From mboxrd@z Thu Jan 1 00:00:00 1970 From: Lucas Stach Subject: Re: [PATCH v4 0/7] add virt-dma support for imx-sdma Date: Thu, 14 Jun 2018 12:22:20 +0200 Message-ID: <1528971740.21021.36.camel@pengutronix.de> References: <1528984982-5074-1-git-send-email-yibin.gong@nxp.com> <1528966400.21021.26.camel@pengutronix.de> <1528999731.10947.20.camel@nxp.com> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 8bit Return-path: In-Reply-To: <1528999731.10947.20.camel@nxp.com> Sender: linux-kernel-owner@vger.kernel.org To: Robin Gong , "s.hauer@pengutronix.de" , "vkoul@kernel.org" , "dan.j.williams@intel.com" , "gregkh@linuxfoundation.org" , "jslaby@suse.com" Cc: "dmaengine@vger.kernel.org" , "linux-kernel@vger.kernel.org" , "linux-arm-kernel@lists.infradead.org" , dl-linux-imx , "linux-serial@vger.kernel.org" List-Id: linux-serial@vger.kernel.org Am Donnerstag, den 14.06.2018, 10:09 +0000 schrieb Robin Gong: > Hi Lucas, > Could you double check again? Still can reproduce lockdep > warning on UART if change > spin_lock_lockirqsave/spinlock_unlock_irqrestore to > spin_lock/spin_unlock in sdma_int_handler as you said without > patch7/7. Yes, I can reproduce the lockdep issue with _irqsave variants in the IRQ handler and 7/7 reverted. It fixes the pcm issue though. > Would you please ask below two more questions? > 1. Does your uart case pass now with my v4 patchset? It seems to work, but the change seems to impact the non-DMA serial somehow. On several boots the system was unable to present a login prompt, so patch 7/7 seems to break other stuff and it's a pretty invasive change to the serial driver, which is known to be non-trivial. I don't have the bandwidth to dig through this patch currently, but I wonder if it couldn't be simplified with the spinlock stuff in the sdma driver fixed. > 2. Do you make some code change in your local('I just gave this > series > a spin')to reproduce your below issue? If yes, could you post your > change? The lockdep splat below was without any changes to your series. Regards, Lucas > On 四, 2018-06-14 at 10:53 +0200, Lucas Stach wrote: > > Hi Robin, > > > > I just gave this series a spin and it seems there is even more > > locking > > fun, see the lockdep output below. After taking a short look it > > seems > > this is caused by using the wrong spinlock variants in > > sdma_int_handler(), those should also use the _irqsave ones. When > > fixing this you might want to reconsider patch 7/7, as it's > > probably > > not needed at all. > > > > Regards, > > Lucas > > > > [   20.479401] > > ======================================================== > > [   20.485769] WARNING: possible irq lock inversion dependency > > detected > > [   20.492140] 4.17.0+ #1538 Not tainted > > [   20.495814] ---------------------------------------------------- > > -- > > -- > > [   20.502181] systemd/1 just changed the state of lock: > > [   20.507247] c0abdafc (&(&substream->self_group.lock)->rlock){- > > ...}, at: snd_pcm_stream_lock+0x54/0x60 > > [   20.516523] but this lock took another, HARDIRQ-unsafe lock in > > the > > past: > > [   20.523234]  (fs_reclaim){+.+.