[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <ZsxTDBm57ga6MkPu@arm.com>
Date: Mon, 26 Aug 2024 13:03:56 +0300
From: Catalin Marinas <catalin.marinas@....com>
To: Steven Price <steven.price@....com>
Cc: kvm@...r.kernel.org, kvmarm@...ts.linux.dev,
Marc Zyngier <maz@...nel.org>, Will Deacon <will@...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>,
Gavin Shan <gshan@...hat.com>,
Shanker Donthineni <sdonthineni@...dia.com>,
Alper Gun <alpergun@...gle.com>
Subject: Re: [PATCH v5 05/19] arm64: Detect if in a realm and set RIPAS RAM
On Mon, Aug 19, 2024 at 02:19:10PM +0100, Steven Price wrote:
> +static bool rsi_version_matches(void)
> +{
> + unsigned long ver_lower, ver_higher;
> + unsigned long ret = rsi_request_version(RSI_ABI_VERSION,
> + &ver_lower,
> + &ver_higher);
> +
> + if (ret == SMCCC_RET_NOT_SUPPORTED)
> + return false;
> +
> + if (ret != RSI_SUCCESS) {
> + pr_err("RME: RMM doesn't support RSI version %lu.%lu. Supported range: %lu.%lu-%lu.%lu\n",
> + RSI_ABI_VERSION_MAJOR, RSI_ABI_VERSION_MINOR,
> + RSI_ABI_VERSION_GET_MAJOR(ver_lower),
> + RSI_ABI_VERSION_GET_MINOR(ver_lower),
> + RSI_ABI_VERSION_GET_MAJOR(ver_higher),
> + RSI_ABI_VERSION_GET_MINOR(ver_higher));
> + return false;
> + }
> +
> + pr_info("RME: Using RSI version %lu.%lu\n",
> + RSI_ABI_VERSION_GET_MAJOR(ver_lower),
> + RSI_ABI_VERSION_GET_MINOR(ver_lower));
> +
> + return true;
> +}
I don't have the spec at hand now (on a plane) but given the possibility
of a 1.0 guest regressing on later RMM versions, I wonder whether we
should simply bail out if it's not an exact version match. I forgot what
the spec says about returned ranges (they were pretty confusing last
time I checked).
> +
> +void __init arm64_rsi_setup_memory(void)
> +{
> + u64 i;
> + phys_addr_t start, end;
> +
> + if (!is_realm_world())
> + return;
> +
> + /*
> + * Iterate over the available memory ranges and convert the state to
> + * protected memory. We should take extra care to ensure that we DO NOT
> + * permit any "DESTROYED" pages to be converted to "RAM".
> + *
> + * BUG_ON is used because if the attempt to switch the memory to
> + * protected has failed here, then future accesses to the memory are
> + * simply going to be reflected as a SEA (Synchronous External Abort)
> + * which we can't handle. Bailing out early prevents the guest limping
> + * on and dying later.
> + */
> + for_each_mem_range(i, &start, &end) {
> + BUG_ON(rsi_set_memory_range_protected_safe(start, end));
> + }
Would it help debugging if we print the memory ranges as well rather
than just a BUG_ON()?
--
Catalin
Powered by blists - more mailing lists