[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <6d6d18be-69ab-41f0-bae5-db7d0a12196d@gmail.com>
Date: Tue, 30 Sep 2025 07:19:06 +0530
From: bhanuseshukumar <bhanuseshukumar@...il.com>
To: Manivannan Sadhasivam <mani@...nel.org>
Cc: Krzysztof WilczyĆski <kwilczynski@...nel.org>,
Kishon Vijay Abraham I <kishon@...nel.org>,
Bjorn Helgaas <bhelgaas@...gle.com>, Petr Mladek <pmladek@...e.com>,
Tejun Heo <tj@...nel.org>, Miri Korenblit
<miriam.rachel.korenblit@...el.com>, Frank Li <Frank.Li@....com>,
linux-pci@...r.kernel.org, linux-kernel@...r.kernel.org,
linux-kernel-mentees@...ts.linux.dev, skhan@...uxfoundation.org
Subject: Re: [PATCH] PCI: endpoint: pci-epf-test: Fix sleeping function being
called from atomic context
On 29/09/25 23:23, Manivannan Sadhasivam wrote:
> On Wed, Sep 17, 2025 at 09:48:17PM +0530, Bhanu Seshu Kumar Valluri wrote:
>> When Root Complex(RC) triggers a Doorbell MSI interrupt to Endpoint(EP) it triggers a warning
>> in the EP. pci_endpoint kselftest target is compiled and used to run the Doorbell test in RC.
>>
>> [ 474.686193] BUG: sleeping function called from invalid context at kernel/locking/mutex.c:271
>> [ 474.694656] in_atomic(): 1, irqs_disabled(): 1, non_block: 0, pid: 0, name: swapper/0
>> [ 474.702473] preempt_count: 10001, expected: 0
>> [ 474.706819] RCU nest depth: 0, expected: 0
>> [ 474.710913] CPU: 0 UID: 0 PID: 0 Comm: swapper/0 Not tainted 6.17.0-rc5-g7aac71907bde #12 PREEMPT
>> [ 474.710926] Hardware name: Texas Instruments AM642 EVM (DT)
>> [ 474.710934] Call trace:
>> [ 474.710940] show_stack+0x20/0x38 (C)
>> [ 474.710969] dump_stack_lvl+0x70/0x88
>> [ 474.710984] dump_stack+0x18/0x28
>> [ 474.710995] __might_resched+0x130/0x158
>> [ 474.711011] __might_sleep+0x70/0x88
>> [ 474.711023] mutex_lock+0x2c/0x80
>> [ 474.711036] pci_epc_get_msi+0x78/0xd8
>> [ 474.711052] pci_epf_test_raise_irq.isra.0+0x74/0x138
>> [ 474.711063] pci_epf_test_doorbell_handler+0x34/0x50
>> [ 474.711072] __handle_irq_event_percpu+0xac/0x1f0
>> [ 474.711086] handle_irq_event+0x54/0xb8
>> [ 474.711096] handle_fasteoi_irq+0x150/0x220
>> [ 474.711110] handle_irq_desc+0x48/0x68
>> [ 474.711121] generic_handle_domain_irq+0x24/0x38
>> [ 474.711131] gic_handle_irq+0x4c/0xc8
>> [ 474.711141] call_on_irq_stack+0x30/0x70
>> [ 474.711151] do_interrupt_handler+0x70/0x98
>> [ 474.711163] el1_interrupt+0x34/0x68
>> [ 474.711176] el1h_64_irq_handler+0x18/0x28
>> [ 474.711189] el1h_64_irq+0x6c/0x70
>> [ 474.711198] default_idle_call+0x10c/0x120 (P)
>> [ 474.711208] do_idle+0x128/0x268
>> [ 474.711220] cpu_startup_entry+0x3c/0x48
>> [ 474.711231] rest_init+0xe0/0xe8
>> [ 474.711240] start_kernel+0x6d4/0x760
>> [ 474.711255] __primary_switched+0x88/0x98
>>
>
> You do not need to use full call trace. Refer:
> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/process/submitting-patches.rst?h=v6.17#n761
>
>> Warnings can be reproduced by following steps below.
>> *On EP side:
>> 1. Configure the pci-epf-test function using steps given below
>> mount -t configfs none /sys/kernel/config
>> cd /sys/kernel/config/pci_ep/
>> mkdir functions/pci_epf_test/func1
>> echo 0x104c > functions/pci_epf_test/func1/vendorid
>> echo 0xb010 > functions/pci_epf_test/func1/deviceid
>> echo 32 > functions/pci_epf_test/func1/msi_interrupts
>> echo 2048 > functions/pci_epf_test/func1/msix_interrupts
>> ln -s functions/pci_epf_test/func1 controllers/f102000.pcie-ep/
>> echo 1 > controllers/f102000.pcie-ep/start
>>
>> *On RC side:
>> 1. Once EP side configuration is done do pci rescan.
>> echo 1 > /sys/bus/pci/rescan
>> 2. Run Doorbell MSI test using pci_endpoint_test kselftest app.
>> ./pci_endpoint_test -r pcie_ep_doorbell.DOORBELL_TEST
>
> This info is already part of the kernel documentation. So it is redundant here.
> It could be probably added in the comment section (where you added the Note).
>
>> Note: Kernel is compiled with CONFIG_DEBUG_KERNEL enabled.
>>
>> The BUG arises because the EP's Doorbell MSI hard interrupt handler is making an
>> indirect call to pci_epc_get_msi, which uses mutex inside, from interrupt context.
>>
>> This patch converts hard irq handler to a threaded irq handler to allow it
>> to call functions that can sleep during bottom half execution. The threaded
>> irq handler is registered with IRQF_ONESHOT and keeps interrupt line disabled
>> until the threaded irq handler completes execution.
>>
>> Fixes: eff0c286aa916221a69126 ("PCI: endpoint: pci-epf-test: Add doorbell test support")
>
> Use 12 char commit SHA.
>
>> -static irqreturn_t pci_epf_test_doorbell_handler(int irq, void *data)
>> +static irqreturn_t pci_epf_test_doorbell_irq_thread(int irq, void *data)
>
> No need to change the function name.
Thank you Mani for your helpful comments on the patch. I will send a v2 patch to address above review comments.
-Bhanu Seshu Kumar Valluri
Powered by blists - more mailing lists