kernel.git/drivers/irqchip, branch linux-5.17.y

irqchip: irq-xtensa-mx: fix initial IRQ affinity

2022-06-09T08:26:26Z

commit a255ee29252066d621df5d6b420bf534c6ba5bc0 upstream. When irq-xtensa-mx chip is used in non-SMP configuration its irq_set_affinity callback is not called leaving IRQ affinity set empty. As a result IRQ delivery does not work in that configuration. Initialize IRQ affinity of the xtensa MX interrupt distributor to CPU 0 for all external IRQ lines. Cc: stable@vger.kernel.org Signed-off-by: Max Filippov Signed-off-by: Greg Kroah-Hartman

irqchip/armada-370-xp: Do not touch Performance Counter Overflow on A375, A38x, A39x

2022-06-09T08:26:26Z

commit a3d66a76348daf559873f19afc912a2a7c2ccdaf upstream. Register ARMADA_370_XP_INT_FABRIC_MASK_OFFS is Armada 370 and XP specific and on new Armada platforms it has different meaning. It does not configure Performance Counter Overflow interrupt masking. So do not touch this register on non-A370/XP platforms (A375, A38x and A39x). Signed-off-by: Pali Rohár Cc: stable@vger.kernel.org Fixes: 28da06dfd9e4 ("irqchip: armada-370-xp: Enable the PMU interrupts") Reviewed-by: Andrew Lunn Signed-off-by: Marc Zyngier Link: https://lore.kernel.org/r/20220425113706.29310-1-pali@kernel.org Signed-off-by: Greg Kroah-Hartman

