Re: [PATCH v2 4/4] mm: fix KSM data corruption

Hi Andrea, On Tue, Aug 01, 2017 at 09:33:41PM +0200, Andrea Arcangeli wrote: > Hello, > > On Tue, Aug 01, 2017 at 02:56:17PM +0900, Minchan Kim wrote: > > CPU0CPU1CPU2CPU3 > > > > Write the same > >

Re: [PATCH v2 4/4] mm: fix KSM data corruption

Hi Andrea, On Tue, Aug 01, 2017 at 09:33:41PM +0200, Andrea Arcangeli wrote: > Hello, > > On Tue, Aug 01, 2017 at 02:56:17PM +0900, Minchan Kim wrote: > > CPU0CPU1CPU2CPU3 > > > > Write the same > >

Re: [PATCH v2 4/4] mm: fix KSM data corruption

On Tue, Aug 01, 2017 at 12:21:41PM -0700, Nadav Amit wrote: > Minchan Kim wrote: > > > Nadav reported KSM can corrupt the user data by the TLB batching race[1]. > > That means data user written can be lost. > > > > Quote from Nadav Amit > > " > > For this race we need 4

Re: [PATCH v2 4/4] mm: fix KSM data corruption

On Tue, Aug 01, 2017 at 12:21:41PM -0700, Nadav Amit wrote: > Minchan Kim wrote: > > > Nadav reported KSM can corrupt the user data by the TLB batching race[1]. > > That means data user written can be lost. > > > > Quote from Nadav Amit > > " > > For this race we need 4 CPUs: > > > > CPU0:

Re: [PATCH v2 4/4] mm: fix KSM data corruption

Hello, On Tue, Aug 01, 2017 at 02:56:17PM +0900, Minchan Kim wrote: > CPU0 CPU1CPU2CPU3 > > Write the same > value on page > > [cache PTE as > dirty in TLB] > > MADV_FREE >

Re: [PATCH v2 4/4] mm: fix KSM data corruption

Hello, On Tue, Aug 01, 2017 at 02:56:17PM +0900, Minchan Kim wrote: > CPU0 CPU1CPU2CPU3 > > Write the same > value on page > > [cache PTE as > dirty in TLB] > > MADV_FREE >

Re: [PATCH v2 4/4] mm: fix KSM data corruption

Minchan Kim wrote: > Nadav reported KSM can corrupt the user data by the TLB batching race[1]. > That means data user written can be lost. > > Quote from Nadav Amit > " > For this race we need 4 CPUs: > > CPU0: Caches a writable and dirty PTE entry, and uses the stale value

Re: [PATCH v2 4/4] mm: fix KSM data corruption

Minchan Kim wrote: > Nadav reported KSM can corrupt the user data by the TLB batching race[1]. > That means data user written can be lost. > > Quote from Nadav Amit > " > For this race we need 4 CPUs: > > CPU0: Caches a writable and dirty PTE entry, and uses the stale value for > write later.

[PATCH v2 4/4] mm: fix KSM data corruption

Nadav reported KSM can corrupt the user data by the TLB batching race[1]. That means data user written can be lost. Quote from Nadav Amit " For this race we need 4 CPUs: CPU0: Caches a writable and dirty PTE entry, and uses the stale value for write later. CPU1: Runs madvise_free on the range

[PATCH v2 4/4] mm: fix KSM data corruption

Nadav reported KSM can corrupt the user data by the TLB batching race[1]. That means data user written can be lost. Quote from Nadav Amit " For this race we need 4 CPUs: CPU0: Caches a writable and dirty PTE entry, and uses the stale value for write later. CPU1: Runs madvise_free on the range