| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
request_irq() is preferred over setup_irq(). Invocations of setup_irq()
occur after memory allocators are ready.
setup_irq() was required in older kernels as the memory allocator was not
available during early boot.
Hence replace setup_irq() by request_irq().
Signed-off-by: afzal mohammed <afzal.mohd.ma@gmail.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lkml.kernel.org/r/b060312689820559121ee0a6456bbc1202fb7ee5.1585320721.git.afzal.mohd.ma@gmail.com
|
|
request_irq() is preferred over setup_irq(). Invocations of setup_irq()
occur after memory allocators are ready.
setup_irq() was required in older kernels as the memory allocator was not
available during early boot.
Hence replace setup_irq() by request_irq().
Signed-off-by: afzal mohammed <afzal.mohd.ma@gmail.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lkml.kernel.org/r/e84ac60de8f747d49ce082659e51595f708c29d4.1585320721.git.afzal.mohd.ma@gmail.com
|
|
request_irq() is preferred over setup_irq(). Invocations of setup_irq()
occur after memory allocators are ready.
setup_irq() was required in older kernels as the memory allocator was not
available during early boot.
Hence replace setup_irq() by request_irq().
Signed-off-by: afzal mohammed <afzal.mohd.ma@gmail.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lkml.kernel.org/r/56e991e920ce5806771fab892574cba89a3d413f.1585320721.git.afzal.mohd.ma@gmail.com
|
|
request_irq() is preferred over setup_irq(). Invocations of setup_irq()
occur after memory allocators are ready.
setup_irq() was required in older kernels as the memory allocator was not
available during early boot.
Hence replace setup_irq() by request_irq().
Signed-off-by: afzal mohammed <afzal.mohd.ma@gmail.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Matt Turner <mattst88@gmail.com>
Link: https://lkml.kernel.org/r/51f8ae7da9f47a23596388141933efa2bdef317b.1585320721.git.afzal.mohd.ma@gmail.com
|
|
Now that we have HW-accelerated SGIs being delivered to VPEs, it
becomes required to map the VPEs on all ITSs instead of relying
on the lazy approach that we would use when using the ITS-list
mechanism.
Signed-off-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Zenghui Yu <yuzenghui@huawei.com>
Link: https://lore.kernel.org/r/20200304203330.4967-17-maz@kernel.org
|
|
Add the SGI configuration entry point for KVM to use.
Signed-off-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Zenghui Yu <yuzenghui@huawei.com>
Link: https://lore.kernel.org/r/20200304203330.4967-16-maz@kernel.org
|
|
Allocate per-VPE SGIs when initializing the GIC-specific part of the
VPE data structure.
Signed-off-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Zenghui Yu <yuzenghui@huawei.com>
Link: https://lore.kernel.org/r/20200304203330.4967-15-maz@kernel.org
|
|
In order to hide some of the differences between v4.0 and v4.1, move
the doorbell management out of the KVM code, and into the GICv4-specific
layer. This allows the calling code to ask for the doorbell when blocking,
and otherwise to leave the doorbell permanently disabled.
This matches the v4.1 code perfectly, and only results in a minor
refactoring of the v4.0 code.
Signed-off-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Zenghui Yu <yuzenghui@huawei.com>
Link: https://lore.kernel.org/r/20200304203330.4967-14-maz@kernel.org
|
|
Just like for vLPIs, there is some configuration information that cannot
be directly communicated through the normal irqchip API, and we have to
use our good old friend set_vcpu_affinity as a side-band communication
mechanism.
This is used to configure group and priority for a given vSGI.
Signed-off-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Zenghui Yu <yuzenghui@huawei.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Link: https://lore.kernel.org/r/20200304203330.4967-13-maz@kernel.org
|
|
To implement the get/set_irqchip_state callbacks (limited to the
PENDING state), we have to use a particular set of hacks:
- Reading the pending state is done by using a pair of new redistributor
registers (GICR_VSGIR, GICR_VSGIPENDR), which allow the 16 interrupts
state to be retrieved.
- Setting the pending state is done by generating it as we'd otherwise do
for a guest (writing to GITS_SGIR).
- Clearing the pending state is done by emitting a VSGI command with the
"clear" bit set.
