linux-kernel - [RFC kvm-unit-tests 10/27] arm: realm: Set RIPAS state for RAM

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Fri, 27 Jan 2023 11:40:51 +0000
From:   Joey Gouly <joey.gouly@....com>
To:     Andrew Jones <andrew.jones@...ux.dev>, <kvmarm@...ts.linux.dev>,
        <kvm@...r.kernel.org>
CC:     <joey.gouly@....com>, Alexandru Elisei <alexandru.elisei@....com>,
        Christoffer Dall <christoffer.dall@....com>,
        Fuad Tabba <tabba@...gle.com>,
        Jean-Philippe Brucker <jean-philippe@...aro.org>,
        Joey Gouly <Joey.Gouly@....com>, Marc Zyngier <maz@...nel.org>,
        Mark Rutland <mark.rutland@....com>,
        Oliver Upton <oliver.upton@...ux.dev>,
        Paolo Bonzini <pbonzini@...hat.com>,
        Quentin Perret <qperret@...gle.com>,
        Steven Price <steven.price@....com>,
        Suzuki K Poulose <suzuki.poulose@....com>,
        "Thomas Huth" <thuth@...hat.com>, Will Deacon <will@...nel.org>,
        Zenghui Yu <yuzenghui@...wei.com>,
        <linux-coco@...ts.linux.dev>, <kvmarm@...ts.cs.columbia.edu>,
        <linux-arm-kernel@...ts.infradead.org>,
        <linux-kernel@...r.kernel.org>
Subject: [RFC kvm-unit-tests 10/27] arm: realm: Set RIPAS state for RAM

From: Suzuki K Poulose <suzuki.poulose@....com>

A Realm must ensure that the "RAM" region is set to RIPAS_RAM, before any
access is made. This patch makes sure that all memory blocks are marked as
RIPAS_RAM. Also, before we relocate the "FDT" and "initrd", make sure the
target location is marked too. This happens before we parse the memory blocks.

It is OK to do this operation on a given IPA multiple times. So, we don't
exclude the inital image areas from the "target" list.

Also, this operation doesn't require the host to commit physical memory to back
the IPAs yet. It can be done on demand via fault handling.

Signed-off-by: Suzuki K Poulose <suzuki.poulose@....com>
Signed-off-by: Joey Gouly <joey.gouly@....com>
---
 lib/arm/setup.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/lib/arm/setup.c b/lib/arm/setup.c
index 36d4d826..7b3dc0b8 100644
--- a/lib/arm/setup.c
+++ b/lib/arm/setup.c
@@ -181,6 +181,7 @@ static void mem_init(phys_addr_t freemem_start)
 	while (r && r->end != mem.end)
 		r = mem_region_find(r->end);
 	assert(r);
+	arm_set_memory_protected(r->start, r->end - r->start);
 
 	/* Ensure our selected freemem range is somewhere in our full range */
 	assert(freemem_start >= mem.start && freemem->end <= mem.end);
@@ -252,6 +253,11 @@ void setup(const void *fdt, phys_addr_t freemem_start)
 
 	/* Move the FDT to the base of free memory */
 	fdt_size = fdt_totalsize(fdt);
+	/*
+	 * Before we touch the memory @freemem, make sure it
+	 * is set to protected for Realms.
+	 */
+	arm_set_memory_protected((unsigned long)freemem, fdt_size);
 	ret = fdt_move(fdt, freemem, fdt_size);
 	assert(ret == 0);
 	ret = dt_init(freemem);
@@ -263,6 +269,7 @@ void setup(const void *fdt, phys_addr_t freemem_start)
 	assert(ret == 0 || ret == -FDT_ERR_NOTFOUND);
 	if (ret == 0) {
 		initrd = freemem;
+		arm_set_memory_protected((unsigned long)initrd, initrd_size);
 		memmove(initrd, tmp, initrd_size);
 		freemem += initrd_size;
 	}
-- 
2.17.1