[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20240709115754.GD13242@willie-the-truck>
Date: Tue, 9 Jul 2024 12:57:54 +0100
From: Will Deacon <will@...nel.org>
To: Steven Price <steven.price@....com>
Cc: kvm@...r.kernel.org, kvmarm@...ts.linux.dev,
Catalin Marinas <catalin.marinas@....com>,
Marc Zyngier <maz@...nel.org>, James Morse <james.morse@....com>,
Oliver Upton <oliver.upton@...ux.dev>,
Suzuki K Poulose <suzuki.poulose@....com>,
Zenghui Yu <yuzenghui@...wei.com>,
linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org,
Joey Gouly <joey.gouly@....com>,
Alexandru Elisei <alexandru.elisei@....com>,
Christoffer Dall <christoffer.dall@....com>,
Fuad Tabba <tabba@...gle.com>, linux-coco@...ts.linux.dev,
Ganapatrao Kulkarni <gankulkarni@...amperecomputing.com>
Subject: Re: [PATCH v4 08/15] arm64: mm: Avoid TLBI when marking pages as
valid
On Mon, Jul 01, 2024 at 10:54:58AM +0100, Steven Price wrote:
> When __change_memory_common() is purely setting the valid bit on a PTE
> (e.g. via the set_memory_valid() call) there is no need for a TLBI as
> either the entry isn't changing (the valid bit was already set) or the
> entry was invalid and so should not have been cached in the TLB.
>
> Signed-off-by: Steven Price <steven.price@....com>
> ---
> v4: New patch
> ---
> arch/arm64/mm/pageattr.c | 8 +++++++-
> 1 file changed, 7 insertions(+), 1 deletion(-)
>
> diff --git a/arch/arm64/mm/pageattr.c b/arch/arm64/mm/pageattr.c
> index 0e270a1c51e6..547a9e0b46c2 100644
> --- a/arch/arm64/mm/pageattr.c
> +++ b/arch/arm64/mm/pageattr.c
> @@ -60,7 +60,13 @@ static int __change_memory_common(unsigned long start, unsigned long size,
> ret = apply_to_page_range(&init_mm, start, size, change_page_range,
> &data);
>
> - flush_tlb_kernel_range(start, start + size);
> + /*
> + * If the memory is being made valid without changing any other bits
> + * then a TLBI isn't required as a non-valid entry cannot be cached in
> + * the TLB.
> + */
> + if (pgprot_val(set_mask) != PTE_VALID || pgprot_val(clear_mask))
> + flush_tlb_kernel_range(start, start + size);
> return ret;
Can you elaborate on when this actually happens, please? It feels like a
case of "Doctor, it hurts when I do this" rather than something we should
be trying to short-circuit in the low-level code.
Will
Powered by blists - more mailing lists