This requires some interesting locking though:
- When talking to the redistributor, we must make sure that the VPE
affinity doesn't change, hence taking the VPE lock.
- At the same time, we must ensure that nobody accesses the same
redistributor's GICR_VSGIR registers for a different VPE, which
would corrupt the reading of the pending bits. We thus take the
per-RD spinlock. Much fun.
Signed-off-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Zenghui Yu <yuzenghui@huawei.com>
Link: https://lore.kernel.org/r/20200304203330.4967-12-maz@kernel.org
|
|
Implement mask/unmask for virtual SGIs by calling into the
configuration helper.
Signed-off-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Zenghui Yu <yuzenghui@huawei.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Link: https://lore.kernel.org/r/20200304203330.4967-11-maz@kernel.org
|
|
The GICv4.1 ITS has yet another new command (VSGI) which allows
a VPE-targeted SGI to be configured (or have its pending state
cleared). Add support for this command and plumb it into the
activate irqdomain callback so that it is ready to be used.
Signed-off-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Zenghui Yu <yuzenghui@huawei.com>
Link: https://lore.kernel.org/r/20200304203330.4967-10-maz@kernel.org
|
|
Since GICv4.1 has the capability to inject 16 SGIs into each VPE,
and that I'm keen not to invent too many specific interfaces to
manipulate these interrupts, let's pretend that each of these SGIs
is an actual Linux interrupt.
For that matter, let's introduce a minimal irqchip and irqdomain
setup that will get fleshed up in the following patches.
Signed-off-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Zenghui Yu <yuzenghui@huawei.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Link: https://lore.kernel.org/r/20200304203330.4967-9-maz@kernel.org
|
|
Sampling the IRQ line state in EOI and retriggering the interrupt to
work around missing level-triggered interrupt support only works for
non-threaded interrupts. Threaded interrupts must be retriggered the
same way in unmask callback.
Signed-off-by: Marek Vasut <marex@denx.de>
[maz: fixed missing static attribute]
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20200323235132.530550-1-marex@denx.de
|
|
irq_domain_update_bus_token should be called after checking for NULL
domain.
Signed-off-by: Liguang Zhang <zhangliguang@linux.alibaba.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/1583983255-44115-1-git-send-email-zhangliguang@linux.alibaba.com
|
|
Using a default domain on DT based platforms is unnecessary.
Signed-off-by: Mubin Sayyed <mubin.usman.sayyed@xilinx.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20200317125600.15913-5-mubin.usman.sayyed@xilinx.com
|
|
Register default arch handler via driver instead of directly pointing to
xilinx intc controller. This patch makes architecture code more generic.
Driver calls generic domain specific irq handler which does the most of
things self. Also get rid of concurrent_irq counting which hasn't been
exported anywhere.
Based on this loop was also optimized by using do/while loop instead of
goto loop.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Stefan Asserhall <stefan.asserhall@xilinx.com>
Link: https://lore.kernel.org/r/20200317125600.15913-4-mubin.usman.sayyed@xilinx.com
|
|
There is no ret filled in case of irq_domain_add_linear() failure.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Stefan Asserhall <stefan.asserhall@xilinx.com>
Link: https://lore.kernel.org/r/20200317125600.15913-3-mubin.usman.sayyed@xilinx.com
|
|
Added support for cascaded interrupt controllers.
Following cascaded configurations have been tested,
- peripheral->xilinx-intc->xilinx-intc->gic->Cortexa53 processor
on zcu102 board
- peripheral->xilinx-intc->xilinx-intc->microblaze processor
on kcu105 board
Signed-off-by: Mubin Sayyed <mubin.usman.sayyed@xilinx.com>
Signed-off-by: Anirudha Sarangi <anirudha.sarangi@xilinx.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20200317125600.15913-2-mubin.usman.sayyed@xilinx.com
|
|
Enable TCU support for Ingenic X1000, which can be supported by
the existing driver.
Signed-off-by: 周琰杰 (Zhou Yanjie) <zhouyanjie@wanyeetech.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/1584456160-40060-3-git-send-email-zhouyanjie@wanyeetech.com
|
|
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
This issue was found with the help of Coccinelle.
[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit 76497732932f ("cxgb3/l2t: Fix undefined behaviour")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20200319214531.GA21326@embeddedor.com
|
|
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
This issue was found with the help of Coccinelle.
