diff options
author | Marc Zyngier <maz@kernel.org> | 2025-02-15 21:52:41 +0300 |
---|---|---|
committer | Greg Kroah-Hartman <gregkh@linuxfoundation.org> | 2025-02-27 15:30:23 +0300 |
commit | ada54b2e81ebb780ad28687ef91af46d016a3fb7 (patch) | |
tree | 209160cb7760359b101f571d5b50573abd8ecbbe /tools/perf/scripts/python/syscall-counts.py | |
parent | 562d40ff3d23d22999e863d8863a811ca50d5cc6 (diff) | |
download | linux-ada54b2e81ebb780ad28687ef91af46d016a3fb7.tar.xz |
irqchip/gic-v3: Fix rk3399 workaround when secure interrupts are enabled
commit 4cb77793842a351b39a030f77caebace3524840e upstream.
Christoph reports that their rk3399 system dies since commit 773c05f417fa1
("irqchip/gic-v3: Work around insecure GIC integrations").
It appears that some rk3399 have secure payloads, and that the firmware
sets SCR_EL3.FIQ==1. Obivously, disabling security in that configuration
leads to even more problems.
Revisit the workaround by:
- making it rk3399 specific
- checking whether Group-0 is available, which is a good proxy
for SCR_EL3.FIQ being 0
- either apply the workaround if Group-0 is available, or disable
pseudo-NMIs if not
Note that this doesn't mean that the secure side is able to receive
interrupts, as all interrupts are made non-secure anyway.
Clearly, nobody ever tested secure interrupts on this platform.
Fixes: 773c05f417fa1 ("irqchip/gic-v3: Work around insecure GIC integrations")
Reported-by: Christoph Fritz <chf.fritz@googlemail.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: Christoph Fritz <chf.fritz@googlemail.com>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/all/20250215185241.3768218-1-maz@kernel.org
Closes: https://lore.kernel.org/r/b1266652fb64857246e8babdf268d0df8f0c36d9.camel@googlemail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Diffstat (limited to 'tools/perf/scripts/python/syscall-counts.py')
0 files changed, 0 insertions, 0 deletions