Netdev Archive on lore.kernel.org help / color / mirror / Atom feed
* Re: [genirq/msi] 495c66aca3: BUG:sleeping_function_called_from_invalid_context_at_kernel/locking/mutex.c [not found] <20211227150535.GA16252@xsang-OptiPlex-9020> @ 2021-12-28 18:40 ` Thomas Gleixner 2021-12-28 19:25 ` Heiner Kallweit 0 siblings, 1 reply; 3+ messages in thread From: Thomas Gleixner @ 2021-12-28 18:40 UTC (permalink / raw) To: kernel test robot Cc: Michael Kelley, Nishanth Menon, Jason Gunthorpe, LKML, x86, lkp, lkp, Heiner Kallweit, nic_swsd, David S. Miller, Jakub Kicinski, netdev On Mon, Dec 27 2021 at 23:05, kernel test robot wrote: > > FYI, we noticed the following commit (built with gcc-9): > > commit: 495c66aca3da704e063fa373fdbe371e71d3f4ee ("genirq/msi: Convert to new functions") > https://git.kernel.org/cgit/linux/kernel/git/tip/tip.git irq/msi > kern :err : [ 126.209306] BUG: sleeping function called from invalid context at kernel/locking/mutex.c:280 > kern :err : [ 126.209308] in_atomic(): 1, irqs_disabled(): 1, non_block: 0, pid: 5183, name: ls > kern :err : [ 126.209311] preempt_count: 2, expected: 0 > kern :warn : [ 126.209312] CPU: 2 PID: 5183 Comm: ls Not tainted 5.16.0-rc5-00091-g495c66aca3da #1 > kern :warn : [ 126.209315] Hardware name: Hewlett-Packard HP Pro 3340 MT/17A1, BIOS 8.07 01/24/2013 > kern :warn : [ 126.209316] Call Trace: > kern :warn : [ 126.209318] <TASK> > kern :warn : [ 126.209319] dump_stack_lvl (lib/dump_stack.c:107) > kern :warn : [ 126.209323] __might_resched.cold (kernel/sched/core.c:9539 kernel/sched/core.c:9492) > kern :warn : [ 126.209326] ? kasan_unpoison (mm/kasan/shadow.c:108 mm/kasan/shadow.c:142) > kern :warn : [ 126.209330] mutex_lock (kernel/locking/mutex.c:280) > kern :warn : [ 126.209335] ? __mutex_lock_slowpath (kernel/locking/mutex.c:279) > kern :warn : [ 126.209339] ? _raw_spin_lock_irqsave (arch/x86/include/asm/atomic.h:202 include/linux/atomic/atomic-instrumented.h:513 include/asm-generic/qspinlock.h:82 include/linux/spinlock.h:185 include/linux/spinlock_api_smp.h:111 kernel/locking/spinlock.c:162) > kern :warn : [ 126.209342] ? _raw_read_unlock_irqrestore (kernel/locking/spinlock.c:161) > kern :warn : [ 126.209344] msi_get_virq (kernel/irq/msi.c:332) > kern :warn : [ 126.209349] pci_irq_vector (drivers/pci/msi/msi.c:1085 drivers/pci/msi/msi.c:1077) > kern :warn : [ 126.209354] rtl8169_netpoll (drivers/net/ethernet/realtek/r8169_main.c:4722) > kern :warn : [ 126.209358] netpoll_poll_dev (net/core/netpoll.c:166 net/core/netpoll.c:195) > kern :warn : [ 126.209363] netpoll_send_skb (net/core/netpoll.c:350 net/core/netpoll.c:376) > kern :warn : [ 126.209367] write_msg (drivers/net/netconsole.c:862 drivers/net/netconsole.c:836) netconsole Fix below. Thanks, tglx --- drivers/net/ethernet/realtek/r8169_main.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) --- a/drivers/net/ethernet/realtek/r8169_main.c +++ b/drivers/net/ethernet/realtek/r8169_main.c @@ -615,6 +615,7 @@ struct rtl8169_private { struct ring_info tx_skb[NUM_TX_DESC]; /* Tx data buffers */ u16 cp_cmd; u32 irq_mask; + int irq; struct clk *clk; struct { @@ -4698,7 +4699,7 @@ static int rtl8169_close(struct net_devi cancel_work_sync(&tp->wk.work); - free_irq(pci_irq_vector(pdev, 0), tp); + free_irq(tp->irq, tp); phy_disconnect(tp->phydev); @@ -4719,7 +4720,7 @@ static void rtl8169_netpoll(struct net_d { struct rtl8169_private *tp = netdev_priv(dev); - rtl8169_interrupt(pci_irq_vector(tp->pci_dev, 0), tp); + rtl8169_interrupt(tp->irq, tp); } #endif @@ -4753,8 +4754,7 @@ static int rtl_open(struct net_device *d rtl_request_firmware(tp); irqflags = pci_dev_msi_enabled(pdev) ? IRQF_NO_THREAD : IRQF_SHARED; - retval = request_irq(pci_irq_vector(pdev, 0), rtl8169_interrupt, - irqflags, dev->name, tp); + retval = request_irq(tp->irq, rtl8169_interrupt, irqflags, dev->name, tp); if (retval < 0) goto err_release_fw_2; @@ -4771,7 +4771,7 @@ static int rtl_open(struct net_device *d return retval; err_free_irq: - free_irq(pci_irq_vector(pdev, 0), tp); + free_irq(tp->irq, tp); err_release_fw_2: rtl_release_firmware(tp); rtl8169_rx_clear(tp); @@ -5341,6 +5341,7 @@ static int rtl_init_one(struct pci_dev * dev_err(&pdev->dev, "Can't allocate interrupt\n"); return rc; } + tp->irq = pci_irq_vector(pdev, 0); INIT_WORK(&tp->wk.work, rtl_task); @@ -5416,8 +5417,7 @@ static int rtl_init_one(struct pci_dev * return rc; netdev_info(dev, "%s, %pM, XID %03x, IRQ %d\n", - rtl_chip_infos[chipset].name, dev->dev_addr, xid, - pci_irq_vector(pdev, 0)); + rtl_chip_infos[chipset].name, dev->dev_addr, xid, tp->irq); if (jumbo_max) netdev_info(dev, "jumbo features [frames: %d bytes, tx checksumming: %s]\n", ^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [genirq/msi] 495c66aca3: BUG:sleeping_function_called_from_invalid_context_at_kernel/locking/mutex.c 2021-12-28 18:40 ` [genirq/msi] 495c66aca3: BUG:sleeping_function_called_from_invalid_context_at_kernel/locking/mutex.c Thomas Gleixner @ 2021-12-28 19:25 ` Heiner Kallweit 2022-01-10 20:43 ` Thomas Gleixner 0 siblings, 1 reply; 3+ messages in thread From: Heiner Kallweit @ 2021-12-28 19:25 UTC (permalink / raw) To: Thomas Gleixner, kernel test robot Cc: Michael Kelley, Nishanth Menon, Jason Gunthorpe, LKML, x86, lkp, lkp, nic_swsd, David S. Miller, Jakub Kicinski, netdev On 28.12.2021 19:40, Thomas Gleixner wrote: > On Mon, Dec 27 2021 at 23:05, kernel test robot wrote: >> >> FYI, we noticed the following commit (built with gcc-9): >> >> commit: 495c66aca3da704e063fa373fdbe371e71d3f4ee ("genirq/msi: Convert to new functions") >> https://git.kernel.org/cgit/linux/kernel/git/tip/tip.git irq/msi >> kern :err : [ 126.209306] BUG: sleeping function called from invalid context at kernel/locking/mutex.c:280 >> kern :err : [ 126.209308] in_atomic(): 1, irqs_disabled(): 1, non_block: 0, pid: 5183, name: ls >> kern :err : [ 126.209311] preempt_count: 2, expected: 0 >> kern :warn : [ 126.209312] CPU: 2 PID: 5183 Comm: ls Not tainted 5.16.0-rc5-00091-g495c66aca3da #1 >> kern :warn : [ 126.209315] Hardware name: Hewlett-Packard HP Pro 3340 MT/17A1, BIOS 8.07 01/24/2013 >> kern :warn : [ 126.209316] Call Trace: >> kern :warn : [ 126.209318] <TASK> >> kern :warn : [ 126.209319] dump_stack_lvl (lib/dump_stack.c:107) >> kern :warn : [ 126.209323] __might_resched.cold (kernel/sched/core.c:9539 kernel/sched/core.c:9492) >> kern :warn : [ 126.209326] ? kasan_unpoison (mm/kasan/shadow.c:108 mm/kasan/shadow.c:142) >> kern :warn : [ 126.209330] mutex_lock (kernel/locking/mutex.c:280) >> kern :warn : [ 126.209335] ? __mutex_lock_slowpath (kernel/locking/mutex.c:279) >> kern :warn : [ 126.209339] ? _raw_spin_lock_irqsave (arch/x86/include/asm/atomic.h:202 include/linux/atomic/atomic-instrumented.h:513 include/asm-generic/qspinlock.h:82 include/linux/spinlock.h:185 include/linux/spinlock_api_smp.h:111 kernel/locking/spinlock.c:162) >> kern :warn : [ 126.209342] ? _raw_read_unlock_irqrestore (kernel/locking/spinlock.c:161) >> kern :warn : [ 126.209344] msi_get_virq (kernel/irq/msi.c:332) >> kern :warn : [ 126.209349] pci_irq_vector (drivers/pci/msi/msi.c:1085 drivers/pci/msi/msi.c:1077) >> kern :warn : [ 126.209354] rtl8169_netpoll (drivers/net/ethernet/realtek/r8169_main.c:4722) >> kern :warn : [ 126.209358] netpoll_poll_dev (net/core/netpoll.c:166 net/core/netpoll.c:195) >> kern :warn : [ 126.209363] netpoll_send_skb (net/core/netpoll.c:350 net/core/netpoll.c:376) >> kern :warn : [ 126.209367] write_msg (drivers/net/netconsole.c:862 drivers/net/netconsole.c:836) netconsole > > Fix below. > > Thanks, > > tglx > --- > drivers/net/ethernet/realtek/r8169_main.c | 14 +++++++------- > 1 