On Thu, 14 Nov 2019 18:12:31 +0100 Joerg Roedel <jroedel@xxxxxxx> wrote: > On Wed, Nov 13, 2019 at 05:55:30PM +0800, Shile Zhang wrote: > > +#if defined(CONFIG_X86_32) && defined(CONFIG_X86_PAE) > > /* > > * First make sure the mappings are removed from all page-tables > > * before they are freed. > > + * > > + * This is only needed on x86-32 with !SHARED_KERNEL_PMD, which is > > + * the case on a PAE kernel with PTI enabled. > > */ > > - vmalloc_sync_all(); > > + if (!SHARED_KERNEL_PMD && boot_cpu_has(X86_FEATURE_PTI)) > > + vmalloc_sync_all(); > > +#endif > > I already submitted another fix for this quite some time ago: > > https://lore.kernel.org/lkml/20191009124418.8286-1-joro@xxxxxxxxxx/ > Your patch is more verbose but looks quite a bit nicer than this patch (Link: http://lkml.kernel.org/r/20191113095530.228959-1-shile.zhang@xxxxxxxxxxxxxxxxx). The separation of sync-for-a-mapping versus sync-for-an-unmapping adds clarity. It's fairly urgent - I consider this to be -stable material. Thomas & co, was that a deliberate skip?