[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit 76497732932f ("cxgb3/l2t: Fix undefined behaviour")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20200319214438.GA21123@embeddedor.com
|
|
Clear its own IRQs before the parent IRQ get enabled, so that the
remaining IRQs do not accidentally interrupt the parent IRQ controller.
This patch also fixes a reboot bug on OX820 SoC, where the remaining
rps-timer IRQ raises a GIC interrupt that is left pending. After that,
the rps-timer IRQ is cleared during driver initialization, and there's
no IRQ left in rps-irq when local_irq_enable() is called, which evokes
an error message "unexpected IRQ trap".
Fixes: bdd272cbb97a ("irqchip: versatile FPGA: support cascaded interrupts from DT")
Signed-off-by: Sungbo Eo <mans0n@gorani.run>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/20200321133842.2408823-1-mans0n@gorani.run
|
|
There is no special reason to set virtual LPI pending table as
non-shareable. If we choose to hard code the shareability without
probing, Inner-Shareable is likely to be a better choice, as the
VPEs can move around and benefit from having the redistributors
snooping each other's cache, if that's something they can do.
Furthermore, Hisilicon hip08 ends up with unspecified errors when
mixing shareability attributes. So let's move to IS attributes for
the VPT. This has also been tested on D05 and didn't show any
regression.
Signed-off-by: Heyi Guo <guoheyi@huawei.com>
[maz: rewrote commit message]
Signed-off-by: Marc Zyngier <maz@kernel.org>
Tested-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20191130073849.38378-1-guoheyi@huawei.com
|
|
One of the new features of GICv4.1 is to allow virtual SGIs to be
directly signaled to a VPE. For that, the ITS has grown a new
64kB page containing only a single register that is used to
signal a SGI to a given VPE.
Add a second mapping covering this new 64kB range, and take this
opportunity to limit the original mapping to 64kB, which is enough
to cover the span of the ITS registers.
Signed-off-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Zenghui Yu <yuzenghui@huawei.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Link: https://lore.kernel.org/r/20200304203330.4967-8-maz@kernel.org
|
|
Tell KVM that we support v4.1. Nothing uses this information so far.
Signed-off-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Zenghui Yu <yuzenghui@huawei.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Link: https://lore.kernel.org/r/20200304203330.4967-7-maz@kernel.org
|
|
The GICv4.1 spec says that it is CONTRAINED UNPREDICTABLE to write to
any of the GICR_INV{LPI,ALL}R registers if GICR_SYNCR.Busy == 1.
To deal with it, we must ensure that only a single invalidation can
happen at a time for a given redistributor. Add a per-RD lock to that
effect and take it around the invalidation/syncr-read to deal with this.
Signed-off-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Zenghui Yu <yuzenghui@huawei.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Link: https://lore.kernel.org/r/20200304203330.4967-6-maz@kernel.org
|
|
In GICv4.1, we emulate a guest-issued INVALL command by a direct write
to GICR_INVALLR. Before we finish the emulation and go back to guest,
let's make sure the physical invalidate operation is actually completed
and no stale data will be left in redistributor. Per the specification,
this can be achieved by polling the GICR_SYNCR.Busy bit (to zero).
Signed-off-by: Zenghui Yu <yuzenghui@huawei.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Link: https://lore.kernel.org/r/20200302092145.899-1-yuzenghui@huawei.com
Link: https://lore.kernel.org/r/20200304203330.4967-5-maz@kernel.org
|
|
Before GICv4.1, all operations would be serialized with the affinity
changes by virtue of using the same ITS command queue. With v4.1, things
change, as invalidations (and a number of other operations) are issued
using the redistributor MMIO frame.
We must thus make sure that these redistributor accesses cannot race
against aginst the affinity change, or we may end-up talking to the
wrong redistributor.
To ensure this, we expand the irq_to_cpuid() helper to take a spinlock
when the LPI is mapped to a vLPI (a new per-VPE lock) on each operation
that requires mutual exclusion.
Signed-off-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Zenghui Yu <yuzenghui@huawei.com>
Link: https://lore.kernel.org/r/20200304203330.4967-4-maz@kernel.org
|
|
In a system that is only sparsly populated with CPUs, we can end-up with
redistributors structures that are not initialized. Let's make sure we
don't try and access those when iterating over them (in this case when
checking we have a L2 VPE table).
Fixes: 4e6437f12d6e ("irqchip/gic-v4.1: Ensure L2 vPE table is allocated at RD level")
Signed-off-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Zenghui Yu <yuzenghui@huawei.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Link: https://lore.kernel.org/r/20200304203330.4967-3-maz@kernel.org
|
|
To allow the direct injection of SGIs into a guest, the GICv4.1
architecture has to sacrifice the Active state so that SGIs look
a lot like LPIs (they are injected by the same mechanism).
In order not to break existing software, the architecture gives
offers guests OSs the choice: SGIs with or without an active
state. It is the hypervisors duty to honor the guest's choice.
For this, the architecture offers a discovery bit indicating whether
the GIC supports GICv4.1 SGIs (GICD_TYPER2.nASSGIcap), and another
bit indicating whether the guest wants Active-less SGIs or not
(controlled by GICD_CTLR.nASSGIreq).
A hypervisor not supporting GICv4.1 SGIs would leave nASSGIcap
clear, and a guest not knowing about GICv4.1 SGIs (or definitely
wanting an Active state) would leave nASSGIreq clear (both being
thankfully backward compatible with older revisions of the GIC).
Since Linux is perfectly happy without an active state on SGIs,
inform the hypervisor that we'll use that if offered.
Signed-off-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Zenghui Yu <yuzenghui@huawei.com>
Link: https://lore.kernel.org/r/20200304203330.4967-2-maz@kernel.org
|
|
Enclose the chained handler with chained_irq_{enter,exit}(), so that the
muxed interrupts get properly acked.
This patch also fixes a reboot bug on OX820 SoC, where the jiffies timer
interrupt is never acked. The kernel waits a clock tick forever in
calibrate_delay_converge(), which leads to a boot hang.
Fixes: c41b16f8c9d9 ("ARM: integrator/versatile: consolidate FPGA IRQ handling code")
Signed-off-by: Sungbo Eo <mans0n@gorani.run>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20200319023448.1479701-1-mans0n@gorani.run
|
|
On a very heavily loaded D05 with GICv4, I managed to trigger the
following lockdep splat:
[ 6022.598864] ======================================================
[ 6022.605031] WARNING: possible circular locking dependency detected
[ 6022.611200] 5.6.0-rc4-00026-geee7c7b0f498 #680 Tainted: G E
[ 6022.618061] ------------------------------------------------------
[ 6022.624227] qemu-system-aar/7569 is trying to acquire lock:
[ 6022.629789] ffff042f97606808 (&p->pi_lock){-.-.}, at: try_to_wake_up+0x54/0x7a0
[ 6022.637102]
[ 6022.637102] but task is already holding lock:
[ 6022.642921] ffff002fae424cf0 (&irq_desc_lock_class){-.-.}, at: __irq_get_desc_lock+0x5c/0x98
[ 6022.651350]
[ 6022.651350] which lock already depends on the new lock.
