From: Joey Gouly <joey.gouly@arm.com>
To: Eric Auger <eric.auger@redhat.com>
Cc: kvm@vger.kernel.org, alexandru.elisei@arm.com,
andrew.jones@linux.dev, kvmarm@lists.linux.dev,
Marc Zyngier <maz@kernel.org>,
Oliver Upton <oliver.upton@linux.dev>
Subject: Re: [kvm-unit-tests PATCH v3 07/10] arm64: selftest: update test for running at EL2
Date: Tue, 2 Dec 2025 12:21:15 +0000 [thread overview]
Message-ID: <20251202122115.GA3921791@e124191.cambridge.arm.com> (raw)
In-Reply-To: <5160dadb-1ff3-487e-bd0b-9f643c3d9ec3@redhat.com>
On Tue, Dec 02, 2025 at 10:16:42AM +0100, Eric Auger wrote:
>
>
> On 9/25/25 4:19 PM, Joey Gouly wrote:
> > From: Alexandru Elisei <alexandru.elisei@arm.com>
> >
> > Remove some hard-coded assumptions that this test is running at EL1.
> >
> > Signed-off-by: Alexandru Elisei <alexandru.elisei@arm.com>
> > Signed-off-by: Joey Gouly <joey.gouly@arm.com>
> > ---
> > arm/selftest.c | 18 +++++++++++++-----
> > 1 file changed, 13 insertions(+), 5 deletions(-)
> >
> > diff --git a/arm/selftest.c b/arm/selftest.c
> > index 1553ed8e..01691389 100644
> > --- a/arm/selftest.c
> > +++ b/arm/selftest.c
> > @@ -232,6 +232,7 @@ static void user_psci_system_off(struct pt_regs *regs)
> > __user_psci_system_off();
> > }
> > #elif defined(__aarch64__)
> > +static unsigned long expected_level;
> >
> > /*
> > * Capture the current register state and execute an instruction
> > @@ -276,8 +277,7 @@ static bool check_regs(struct pt_regs *regs)
> > {
> > unsigned i;
> >
> > - /* exception handlers should always run in EL1 */
> > - if (current_level() != CurrentEL_EL1)
> > + if (current_level() != expected_level)
> > return false;
> >
> > for (i = 0; i < ARRAY_SIZE(regs->regs); ++i) {
> > @@ -301,7 +301,11 @@ static enum vector check_vector_prep(void)
> > return EL0_SYNC_64;
> >
> > asm volatile("mrs %0, daif" : "=r" (daif) ::);
> > - expected_regs.pstate = daif | PSR_MODE_EL1h;
> > + expected_regs.pstate = daif;
> > + if (current_level() == CurrentEL_EL1)
> > + expected_regs.pstate |= PSR_MODE_EL1h;
> > + else
> > + expected_regs.pstate |= PSR_MODE_EL2h;
> > return EL1H_SYNC;
> > }
> >
> > @@ -317,8 +321,8 @@ static bool check_und(void)
> >
> > install_exception_handler(v, ESR_EL1_EC_UNKNOWN, unknown_handler);
> >
> > - /* try to read an el2 sysreg from el0/1 */
> > - test_exception("", "mrs x0, sctlr_el2", "", "x0");
> > + /* try to read an el3 sysreg from el0/1/2 */
> > + test_exception("", "mrs x0, sctlr_el3", "", "x0");
> >
> > install_exception_handler(v, ESR_EL1_EC_UNKNOWN, NULL);
> >
> > @@ -429,6 +433,10 @@ int main(int argc, char **argv)
> > if (argc < 2)
> > report_abort("no test specified");
> >
> > +#if defined(__aarch64__)
> > + expected_level = current_level();
> nit I would directly use current_level() in the calling function,
> check_regs() to avoid that #ifdef
I can't move it into check_regs() because that's what's checking the exception
level of the handler is what the expected_level is.