irqchip/gic-v3: Fix priority mask handling

2022-06-09T08:25:53Z

[ Upstream commit 614ab80c96474682157cabb14f8c8602b3422e90 ] When a kernel is built with CONFIG_ARM64_PSEUDO_NMI=y and pseudo-NMIs are enabled at runtime, GICv3's gic_handle_irq() can leave DAIF and ICC_PMR_EL1 in an unexpected state in some cases, breaking subsequent usage of local_irq_enable() and resulting in softirqs being run with IRQs erroneously masked (possibly resulting in deadlocks). This can happen when an IRQ exception is taken from a context where regular IRQs were unmasked, and either: (1) ICC_IAR1_EL1 indicates a special INTID (e.g. as a result of an IRQ being withdrawn since the IRQ exception was taken). (2) ICC_IAR1_EL1 and ICC_RPR_EL1 indicate an NMI was acknowledged. When an NMI is taken from a context where regular IRQs were masked, there is no problem. When CONFIG_ARM64_DEBUG_PRIORITY_MASKING=y, this can be detected with perf, e.g. | # ./perf record -a -g -e cycles:k ls -alR / > /dev/null 2>&1 | ------------[ cut here ]------------ | WARNING: CPU: 0 PID: 14 at arch/arm64/include/asm/irqflags.h:32 arch_local_irq_enable+0x4c/0x6c | Modules linked in: | CPU: 0 PID: 14 Comm: ksoftirqd/0 Not tainted 5.18.0-rc5-00004-g876c38e3d20b #12 | Hardware name: linux,dummy-virt (DT) | pstate: 204000c5 (nzCv daIF +PAN -UAO -TCO -DIT -SSBS BTYPE=--) | pc : arch_local_irq_enable+0x4c/0x6c | lr : __do_softirq+0x110/0x5d8 | sp : ffff8000080bbbc0 | pmr_save: 000000f0 | x29: ffff8000080bbbc0 x28: ffff316ac3a6ca40 x27: 0000000000000000 | x26: 0000000000000000 x25: ffffa04611c06008 x24: ffffa04611c06008 | x23: 0000000040400005 x22: 0000000000000200 x21: ffff8000080bbe20 | x20: ffffa0460fe10320 x19: 0000000000000009 x18: 0000000000000000 | x17: ffff91252dfa9000 x16: ffff800008004000 x15: 0000000000004000 | x14: 0000000000000028 x13: ffffa0460fe17578 x12: ffffa0460fed4294 | x11: ffffa0460fedc168 x10: ffffffffffffff80 x9 : ffffa0460fe10a70 | x8 : ffffa0460fedc168 x7 : 000000000000b762 x6 : 00000000057c3bdf | x5 : ffff8000080bbb18 x4 : 0000000000000000 x3 : 0000000000000001 | x2 : ffff91252dfa9000 x1 : 0000000000000060 x0 : 00000000000000f0 | Call trace: | arch_local_irq_enable+0x4c/0x6c | __irq_exit_rcu+0x180/0x1ac | irq_exit_rcu+0x1c/0x44 | el1_interrupt+0x4c/0xe4 | el1h_64_irq_handler+0x18/0x24 | el1h_64_irq+0x74/0x78 | smpboot_thread_fn+0x68/0x2c0 | kthread+0x124/0x130 | ret_from_fork+0x10/0x20 | irq event stamp: 193241 | hardirqs last enabled at (193240): [] __do_softirq+0x10c/0x5d8 | hardirqs last disabled at (193241): [] el1_dbg+0x24/0x90 | softirqs last enabled at (193234): [] __do_softirq+0x470/0x5d8 | softirqs last disabled at (193239): [] __irq_exit_rcu+0x180/0x1ac | ---[ end trace 0000000000000000 ]--- The necessary manipulation of DAIF and ICC_PMR_EL1 depends on the interrupted context, but the structure of gic_handle_irq() makes this also depend on whether the GIC reports an IRQ, NMI, or special INTID: * When the interrupted context had regular IRQs masked (and hence the interrupt must be an NMI), the entry code performs the NMI entry/exit and gic_handle_irq() should return with DAIF and ICC_PMR_EL1 unchanged. This is handled correctly today. * When the interrupted context had regular IRQs unmasked, the entry code performs IRQ entry/exit, but expects gic_handle_irq() to always update ICC_PMR_EL1 and DAIF.IF to unmask NMIs (but not regular IRQs) prior to returning (which it must do prior to invoking any regular IRQ handler). This unbalanced calling convention is necessary because we don't know whether an NMI has been taken until acknowledged by a read from ICC_IAR1_EL1, and so we need to perform the read with NMI masked in case an NMI has been taken (and needs to be handled with NMIs masked). Unfortunately, this is not handled consistently: - When ICC_IAR1_EL1 reports a special INTID, gic_handle_irq() returns immediately without manipulating ICC_PMR_EL1 and DAIF. - When RPR_EL1 indicates an NMI, gic_handle_irq() calls gic_handle_nmi() to invoke the NMI handler, then returns without manipulating ICC_PMR_EL1 and DAIF. - For regular IRQs, gic_handle_irq() manipulates ICC_PMR_EL1 and DAIF prior to invoking the IRQ handler. There were related problems with special INTID handling in the past, where if an exception was taken from a context with regular IRQs masked and ICC_IAR_EL1 reported a special INTID, gic_handle_irq() would erroneously unmask NMIs in NMI context permitted an unexpected nested NMI. That case specifically was fixed by commit: a97709f563a078e2 ("irqchip/gic-v3: Do not enable irqs when handling spurious interrups") ... but unfortunately that commit added an inverse problem, where if an exception was taken from a context with regular IRQs *unmasked* and ICC_IAR_EL1 reported a special INTID, gic_handle_irq() would erroneously fail to unmask NMIs (and consequently regular IRQs could not be unmasked during softirq processing). Before and after that commit, if an NMI was taken from a context with regular IRQs unmasked gic_handle_irq() would not unmask NMIs prior to returning, leading to the same problem with softirq handling. This patch fixes this by restructuring gic_handle_irq(), splitting it into separate irqson/irqsoff helper functions which consistently perform the DAIF + ICC_PMR1_EL1 manipulation based upon the interrupted context, regardless of the event indicated by ICC_IAR1_EL1. The special INTID handling is moved into the low-level IRQ/NMI handler invocation helper functions, so that early returns don't prevent the required manipulation of DAIF + ICC_PMR_EL1. Fixes: f32c926651dcd168 ("irqchip/gic-v3: Handle pseudo-NMIs") Signed-off-by: Mark Rutland Cc: Marc Zyngier Cc: Thomas Gleixner Signed-off-by: Marc Zyngier Link: https://lore.kernel.org/r/20220513133038.226182-4-mark.rutland@arm.com Signed-off-by: Sasha Levin