[ 6022.651350]
[ 6022.659512]
[ 6022.659512] the existing dependency chain (in reverse order) is:
[ 6022.666980]
[ 6022.666980] -> #2 (&irq_desc_lock_class){-.-.}:
[ 6022.672983] _raw_spin_lock_irqsave+0x50/0x78
[ 6022.677848] __irq_get_desc_lock+0x5c/0x98
[ 6022.682453] irq_set_vcpu_affinity+0x40/0xc0
[ 6022.687236] its_make_vpe_non_resident+0x6c/0xb8
[ 6022.692364] vgic_v4_put+0x54/0x70
[ 6022.696273] vgic_v3_put+0x20/0xd8
[ 6022.700183] kvm_vgic_put+0x30/0x48
[ 6022.704182] kvm_arch_vcpu_put+0x34/0x50
[ 6022.708614] kvm_sched_out+0x34/0x50
[ 6022.712700] __schedule+0x4bc/0x7f8
[ 6022.716697] schedule+0x50/0xd8
[ 6022.720347] kvm_arch_vcpu_ioctl_run+0x5f0/0x978
[ 6022.725473] kvm_vcpu_ioctl+0x3d4/0x8f8
[ 6022.729820] ksys_ioctl+0x90/0xd0
[ 6022.733642] __arm64_sys_ioctl+0x24/0x30
[ 6022.738074] el0_svc_common.constprop.3+0xa8/0x1e8
[ 6022.743373] do_el0_svc+0x28/0x88
[ 6022.747198] el0_svc+0x14/0x40
[ 6022.750761] el0_sync_handler+0x124/0x2b8
[ 6022.755278] el0_sync+0x140/0x180
[ 6022.759100]
[ 6022.759100] -> #1 (&rq->lock){-.-.}:
[ 6022.764143] _raw_spin_lock+0x38/0x50
[ 6022.768314] task_fork_fair+0x40/0x128
[ 6022.772572] sched_fork+0xe0/0x210
[ 6022.776484] copy_process+0x8c4/0x18d8
[ 6022.780742] _do_fork+0x88/0x6d8
[ 6022.784478] kernel_thread+0x64/0x88
[ 6022.788563] rest_init+0x30/0x270
[ 6022.792390] arch_call_rest_init+0x14/0x1c
[ 6022.796995] start_kernel+0x498/0x4c4
[ 6022.801164]
[ 6022.801164] -> #0 (&p->pi_lock){-.-.}:
[ 6022.806382] __lock_acquire+0xdd8/0x15c8
[ 6022.810813] lock_acquire+0xd0/0x218
[ 6022.814896] _raw_spin_lock_irqsave+0x50/0x78
[ 6022.819761] try_to_wake_up+0x54/0x7a0
[ 6022.824018] wake_up_process+0x1c/0x28
[ 6022.828276] wakeup_softirqd+0x38/0x40
[ 6022.832533] __tasklet_schedule_common+0xc4/0xf0
[ 6022.837658] __tasklet_schedule+0x24/0x30
[ 6022.842176] check_irq_resend+0xc8/0x158
[ 6022.846609] irq_startup+0x74/0x128
[ 6022.850606] __enable_irq+0x6c/0x78
[ 6022.854602] enable_irq+0x54/0xa0
[ 6022.858431] its_make_vpe_non_resident+0xa4/0xb8
[ 6022.863557] vgic_v4_put+0x54/0x70
[ 6022.867469] kvm_arch_vcpu_blocking+0x28/0x38
[ 6022.872336] kvm_vcpu_block+0x48/0x490
[ 6022.876594] kvm_handle_wfx+0x18c/0x310
[ 6022.880938] handle_exit+0x138/0x198
[ 6022.885022] kvm_arch_vcpu_ioctl_run+0x4d4/0x978
[ 6022.890148] kvm_vcpu_ioctl+0x3d4/0x8f8
[ 6022.894494] ksys_ioctl+0x90/0xd0
[ 6022.898317] __arm64_sys_ioctl+0x24/0x30
[ 6022.902748] el0_svc_common.constprop.3+0xa8/0x1e8
[ 6022.908046] do_el0_svc+0x28/0x88
[ 6022.911871] el0_svc+0x14/0x40
[ 6022.915434] el0_sync_handler+0x124/0x2b8
[ 6022.919951] el0_sync+0x140/0x180
[ 6022.923773]
[ 6022.923773] other info that might help us debug this:
[ 6022.923773]
[ 6022.931762] Chain exists of:
[ 6022.931762] &p->pi_lock --> &rq->lock --> &irq_desc_lock_class
[ 6022.931762]
[ 6022.942101] Possible unsafe locking scenario:
[ 6022.942101]
[ 6022.948007] CPU0 CPU1
[ 6022.952523] ---- ----
[ 6022.957039] lock(&irq_desc_lock_class);
[ 6022.961036] lock(&rq->lock);
[ 6022.966595] lock(&irq_desc_lock_class);
[ 6022.973109] lock(&p->pi_lock);
[ 6022.976324]
[ 6022.976324] *** DEADLOCK ***
This is happening because we have a pending doorbell that requires
retrigger. As SW retriggering is done in a tasklet, we trigger the
circular dependency above.
The easy cop-out is to provide a retrigger callback that doesn't
require acquiring any extra lock.
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20200310184921.23552-5-maz@kernel.org
|
|
The irq_retrigger callback is supposed to return 0 when retrigger
has failed, and a non-zero value otherwise. Tell the core code
that the driver has succedded in using the HW to retrigger the
interrupt (if ever).
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20200310184921.23552-4-maz@kernel.org
|
|
The irq_retrigger callback is supposed to return 0 when retrigger
has failed, and a non-zero value otherwise. Tell the core code
that the driver has succedded in using the HW to retrigger the
interrupt.
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20200310184921.23552-3-maz@kernel.org
|
|
The irq_retrigger callback is supposed to return 0 when retrigger
has failed, and a non-zero value otherwise. Tell the core code
that the driver has succedded in using the HW to retrigger the
interrupt.
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20200310184921.23552-2-maz@kernel.org
|
|
The GICv3 ITS driver assumes that once it has latched on a page size for
a given BASER register, it can use the same page size as the maximum
page size for all subsequent BASER registers.
Although it worked so far, nothing in the architecture guarantees this,
and Nianyao Tang hit this problem on some undisclosed implementation.
Let's bite the bullet and probe the the supported page size on all BASER
registers before starting to populate the tables. This simplifies the
setup a bit, at the expense of a few additional MMIO accesses.
Signed-off-by: Marc Zyngier <maz@kernel.org>
Reported-by: Nianyao Tang <tangnianyao@huawei.com>
Tested-by: Nianyao Tang <tangnianyao@huawei.com>
Link: https://lore.kernel.org/r/1584089195-63897-1-git-send-email-zhangshaokun@hisilicon.com
|
|
Per the spec, the BCM2835's IRQs are all disabled when coming out of
power-on reset. Its IRQ driver assumes that's still the case when the
kernel boots and does not perform any initialization of the registers.
However the Raspberry Pi Foundation's bootloader leaves the USB
interrupt enabled when handing over control to the kernel.
Quiesce IRQs and the FIQ if they were left enabled and log a message to
let users know that they should update the bootloader once a fixed
version is released.
If the USB interrupt is not quiesced and the USB driver later on claims
the FIQ (as it does on the Raspberry Pi Foundation's downstream kernel),
interrupt latency for all other peripherals increases and occasional
lockups occur. That's because both the FIQ and the normal USB interrupt
fire simultaneously:
On a multicore Raspberry Pi, if normal interrupts are routed to CPU 0
and the FIQ to CPU 1 (hardcoded in the Foundation's kernel), then a USB
interrupt causes CPU 0 to spin in bcm2836_chained_handle_irq() until the
FIQ on CPU 1 has cleared it. Other peripherals' interrupts are starved
as long. I've seen CPU 0 blocked for up to 2.9 msec. eMMC throughput
on a Compute Module 3 irregularly dips to 23.0 MB/s without this commit
but remains relatively constant at 23.5 MB/s with this commit.
The lockups occur when CPU 0 receives a USB interrupt while holding a
lock which CPU 1 is trying to acquire while the FIQ is temporarily
disabled on CPU 1. At best users get RCU CPU stall warnings, but most
of the time the system just freezes.
Fixes: 89214f009c1d ("ARM: bcm2835: add interrupt controller driver")
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/f97868ba4e9b86ddad71f44ec9d8b3b7d8daa1ea.1582618537.git.lukas@wunner.de
|
|
Current, PLIC driver can support only 1 PLIC on the board. However,
there can be multiple PLICs present on a two socket systems in RISC-V.
Modify the driver so that each PLIC handler can have a information
about individual PLIC registers and an irqdomain associated with it.
Tested on two socket RISC-V system based on VCU118 FPGA connected via
OmniXtend protocol.
Signed-off-by: Atish Patra <atish.patra@wdc.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Anup Patel <anup@brainfault.org>
Link: https://lore.kernel.org/r/20200302231146.15530-3-atish.patra@wdc.com
|
|
Currently, PLIC threshold is only initialized once in the beginning.
However, threshold can be set to disabled if a CPU is marked offline with
CPU hotplug feature. This will not allow to change the irq affinity to a
CPU that just came online.
Add PLIC specific CPU hotplug callbacks and enable the threshold when a CPU
comes online. Take this opportunity to move the external interrupt enable
code from trap init to PLIC driver as well. On cpu offline path, the driver
performs the exact opposite operations i.e. disable the interrupt and
the threshold.