Something like this (untested) would work:
diff --git a/arm/selftest.c b/arm/selftest.c
index 01691389..f173bc99 100644
--- a/arm/selftest.c
+++ b/arm/selftest.c
@@ -215,6 +215,7 @@ static void pabt_handler(struct pt_regs *regs)
static bool check_pabt(void)
{
+ expected_level = current_level();
install_exception_handler(EXCPTN_PABT, pabt_handler);
test_exception("ldr r9, =check_pabt_invalid_paddr\n"
@@ -318,6 +319,7 @@ static void unknown_handler(struct pt_regs *regs, unsigned int esr __unused)
static bool check_und(void)
{
enum vector v = check_vector_prep();
+ expected_level = current_level();
install_exception_handler(v, ESR_EL1_EC_UNKNOWN, unknown_handler);
@@ -340,6 +342,7 @@ static void svc_handler(struct pt_regs *regs, unsigned int esr)
static bool check_svc(void)
{
enum vector v = check_vector_prep();
+ expected_level = current_level();
install_exception_handler(v, ESR_EL1_EC_SVC64, svc_handler);
@@ -433,10 +436,6 @@ int main(int argc, char **argv)
if (argc < 2)
report_abort("no test specified");
-#if defined(__aarch64__)
- expected_level = current_level();
-#endif
-
report_prefix_push(argv[1]);
if (strcmp(argv[1], "setup") == 0) {
Is that preferable than an #ifdef?
Thanks,
Joey
>
> Eric
> > +#endif
> > +
> > report_prefix_push(argv[1]);
> >
> > if (strcmp(argv[1], "setup") == 0) {
>
next prev parent reply other threads:[~2025-12-02 12:21 UTC|newest]
Thread overview: 38+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-09-25 14:19 [kvm-unit-tests PATCH v3 00/10] arm64: EL2 support Joey Gouly
2025-09-25 14:19 ` [kvm-unit-tests PATCH v3 01/10] arm64: drop to EL1 if booted at EL2 Joey Gouly
2025-11-27 17:07 ` Eric Auger
2025-11-28 15:11 ` Joey Gouly
2025-09-25 14:19 ` [kvm-unit-tests PATCH v3 02/10] arm64: efi: initialise SCTLR_ELx fully Joey Gouly
2025-11-27 16:49 ` Eric Auger
2025-11-28 15:18 ` Joey Gouly
2025-09-25 14:19 ` [kvm-unit-tests PATCH v3 03/10] arm64: efi: initialise the EL Joey Gouly
2025-11-27 17:08 ` Eric Auger
2025-09-25 14:19 ` [kvm-unit-tests PATCH v3 04/10] arm64: timer: use hypervisor timers when at EL2 Joey Gouly
2025-12-02 8:36 ` Eric Auger
2025-09-25 14:19 ` [kvm-unit-tests PATCH v3 05/10] arm64: micro-bench: fix timer IRQ Joey Gouly
2025-12-02 8:36 ` Eric Auger
2025-09-25 14:19 ` [kvm-unit-tests PATCH v3 06/10] arm64: micro-bench: use smc when at EL2 Joey Gouly
2025-12-02 9:11 ` Eric Auger
2025-09-25 14:19 ` [kvm-unit-tests PATCH v3 07/10] arm64: selftest: update test for running " Joey Gouly
2025-12-02 9:16 ` Eric Auger
2025-12-02 12:21 ` Joey Gouly [this message]
2025-12-02 12:30 ` Eric Auger
2025-09-25 14:19 ` [kvm-unit-tests PATCH v3 08/10] arm64: pmu: count EL2 cycles Joey Gouly
2025-12-02 10:31 ` Eric Auger
2025-09-25 14:19 ` [kvm-unit-tests PATCH v3 09/10] arm64: run at EL2 if supported Joey Gouly
2025-12-02 10:35 ` Eric Auger
2025-09-25 14:19 ` [kvm-unit-tests PATCH v3 10/10] arm64: add EL2 environment variable Joey Gouly
2025-11-27 10:34 ` Eric Auger
2025-11-27 10:40 ` Joey Gouly
2025-12-01 23:34 ` Andrew Jones
2025-11-19 13:18 ` [kvm-unit-tests PATCH v3 00/10] arm64: EL2 support Joey Gouly
2025-11-19 13:48 ` Nadav Amit
2025-11-19 14:02 ` Joey Gouly
2025-11-19 15:34 ` Andrew Jones
2025-11-19 15:34 ` Marc Zyngier
2025-11-27 10:04 ` Eric Auger
2025-11-27 11:08 ` Joey Gouly
2025-11-27 12:04 ` Eric Auger
2025-11-27 14:52 ` Joey Gouly
2025-12-01 23:16 ` Andrew Jones
2025-12-02 14:22 ` Joey Gouly
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20251202122115.GA3921791@e124191.cambridge.arm.com \
--to=joey.gouly@arm.com \
--cc=alexandru.elisei@arm.com \
--cc=andrew.jones@linux.dev \
--cc=eric.auger@redhat.com \
--cc=kvm@vger.kernel.org \
--cc=kvmarm@lists.linux.dev \
--cc=maz@kernel.org \
--cc=oliver.upton@linux.dev \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox