linux-kernel - Re: [PATCH v3 4/4] KVM: selftests: arm64: Skip all 32 bit IDs when set_id

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]

Message-ID: <90fb25bd-fa79-4980-902b-497346a3e6d5@sirena.org.uk>
Date: Mon, 5 Jan 2026 16:45:50 +0000
From: Mark Brown <broonie@...nel.org>
To: Ben Horgan <ben.horgan@....com>
Cc: Marc Zyngier <maz@...nel.org>, Joey Gouly <joey.gouly@....com>,
	Suzuki K Poulose <suzuki.poulose@....com>,
	Paolo Bonzini <pbonzini@...hat.com>, Shuah Khan <shuah@...nel.org>,
	Oliver Upton <oupton@...nel.org>,
	linux-arm-kernel@...ts.infradead.org, kvmarm@...ts.linux.dev,
	kvm@...r.kernel.org, linux-kselftest@...r.kernel.org,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH v3 4/4] KVM: selftests: arm64: Skip all 32 bit IDs when
 set_id_regs is aarch64 only

On Fri, Jan 02, 2026 at 02:50:07PM +0000, Ben Horgan wrote:
> On 12/19/25 19:28, Mark Brown wrote:

> > +			if (aarch64_only && sys_reg_CRm(reg_id) < 4) {
> > +				ksft_test_result_skip("%s value seen in guest\n",
> > +						      get_reg_name(reg_id));
> > +				break;
> > +			}
> > +

> Unnecessary? The decision for which regs are testing is made in
> guest_code().

The guest code has a fixed list of registers it reads blindly and we
skip the write for these so our expected value won't be something we
explicitly set.  The actual test is done here in the host code and it
seems both more maintainable to keep the skip adjacent to the live test
and clearer to be more explicit about nothing actually being tested.

> > +	if (aarch64_only && sys_reg_CRm(encoding) < 4) {

> Doesn't this exclude more registers than needed? E.g. MIDR?

Yes, I took this test from somewhere without thinking about it properly
- it's been so long I can't remember but it was clearly wrong in this
context.  I'll update.

Download attachment "signature.asc" of type "application/pgp-signature" (489 bytes)