From mboxrd@z Thu Jan 1 00:00:00 1970 From: will.deacon@arm.com (Will Deacon) Date: Tue, 6 Nov 2018 22:04:03 +0000 Subject: [PATCH v3 1/2] arm64: mm: purge lazily unmapped vm regions before changing permissions In-Reply-To: <20181106214404.2497-2-ard.biesheuvel@linaro.org> References: <20181106214404.2497-1-ard.biesheuvel@linaro.org> <20181106214404.2497-2-ard.biesheuvel@linaro.org> Message-ID: <20181106220359.GA32210@brain-police> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Tue, Nov 06, 2018 at 10:44:03PM +0100, Ard Biesheuvel wrote: > Call vm_unmap_aliases() every time we apply any changes to permission > attributes of mappings in the vmalloc region. This avoids any potential > issues resulting from lingering writable or executable aliases of > mappings that should be read-only or non-executable, respectively. > > Signed-off-by: Ard Biesheuvel > --- > arch/arm64/mm/pageattr.c | 6 ++++++ > 1 file changed, 6 insertions(+) > > diff --git a/arch/arm64/mm/pageattr.c b/arch/arm64/mm/pageattr.c > index a56359373d8b..f8cf5bc1d1f8 100644 > --- a/arch/arm64/mm/pageattr.c > +++ b/arch/arm64/mm/pageattr.c > @@ -93,6 +93,12 @@ static int change_memory_common(unsigned long addr, int numpages, > if (!numpages) > return 0; > > + /* > + * Get rid of potentially aliasing lazily unmapped vm areas that may > + * have permissions set that deviate from the ones we are setting here. > + */ > + vm_unmap_aliases(); So this might_sleep(), which I don't think is currently the case for our set_memory_XX() functions. However, a quick look at the existing callsites indicates that's ok and matches the x86 implementation, so: Acked-by: Will Deacon Will