Signed-off-by: Atish Patra <atish.patra@wdc.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Anup Patel <anup@brainfault.org>
Link: https://lore.kernel.org/r/20200302231146.15530-2-atish.patra@wdc.com
|
|
Modern x86 laptops are starting to use GPIO pins as interrupts more
and more, e.g. touchpads and touchscreens have almost all moved away
from PS/2 and USB to using I2C with a GPIO pin as interrupt.
Modern x86 laptops also have almost all moved to using s2idle instead
of using the system S3 ACPI power state to suspend.
The Intel and AMD pinctrl drivers do not define irq_retrigger handlers
for the irqchips they register, this is causing edge triggered interrupts
which happen while suspended using s2idle to get lost.
One specific example of this is the lid switch on some devices, lid
switches used to be handled by the embedded-controller, but now the
lid open/closed sensor is sometimes directly connected to a GPIO pin.
On most devices the ACPI code for this looks like this:
Method (_E00, ...) {
Notify (LID0, 0x80) // Status Change
}
Where _E00 is an ACPI event handler for changes on both edges of the GPIO
connected to the lid sensor, this event handler is then combined with an
_LID method which directly reads the pin. When the device is resumed by
opening the lid, the GPIO interrupt will wake the system, but because the
pinctrl irqchip doesn't have an irq_retrigger handler, the Notify will not
happen. This is not a problem in the case the _LID method directly reads
the GPIO, because the drivers/acpi/button.c code will call _LID on resume
anyways.
But some devices have an event handler for the GPIO connected to the
lid sensor which looks like this:
Method (_E00, ...) {
if (LID_GPIO == One)
LIDS = One
else
LIDS = Zero
Notify (LID0, 0x80) // Status Change
}
And the _LID method returns the cached LIDS value, since on open we
do not re-run the edge-interrupt handler when we re-enable IRQS on resume
(because of the missing irq_retrigger handler), _LID now will keep
reporting closed, as LIDS was never changed to reflect the open status,
this causes userspace to re-resume the laptop again shortly after opening
the lid.
The Intel GPIO controllers do not allow implementing irq_retrigger without
emulating it in software, at which point we are better of just using the
generic HARDIRQS_SW_RESEND mechanism rather then re-implementing software
emulation for this separately in aprox. 14 different pinctrl drivers.
Select HARDIRQS_SW_RESEND to solve the problem of edge-triggered GPIO
interrupts not being re-triggered on resume when they were triggered during
suspend (s2idle) and/or when they were the cause of the wakeup.
This requires
008f1d60fe25 ("x86/apic/vector: Force interupt handler invocation to irq context")
c16816acd086 ("genirq: Add protection against unsafe usage of generic_handle_irq()")
to protect the APIC based interrupts from being wreckaged by a software
resend.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lkml.kernel.org/r/20200123210242.53367-1-hdegoede@redhat.com
|
|
request_irq() is preferred over setup_irq(). Invocations of setup_irq()
occur after memory allocators are ready.
Per tglx[1], setup_irq() existed in olden days when allocators were not
ready by the time early interrupts were initialized.
Hence replace setup_irq() by request_irq().
[1] https://lkml.kernel.org/r/alpine.DEB.2.20.1710191609480.1971@nanos
Signed-off-by: afzal mohammed <afzal.mohd.ma@gmail.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20200304004839.4729-1-afzal.mohd.ma@gmail.com
|
|
Restore alignment of the continuation of the devm_ioremap() call in
intc_irqpin_probe().
Fixes: 4bdc0d676a643140 ("remove ioremap_nocache and devm_ioremap_nocache")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20200212084744.9376-1-geert+renesas@glider.be
|
|
Add COMPILE_TEST support to IMX_INTMUX driver for better compile
testing coverage.
Signed-off-by: Anson Huang <Anson.Huang@nxp.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/1583588547-7164-1-git-send-email-Anson.Huang@nxp.com
|
|
GICR_SYNCR is a 32bit register, so it is better to access it with
32bit access width, though we have not seen any real problem.