irqchip/gic-v3: Refactor ISB + EOIR at ack time

2022-06-09T08:25:53Z

[ Upstream commit 6efb50923771f392122f5ce69dfc43b08f16e449 ] There are cases where a context synchronization event is necessary between an IRQ being raised and being handled, and there are races such that we cannot rely upon the exception entry being subsequent to the interrupt being raised. To fix this, we place an ISB between a read of IAR and the subsequent invocation of an IRQ handler. When EOI mode 1 is in use, we need to EOI an interrupt prior to invoking its handler, and we have a write to EOIR for this. As this write to EOIR requires an ISB, and this is provided by the gic_write_eoir() helper, we omit the usual ISB in this case, with the logic being: | if (static_branch_likely(&supports_deactivate_key)) | gic_write_eoir(irqnr); | else | isb(); This is somewhat opaque, and it would be a little clearer if there were an unconditional ISB, with only the write to EOIR being conditional, e.g. | if (static_branch_likely(&supports_deactivate_key)) | write_gicreg(irqnr, ICC_EOIR1_EL1); | | isb(); This patch rewrites the code that way, with this logic factored into a new helper function with comments explaining what the ISB is for, as were originally laid out in commit: 39a06b67c2c1256b ("irqchip/gic: Ensure we have an ISB between ack and ->handle_irq") Note that since then, we removed the IAR polling in commit: 342677d70ab92142 ("irqchip/gic-v3: Remove acknowledge loop") ... which removed one of the two race conditions. For consistency, other portions of the driver are made to manipulate EOIR using write_gicreg() and explcit ISBs, and the gic_write_eoir() helper function is removed. There should be no functional change as a result of this patch. Signed-off-by: Mark Rutland Cc: Marc Zyngier Cc: Thomas Gleixner Cc: Will Deacon Signed-off-by: Marc Zyngier Link: https://lore.kernel.org/r/20220513133038.226182-3-mark.rutland@arm.com Signed-off-by: Sasha Levin

irqchip/gic-v3: Ensure pseudo-NMIs have an ISB between ack and handling

2022-06-09T08:25:53Z

[ Upstream commit adf14453d2c037ab529040c1186ea32e277e783a ] There are cases where a context synchronization event is necessary between an IRQ being raised and being handled, and there are races such that we cannot rely upon the exception entry being subsequent to the interrupt being raised. We identified and fixes this for regular IRQs in commit: 39a06b67c2c1256b ("irqchip/gic: Ensure we have an ISB between ack and ->handle_irq") Unfortunately, we forgot to do the same for psuedo-NMIs when support for those was added in commit: f32c926651dcd168 ("irqchip/gic-v3: Handle pseudo-NMIs") Which means that when pseudo-NMIs are used for PMU support, we'll hit the same problem. Apply the same fix as for regular IRQs. Note that when EOI mode 1 is in use, the call to gic_write_eoir() will provide an ISB. Fixes: f32c926651dcd168 ("irqchip/gic-v3: Handle pseudo-NMIs") Signed-off-by: Mark Rutland Cc: Marc Zyngier Cc: Thomas Gleixner Cc: Will Deacon Signed-off-by: Marc Zyngier Link: https://lore.kernel.org/r/20220513133038.226182-2-mark.rutland@arm.com Signed-off-by: Sasha Levin

irqchip/aspeed-scu-ic: Fix irq_of_parse_and_map() return value

2022-06-09T08:25:47Z

[ Upstream commit f03a9670d27d23fe734a456f16e2579b21ec02b4 ] The irq_of_parse_and_map() returns 0 on failure, not a negative ERRNO. Fixes: 04f605906ff0 ("irqchip: Add Aspeed SCU interrupt controller") Signed-off-by: Krzysztof Kozlowski Signed-off-by: Marc Zyngier Link: https://lore.kernel.org/r/20220423094227.33148-2-krzysztof.kozlowski@linaro.org Signed-off-by: Sasha Levin

