From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Message-ID: <56A6CE90.4040701@cn.fujitsu.com> Date: Tue, 26 Jan 2016 09:40:32 +0800 From: Chen Fan MIME-Version: 1.0 To: Bjorn Helgaas CC: , , , , , , , , , , , , Thomas Gleixner Subject: Re: [PATCH v2] pci: fix unavailable irq number 255 reported by BIOS References: <1453705178-27389-1-git-send-email-chen.fan.fnst@cn.fujitsu.com> <20160125205803.GA10272@localhost> In-Reply-To: <20160125205803.GA10272@localhost> Content-Type: text/plain; charset="windows-1252"; format=flowed Sender: linux-acpi-owner@vger.kernel.org List-ID: On 01/26/2016 04:58 AM, Bjorn Helgaas wrote: > [+cc Thomas] > > On Mon, Jan 25, 2016 at 02:59:38PM +0800, Chen Fan wrote: >> In our environment, when enable Secure boot, we found an abnormal >> phenomenon as following call trace shows. after investigation, we >> found the firmware assigned an irq number 255 which means unknown >> or no connection in PCI local spec for i801_smbus, meanwhile the >> ACPI didn't configure the pci irq routing. and the 255 irq number >> was assigned for megasa msix without IRQF_SHARED. then in this case >> during i801_smbus probe, the i801_smbus driver would request irq with >> bad irq number 255. but the 255 irq number was assigned for memgasa >> with MSIX enable. which will cause request_irq fails as call trace >> shows, here we use ~0U as invalid IRQ to identify the 0xff IRQ specified >> by BIOS. >> >> See the call trace: > Maybe you missed my suggestion that the timestamps aren't useful; > here's my suggestion again in more detail: > > Changelogs are written once, but read dozens or hundreds of time, so > stripping out irrelevant details shows consideration for the readers. Got it, thanks for your correction, I will remake it as you suggest. > >> [ 32.459195] ipmi device interface >> [ 32.612907] shpchp: Standard Hot Plug PCI Controller Driver version: 0.4 >> [ 32.800459] ixgbe: Intel(R) 10 Gigabit PCI Express Network Driver - version 4.0.1-k-rh >> [ 32.818319] ixgbe: Copyright (c) 1999-2014 Intel Corporation. >> [ 32.844009] lpc_ich 0001:80:1f.0: I/O space for ACPI uninitialized > I think the lines above are completely irrelevant. > >> [ 32.850093] i801_smbus 0000:00:1f.3: enabling device (0140 -> 0143) >> [ 32.851134] i801_smbus 0000:00:1f.3: can't derive routing for PCI INT C >> [ 32.851136] i801_smbus 0000:00:1f.3: PCI INT C: no GSI >> [ 32.851164] genirq: Flags mismatch irq 255. 00000080 (i801_smbus) vs. 00000000 (megasa > These are useful, but the timestamps ("[ 32.850093]") are not. > >> [ 32.851168] CPU: 0 PID: 2487 Comm: kworker/0:1 Not tainted 3.10.0-229.el7.x86_64 #1 >> [ 32.851170] Hardware name: FUJITSU PRIMEQUEST 2800E2/D3736, BIOS PRIMEQUEST 2000 Serie5 > These are probably useful; it's nice to know what kernel and hardware > is involved. > >> [ 32.851178] Workqueue: events work_for_cpu_fn >> [ 32.851208] ffff88086c330b00 00000000e233a9df ffff88086d57bca0 ffffffff81603f36 >> [ 32.851227] ffff88086d57bcf8 ffffffff8110d23a ffff88686fe02000 0000000000000246 >> [ 32.851246] ffff88086a9a8c00 00000000e233a9df ffffffffa00ad220 0000000000000080 > I doubt these are useful. > >> [ 32.851247] Call Trace: >> [ 32.851261] [] dump_stack+0x19/0x1b >> [ 32.851271] [] __setup_irq+0x54a/0x570 >> [ 32.851282] [] ? i801_check_pre.isra.5+0xe0/0xe0 [i2c_i801] >> [ 32.851289] [] request_threaded_irq+0xcc/0x170 >> [ 32.851298] [] i801_probe+0x32f/0x508 [i2c_i801] >> [ 32.851308] [] local_pci_probe+0x45/0xa0 > The above might be useful, but the addresses ("[]") > are not, and you should go through them manually and strip out the > lines that are junk from the stack. For example, I don't think > request_threaded_irq() really calls