} > > [   20.523253]  > > [   20.523253]  > > [   20.523253] and interrupts could create inverse lock ordering > > between them. > > [   20.523253]  > > [   20.537804]  > > [   20.537804] other info that might help us debug this: > > [   20.544344]  Possible interrupt unsafe locking scenario: > > [   20.544344]  > > [   20.551144]        CPU0                    CPU1 > > [   20.555686]        ----                    ---- > > [   20.560224]   lock(fs_reclaim); > > [   20.563386]                                local_irq_disable(); > > [   20.569315]                                lock(&(&substream- > > > self_group.lock)->rlock); > > > > [   20.577337]                                lock(fs_reclaim); > > [   20.583014]    > > [   20.585643]     lock(&(&substream->self_group.lock)->rlock); > > [   20.591322]  > > [   20.591322]  *** DEADLOCK *** > > [   20.591322]  > > [   20.597260] 1 lock held by systemd/1: > > [   20.607806]  #0: ab23d11c (snd_pcm_link_rwlock){.-..}, at: > > snd_pcm_stream_lock+0x4c/0x60 > > [   20.615951]  > > [   20.615951] the shortest dependencies between 2nd lock and 1st > > lock: > > [   20.623799]  -> (fs_reclaim){+.+.} ops: 248474 { > > [   20.628456]     HARDIRQ-ON-W at: > > [   20.631716]                       lock_acquire+0x260/0x29c > > [   20.637223]                       fs_reclaim_acquire+0x48/0x58 > > [   20.643075]                       kmem_cache_alloc_trace+0x3c/0x > > 36 > > 4 > > [   20.649366]                       alloc_worker.constprop.15+0x28 > > /0 > > x64 > > [   20.655824]                       init_rescuer.part.5+0x20/0xa4 > > [   20.661764]                       workqueue_init+0x124/0x1f8 > > [   20.667446]                       kernel_init_freeable+0x60/0x55 > > 0 > > [   20.673561]                       kernel_init+0x18/0x120 > > [   20.678890]                       ret_from_fork+0x14/0x20 > > [   20.684299]                         (null) > > [   20.688408]     SOFTIRQ-ON-W at: > > [   20.691659]                       lock_acquire+0x260/0x29c > > [   20.697158]                       fs_reclaim_acquire+0x48/0x58 > > [   20.703006]                       kmem_cache_alloc_trace+0x3c/0x > > 36 > > 4 > > [   20.709288]                       alloc_worker.constprop.15+0x28 > > /0 > > x64 > > [   20.709301]                       init_rescuer.part.5+0x20/0xa4 > > [   20.720717]                       workqueue_init+0x124/0x1f8 > > [   20.720729]                       kernel_init_freeable+0x60/0x55 > > 0 > > [   20.720738]                       kernel_init+0x18/0x120 > > [   20.720746]                       ret_from_fork+0x14/0x20 > > [   20.720751]                         (null) > > [   20.720756]     INITIAL USE at: > > [   20.720770]                      lock_acquire+0x260/0x29c > > [   20.720782]                      fs_reclaim_acquire+0x48/0x58 > > [   20.774374]                      kmem_cache_alloc_trace+0x3c/0x3 > > 64 > > [   20.780574]                      alloc_worker.constprop.15+0x28/ > > 0x > > 64 > > [   20.786945]                      init_rescuer.part.5+0x20/0xa4 > > [   20.792794]                      workqueue_init+0x124/0x1f8 > > [   20.798384]                      kernel_init_freeable+0x60/0x550 > > [   20.804406]                      kernel_init+0x18/0x120 > > [   20.809648]                      ret_from_fork+0x14/0x20 > > [   20.814971]                        (null) > > [   20.818992]   } > > [   20.820768]   ... key      at: [<80e22074>] > > __fs_reclaim_map+0x0/0x10 > > [   20.827220]   ... acquired at: > > [   20.830289]    fs_reclaim_acquire+0x48/0x58 > > [   20.834487]    kmem_cache_alloc_trace+0x3c/0x364 > > [   20.839123]    sdma_transfer_init+0x44/0x130 > > [   20.843409]    sdma_prep_dma_cyclic+0x78/0x21c > > [   20.847869]    snd_dmaengine_pcm_trigger+0xdc/0x184 > > [   20.852764]    soc_pcm_trigger+0x164/0x190 > > [   20.856876]    snd_pcm_do_start+0x34/0x40 > > [   20.860902]    snd_pcm_action_single+0x48/0x74 > > [   20.865360]    snd_pcm_action+0x34/0xfc > > [   20.869213]    snd_pcm_ioctl+0x910/0x10ec > > [   20.873241]    vfs_ioctl+0x30/0x44 > > [   20.876657]    do_vfs_ioctl+0xac/0x974 > > [   20.880421]    ksys_ioctl+0x48/0x64 > > [   20.883923]    sys_ioctl+0x18/0x1c > > [   20.887340]    ret_fast_syscall+0x0/0x28 > > [   20.891277]    0x7289bcbc > > [   20.893909]  > > [   20.895410] -> (&(&substream->self_group.lock)->rlock){-...} > > ops: > > 59 { > > [   20.901977]    IN-HARDIRQ-W at: > > [   20.905143]                     lock_acquire+0x260/0x29c > > [   20.910473]                     _raw_spin_lock+0x48/0x58 > > [   20.915801]                     snd_pcm_stream_lock+0x54/0x60 > > [   20.921561]                     _snd_pcm_stream_lock_irqsave+0x4 > > 0/ > > 0x48 > > [   20.928107]                     snd_pcm_period_elapsed+0x2c/0xa4 > > [   20.934127]                     dmaengine_pcm_dma_complete+0x54/ > > 0x > > 58 > > [   20.940498]                     sdma_int_handler+0x1dc/0x2a8 > > [   20.946179]                     __handle_irq_event_percpu+0x1fc/ > > 0x > > 498 > > [   20.952635]                     handle_irq_event_percpu+0x38/0x8 > > c > > [   20.958742]                     handle_irq_event+0x48/0x6c > > [   20.964242]                     handle_fasteoi_irq+0xc4/0x138 > > [   20.970006]                     generic_handle_irq+0x28/0x38 > > [   20.975681]                     __handle_domain_irq+0xb0/0xc4 > > [   20.981443]                     gic_handle_irq+0x68/0xa0 > > [   20.986769]                     __irq_svc+0x70/0xb0 > > [   20.991662]                     _raw_spin_unlock_irq+0x38/0x6c > > [   20.997511]                     task_work_run+0x90/0xb8 > > [   21.002751]                     do_work_pending+0xc8/0xd0 > > [   21.008164]                     slow_work_pending+0xc/0x20 > > [   21.013661]                     0x76c77e86 > > [   21.017768]    INITIAL USE at: > > [   21.020844]                    lock_acquire+0x260/0x29c > > [   21.026086]                    _raw_spin_lock+0x48/0x58 > > [   21.031328]                    snd_pcm_stream_lock+0x54/0x60 > > [   21.037002]                    snd_pcm_stream_lock_irq+0x38/0x3c > > [   21.043023]                    snd_pcm_sync_ptr+0x214/0x260 > > [   21.048609]                    snd_pcm_ioctl+0xbe0/0x10ec > > [   21.054027]                    vfs_ioctl+0x30/0x44 > > [   21.058832]                    do_vfs_ioctl+0xac/0x974 > > [   21.063984]                    ksys_ioctl+0x48/0x64 > > [   21.068875]                    sys_ioctl+0x18/0x1c > > [   21.073679]                    ret_fast_syscall+0x0/0x28 > > [   21.079004]                    0x7e9026dc > > [   21.083023]  } > > [   21.084710]  ... key      at: [<8162a6e4>] __key.31798+0x0/0x8 > > [   21.090552]  ... acquired at: > > [   21.093537]    mark_lock+0x3a4/0x69c > > [   21.097128]    __lock_acquire+0x420/0x16d4 > > [   21.101239]    lock_acquire+0x260/0x29c > > [   21.105091]    _raw_spin_lock+0x48/0x58 > > [   21.108940]    snd_pcm_stream_lock+0x54/0x60 > > [   21.113226]    _snd_pcm_stream_lock_irqsave+0x40/0x48 > > [   21.118296]    snd_pcm_period_elapsed+0x2c/0xa4 > > [   21.122841]    dmaengine_pcm_dma_complete+0x54/0x58 > > [   21.127735]    sdma_int_handler+0x1dc/0x2a8 > > [   21.131937]    __handle_irq_event_percpu+0x1fc/0x498 > > [   21.136915]    handle_irq_event_percpu+0x38/0x8c > > [   