irqchip/aspeed-i2c-ic: Fix irq_of_parse_and_map() return value

2022-06-09T08:25:47Z

[ Upstream commit 50f0f26e7c8665763d0d7d3372dbcf191f94d077 ] The irq_of_parse_and_map() returns 0 on failure, not a negative ERRNO. Fixes: f48e699ddf70 ("irqchip/aspeed-i2c-ic: Add I2C IRQ controller for Aspeed") Signed-off-by: Krzysztof Kozlowski Signed-off-by: Marc Zyngier Link: https://lore.kernel.org/r/20220423094227.33148-1-krzysztof.kozlowski@linaro.org Signed-off-by: Sasha Levin

irqchip/exiu: Fix acknowledgment of edge triggered interrupts

2022-06-09T08:25:46Z

[ Upstream commit 4efc851c36e389f7ed432edac0149acc5f94b0c7 ] Currently the EXIU uses the fasteoi interrupt flow that is configured by it's parent (irq-gic-v3.c). With this flow the only chance to clear the interrupt request happens during .irq_eoi() and (obviously) this happens after the interrupt handler has run. EXIU requires edge triggered interrupts to be acked prior to interrupt handling. Without this we risk incorrect interrupt dismissal when a new interrupt is delivered after the handler reads and acknowledges the peripheral but before the irq_eoi() takes place. Fix this by clearing the interrupt request from .irq_ack() if we are configured for edge triggered interrupts. This requires adopting the fasteoi-ack flow instead of the fasteoi to ensure the ack gets called. These changes have been tested using the power button on a Developerbox/SC2A11 combined with some hackery in gpio-keys so I can play with the different trigger mode [and an mdelay(500) so I can can check what happens on a double click in both modes]. Fixes: 706cffc1b912 ("irqchip/exiu: Add support for Socionext Synquacer EXIU controller") Signed-off-by: Daniel Thompson Reviewed-by: Ard Biesheuvel Signed-off-by: Marc Zyngier Link: https://lore.kernel.org/r/20220503134541.2566457-1-daniel.thompson@linaro.org Signed-off-by: Sasha Levin

irqchip/gic, gic-v3: Prevent GSI to SGI translations

2022-04-13T17:27:42Z

commit 544808f7e21cb9ccdb8f3aa7de594c05b1419061 upstream. At the moment the GIC IRQ domain translation routine happily converts ACPI table GSI numbers below 16 to GIC SGIs (Software Generated Interrupts aka IPIs). On the Devicetree side we explicitly forbid this translation, actually the function will never return HWIRQs below 16 when using a DT based domain translation. We expect SGIs to be handled in the first part of the function, and any further occurrence should be treated as a firmware bug, so add a check and print to report this explicitly and avoid lengthy debug sessions. Fixes: 64b499d8df40 ("irqchip/gic-v3: Configure SGIs as standard interrupts") Signed-off-by: Andre Przywara Signed-off-by: Marc Zyngier Link: https://lore.kernel.org/r/20220404110842.2882446-1-andre.przywara@arm.com Signed-off-by: Greg Kroah-Hartman

irqchip/gic-v4: Wait for GICR_VPENDBASER.Dirty to clear before descheduling

2022-04-13T17:27:42Z

commit af27e41612ec7e5b4783f589b753a7c31a37aac8 upstream. The way KVM drives GICv4.{0,1} is as follows: - vcpu_load() makes the VPE resident, instructing the RD to start scanning for interrupts - just before entering the guest, we check that the RD has finished scanning and that we can start running the vcpu - on preemption, we deschedule the VPE by making it invalid on the RD However, we are preemptible between the first two steps. If it so happens *and* that the RD was still scanning, we nonetheless write to the GICR_VPENDBASER register while Dirty is set, and bad things happen (we're in UNPRED land). This affects both the 4.0 and 4.1 implementations. Make sure Dirty is cleared before performing the deschedule, meaning that its_clear_vpend_valid() becomes a sort of full VPE residency barrier. Reported-by: Jingyi Wang Tested-by: Nianyao Tang Signed-off-by: Marc Zyngier Fixes: 57e3cebd022f ("KVM: arm64: Delay the polling of the GICR_VPENDBASER.Dirty bit") Link: https://lore.kernel.org/r/4aae10ba-b39a-5f84-754b-69c2eb0a2c03@huawei.com Signed-off-by: Greg Kroah-Hartman