_range_start and invalidate_range_end)
breaks semantics of set_pte_at_notify. When calls to set_pte_at_notify
are wrapped with mmu_notifier_invalidate_range_start and
Commit 6bdb913f0a70a4dfb7f066fb15e2d6f960701d00 (mm: wrap calls to
set_pte_at_notify with invalidate_range_start and invalidate_range_end)
breaks semantics of set_pte_at_notify. When calls to set_pte_at_notify
are wrapped with mmu_notifier_invalidate_range_start and
18 matches
Mail list logo