Signed-off-by: Heyi Guo <guoheyi@huawei.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20200225090023.28020-1-guoheyi@huawei.com
|
|
GPIO hardware block is directly linked to EXTI block but EXTI handles
external interrupts only on edge. To be able to handle GPIO interrupt on
level a "hack" is done in gpio irq chip: parent interrupt (exti irq chip)
is retriggered following interrupt type and gpio line value.
Signed-off-by: Alexandre Torgue <alexandre.torgue@st.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Tested-by: Marek Vasut <marex@denx.de>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://lore.kernel.org/r/20200219143229.18084-3-alexandre.torgue@st.com
|
|
This commit introduces retrigger support for stm32_ext_h chip.
It consists to rise the GIC interrupt mapped to an EXTI line.
Signed-off-by: Alexandre Torgue <alexandre.torgue@st.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Tested-by: Marek Vasut <marex@denx.de>
Link: https://lore.kernel.org/r/20200219143229.18084-2-alexandre.torgue@st.com
|
|
When transitioning some elder platforms to device tree it
becomes necessary to cascade VIC IRQ chips off another
interrupt controller.
Tested with the cascaded VIC on the Integrator/AP attached
logic module IM-PD1.
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20200219153543.137153-1-linus.walleij@linaro.org
|
|
irq_domain_alloc_irqs_hierarchy() has 3 call sites in the compilation unit
but only one of them checks for the pointer which is being dereferenced
inside the called function. Move the check into the function. This allows
for catching the error instead of the following crash:
Unable to handle kernel NULL pointer dereference at virtual address 00000000
PC is at 0x0
LR is at gpiochip_hierarchy_irq_domain_alloc+0x11f/0x140
...
[<c06c23ff>] (gpiochip_hierarchy_irq_domain_alloc)
[<c0462a89>] (__irq_domain_alloc_irqs)
[<c0462dad>] (irq_create_fwspec_mapping)
[<c06c2251>] (gpiochip_to_irq)
[<c06c1c9b>] (gpiod_to_irq)
[<bf973073>] (gpio_irqs_init [gpio_irqs])
[<bf974048>] (gpio_irqs_exit+0xecc/0xe84 [gpio_irqs])
Code: bad PC value
Signed-off-by: Alexander Sverdlin <alexander.sverdlin@nokia.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lkml.kernel.org/r/20200306174720.82604-1-alexander.sverdlin@nokia.com
|
|
The AER error injection mechanism just blindly abuses generic_handle_irq()
which is really not meant for consumption by random drivers. The include of
linux/irq.h should have been a red flag in the first place. Driver code,
unless implementing interrupt chips or low level hypervisor functionality
has absolutely no business with that.
Invoking generic_handle_irq() from non interrupt handling context can have
nasty side effects at least on x86 due to the hardware trainwreck which
makes interrupt affinity changes a fragile beast. Sathyanarayanan triggered
a NULL pointer dereference in the low level APIC code that way. While the
particular pointer could be checked this would only paper over the issue
because there are other ways to trigger warnings or silently corrupt state.
Invoke the new irq_inject_interrupt() mechanism, which has the necessary
sanity checks in place and injects the interrupt via the irq_retrigger()
mechanism, which is at least halfways safe vs. the fragile x86 affinity
change mechanics.
It's safe on x86 as it does not corrupt state, but it still can cause a
premature completion of an interrupt affinity change causing the interrupt
line to become stale. Very unlikely, but possible.
For regular operations this is a non issue as AER error injection is meant
for debugging and testing and not for usage on production systems. People
using this should better know what they are doing.
Fixes: 390e2db82480 ("PCI/AER: Abstract AER interrupt handling")
Reported-by: sathyanarayanan.kuppuswamy@linux.intel.com
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Reviewed-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Cc: Bjorn Helgaas <bhelgaas@google.com>
Link: https://lkml.kernel.org/r/20200306130624.098374457@linutronix.de
|
afzal mohammed
Marc Zyngier
Marek Vasut
luanshi
Mubin Sayyed
Michal Simek
周琰杰 (Zhou Yanjie)
Gustavo A. R. Silva
Sungbo Eo
Heyi Guo
Zenghui Yu
Lukas Wunner
Atish Patra
Hans de Goede
Geert Uytterhoeven
Anson Huang
Alexandre Torgue
Linus Walleij
Alexander Sverdlin
Thomas Gleixner