21.141547]    handle_irq_event+0x48/0x6c > > [   21.145570]    handle_fasteoi_irq+0xc4/0x138 > > [   21.149854]    generic_handle_irq+0x28/0x38 > > [   21.154052]    __handle_domain_irq+0xb0/0xc4 > > [   21.158335]    gic_handle_irq+0x68/0xa0 > > [   21.162184]    __irq_svc+0x70/0xb0 > > [   21.165601]    _raw_spin_unlock_irq+0x38/0x6c > > [   21.169973]    task_work_run+0x90/0xb8 > > [   21.173735]    do_work_pending+0xc8/0xd0 > > [   21.177670]    slow_work_pending+0xc/0x20 > > [   21.181691]    0x76c77e86 > > [   21.184320]  > > [   21.185821]  > > [   21.185821] stack backtrace: > > [   21.190198] CPU: 0 PID: 1 Comm: systemd Not tainted 4.17.0+ > > #1538 > > [   21.196303] Hardware name: Freescale i.MX6 Quad/DualLite (Device > > Tree) > > [   21.202841] Backtrace:  > > [   21.205314] [<8010e318>] (dump_backtrace) from [<8010e604>] > > (show_stack+0x20/0x24) > > [   21.212900]  r7:80e9f3d0 r6:00000000 r5:60070193 r4:80e9f3d0 > > [   21.218581] [<8010e5e4>] (show_stack) from [<8099b660>] > > (dump_stack+0xa4/0xd8) > > [   21.225825] [<8099b5bc>] (dump_stack) from [<8017b52c>] > > (print_irq_inversion_bug+0x15c/0x1fc) > > [   21.234368]  r9:814da818 r8:00000001 r7:ee926c00 r6:00000000 > > r5:ee915bb0 r4:814da818 > > [   21.242133] [<8017b3d0>] (print_irq_inversion_bug) from > > [<8017b6dc>] (check_usage_forwards+0x110/0x13c) > > [   21.251544]  r9:00000002 r8:80bfd3e2 r7:ee926c00 r6:ee927148 > > r5:80e08488 r4:00000001 > > [   21.259306] [<8017b5cc>] (check_usage_forwards) from > > [<8017c2a4>] > > (mark_lock+0x3a4/0x69c) > > [   21.267500]  r9:ee926c00 r8:80a03cd8 r7:00000101 r6:00000002 > > r5:00000000 r4:ee927148 > > [   21.275263] [<8017bf00>] (mark_lock) from [<8017cf68>] > > (__lock_acquire+0x420/0x16d4) > > [   21.283023]  r10:ee927148 r9:ed4620e4 r8:ee926c00 r7:00000000 > > r6:00000001 r5:00000001 > > [   21.290863]  r4:00000490 > > [   21.293416] [<8017cb48>] (__lock_acquire) from [<8017ed58>] > > (lock_acquire+0x260/0x29c) > > [   21.301350]  r10:00000001 r9:80e084a4 r8:00000000 r7:00000000 > > r6:00000000 r5:ed4620e4 > > [   21.309189]  r4:00000000 > > [   21.311746] [<8017eaf8>] (lock_acquire) from [<809b74f0>] > > (_raw_spin_lock+0x48/0x58) > > [   21.319506]  r10:ee0a4714 r9:ed457100 r8:ee0a46c8 r7:ee0a4714 > > r6:ee0a4010 r5:807847b0 > > [   21.327346]  r4:ed4620d4 > > [   21.329902] [<809b74a8>] (_raw_spin_lock) from [<807847b0>] > > (snd_pcm_stream_lock+0x54/0x60) > > [   21.338265]  r5:ed462000 r4:ed462000 > > [   21.341863] [<8078475c>] (snd_pcm_stream_lock) from [<80784838>] > > (_snd_pcm_stream_lock_irqsave+0x40/0x48) > > [   21.351440]  r5:ed462000 r4:60070193 > > [   21.355042] [<807847f8>] (_snd_pcm_stream_lock_irqsave) from > > [<8078b044>] (snd_pcm_period_elapsed+0x2c/0xa4) > > [   21.364881]  r5:ee3ef000 r4:ed462000 > > [   21.368478] [<8078b018>] (snd_pcm_period_elapsed) from > > [<8078d7b4>] (dmaengine_pcm_dma_complete+0x54/0x58) > > [   21.378148]  r7:ee0a4714 r6:ee0a4010 r5:00000007 r4:ee0a46bc > > [   21.383827] [<8078d760>] (dmaengine_pcm_dma_complete) from > > [<80504c0c>] (sdma_int_handler+0x1dc/0x2a8) > > [   21.393157] [<80504a30>] (sdma_int_handler) from [<8018cd28>] > > (__handle_irq_event_percpu+0x1fc/0x498) > > [   21.402393]  r10:00000000 r9:eeafd400 r8:80e084a4 r7:00000038 > > r6:00000038 r5:80ea3c12 > > [   21.410233]  r4:ee2b5d40 > > [   21.412787] [<8018cb2c>] (__handle_irq_event_percpu) from > > [<8018cffc>] (handle_irq_event_percpu+0x38/0x8c) > > [   21.422457]  r10:00000000 r9:