file changed, 7 insertions(+), 7 deletions(-) > > --- a/drivers/net/ethernet/realtek/r8169_main.c > +++ b/drivers/net/ethernet/realtek/r8169_main.c > @@ -615,6 +615,7 @@ struct rtl8169_private { > struct ring_info tx_skb[NUM_TX_DESC]; /* Tx data buffers */ > u16 cp_cmd; > u32 irq_mask; > + int irq; > struct clk *clk; > > struct { > @@ -4698,7 +4699,7 @@ static int rtl8169_close(struct net_devi > > cancel_work_sync(&tp->wk.work); > > - free_irq(pci_irq_vector(pdev, 0), tp); > + free_irq(tp->irq, tp); > > phy_disconnect(tp->phydev); > > @@ -4719,7 +4720,7 @@ static void rtl8169_netpoll(struct net_d > { > struct rtl8169_private *tp = netdev_priv(dev); > > - rtl8169_interrupt(pci_irq_vector(tp->pci_dev, 0), tp); > + rtl8169_interrupt(tp->irq, tp); > } > #endif > > @@ -4753,8 +4754,7 @@ static int rtl_open(struct net_device *d > rtl_request_firmware(tp); > > irqflags = pci_dev_msi_enabled(pdev) ? IRQF_NO_THREAD : IRQF_SHARED; > - retval = request_irq(pci_irq_vector(pdev, 0), rtl8169_interrupt, > - irqflags, dev->name, tp); > + retval = request_irq(tp->irq, rtl8169_interrupt, irqflags, dev->name, tp); > if (retval < 0) > goto err_release_fw_2; > > @@ -4771,7 +4771,7 @@ static int rtl_open(struct net_device *d > return retval; > > err_free_irq: > - free_irq(pci_irq_vector(pdev, 0), tp); > + free_irq(tp->irq, tp); > err_release_fw_2: > rtl_release_firmware(tp); > rtl8169_rx_clear(tp); > @@ -5341,6 +5341,7 @@ static int rtl_init_one(struct pci_dev * > dev_err(&pdev->dev, "Can't allocate interrupt\n"); > return rc; > } > + tp->irq = pci_irq_vector(pdev, 0); > > INIT_WORK(&tp->wk.work, rtl_task); > > @@ -5416,8 +5417,7 @@ static int rtl_init_one(struct pci_dev * > return rc; > > netdev_info(dev, "%s, %pM, XID %03x, IRQ %d\n", > - rtl_chip_infos[chipset].name, dev->dev_addr, xid, > - pci_irq_vector(pdev, 0)); > + rtl_chip_infos[chipset].name, dev->dev_addr, xid, tp->irq); > > if (jumbo_max) > netdev_info(dev, "jumbo features [frames: %d bytes, tx checksumming: %s]\n", Thanks for the patch, I'll submit it with your SoB. Apart from pci_irq_vector() incl. underlying msi_get_virq(), are there more functions that must not be called from atomic context any longer? Maybe the new constraint should be added to kernel-doc of affected functions? ^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [genirq/msi] 495c66aca3: BUG:sleeping_function_called_from_invalid_context_at_kernel/locking/mutex.c 2021-12-28 19:25 ` Heiner Kallweit @ 2022-01-10 20:43 ` Thomas Gleixner 0 siblings, 0 replies; 3+ messages in thread From: Thomas Gleixner @ 2022-01-10 20:43 UTC (permalink / raw) To: Heiner Kallweit, kernel test robot Cc: Michael Kelley, Nishanth Menon, Jason Gunthorpe, LKML, x86, lkp, lkp, nic_swsd, David S. Miller, Jakub Kicinski, netdev On Tue, Dec 28 2021 at 20:25, Heiner Kallweit wrote: > On 28.12.2021 19:40, Thomas Gleixner wrote: >> if (jumbo_max) >> netdev_info(dev, "jumbo features [frames: %d bytes, tx checksumming: %s]\n", > > Thanks for the patch, I'll submit it with your SoB. Thank you! > Apart from pci_irq_vector() incl. underlying msi_get_virq(), are there > more functions that must not be called from atomic context any longer? I don't think so. The allocation/free functions were always restricted to preemptible context. > Maybe the new constraint should be added to kernel-doc of affected > functions? Probably. Thanks, tglx ^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2022-01-10 20:43 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
[not found] <20211227150535.GA16252@xsang-OptiPlex-9020>
2021-12-28 18:40 ` [genirq/msi] 495c66aca3: BUG:sleeping_function_called_from_invalid_context_at_kernel/locking/mutex.c Thomas Gleixner
2021-12-28 19:25 ` Heiner Kallweit
2022-01-10 20:43 ` Thomas Gleixner
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).