i801_check_pre(). > >> [ 32.851315] [] work_for_cpu_fn+0x14/0x20 >> [ 32.851323] [] process_one_work+0x17b/0x470 >> [ 32.851330] [] worker_thread+0x293/0x400 >> [ 32.851338] [] ? rescuer_thread+0x400/0x400 >> [ 32.851346] [] kthread+0xcf/0xe0 >> [ 32.851353] [] ? kthread_create_on_node+0x140/0x140 >> [ 32.851362] [] ret_from_fork+0x7c/0xb0 >> [ 32.851369] [] ? kthread_create_on_node+0x140/0x140 > The lines above are completely useless. > >> [ 32.851373] i801_smbus 0000:00:1f.3: Failed to allocate irq 255: -16 >> [ 32.851435] i801_smbus: probe of 0000:00:1f.3 failed with error -16 >> [ 33.180145] ixgbe 0000:5a:00.0: Multiq[ 33.240538] ixgbe 0000:5a:00.0: (PCI Express:03:e0 >> [ 33.280826] ixgbe 0000:5a:00.0: MAC: 3, PHY: 0, PBA No: 000000-000 > These ixgbe entries are irrelevant. > >> Signed-off-by: Chen Fan >> --- >> arch/x86/include/asm/irq_vectors.h | 2 ++ >> drivers/acpi/pci_irq.c | 11 ++++++++++- >> include/linux/interrupt.h | 9 +++++++++ >> 3 files changed, 21 insertions(+), 1 deletion(-) >> >> diff --git a/arch/x86/include/asm/irq_vectors.h b/arch/x86/include/asm/irq_vectors.h >> index 6ca9fd6..b616d69 100644 >> --- a/arch/x86/include/asm/irq_vectors.h >> +++ b/arch/x86/include/asm/irq_vectors.h >> @@ -146,4 +146,6 @@ >> #define NR_IRQS NR_IRQS_LEGACY >> #endif >> >> +#define IRQ_INVALID (~0U) > If this is a good idea (I cc'd Thomas, the IRQ maintainer, for his > thoughts), I'd like to see this in a more generic place so it isn't > x86-specific. > >> #endif /* _ASM_X86_IRQ_VECTORS_H */ >> diff --git a/drivers/acpi/pci_irq.c b/drivers/acpi/pci_irq.c >> index d30184c..819eb23 100644 >> --- a/drivers/acpi/pci_irq.c >> +++ b/drivers/acpi/pci_irq.c >> @@ -33,6 +33,7 @@ >> #include >> #include >> #include >> +#include >> >> #define PREFIX "ACPI: " >> >> @@ -436,7 +437,15 @@ int acpi_pci_irq_enable(struct pci_dev *dev) >> * driver reported one, then use it. Exit in any case. >> */ >> if (gsi < 0) { >> - if (acpi_isa_register_gsi(dev)) >> +#ifdef CONFIG_X86 >> + /* >> + * The Interrupt Line value of 0xff is defined to mean "unknown" >> + * or "no connection" (PCI 3.0, Section 6.2.4, footnote on page >> + * 223), using ~0U as invalid IRQ. >> + */ >> + dev->irq = (dev->irq == 0xff) ? IRQ_INVALID : dev->irq; > It's much simpler and clearer to write: > > if (dev->irq == 0xff) > dev->irq = IRQ_INVALID; > >> +#endif >> + if (!irq_is_valid(dev->irq) || acpi_isa_register_gsi(dev)) >> dev_warn(&dev->dev, "PCI INT %c: no GSI\n", >> pin_name(pin)); >> >> diff --git a/include/linux/interrupt.h b/include/linux/interrupt.h >> index cb30edb..2f0d46b 100644 >> --- a/include/linux/interrupt.h >> +++ b/include/linux/interrupt.h >> @@ -198,6 +198,15 @@ extern void enable_percpu_irq(unsigned int irq, unsigned int type); >> extern bool irq_percpu_is_enabled(unsigned int irq); >> extern void irq_wake_thread(unsigned int irq, void *dev_id); >> >> +static inline bool irq_is_valid(unsigned int irq) >> +{ >> +#ifdef CONFIG_X86 >> + if (irq == IRQ_INVALID) >> + return false; >> +#endif >> + return true; >> +} > I don't like the x86 ifdef. I'd prefer: > > static inline bool irq_valid(unsigned int irq) > { > if (irq < NR_IRQS) > return true; > return false; > } > > This could be used in many of the places that currently use NR_IRQS. > > My suggestion: > > - patch 1: Add IRQ_INVALID and irq_valid() as generic things > - patch 2: Use irq_valid() in all the places where it can obviously > replace NR_IRQS > - patch 3: Add the acpi_pci_irq_enable() check. This is now a > trivial patch, basically just this: > > + #ifdef CONFIG_X86 > + if (dev->irq == 0xff) > + dev->irq = IRQ_INVALID; > + #endif > + if (!irq_valid(dev->irq) ... this will be more useful. Thanks, Chen > > Bjorn > > > . >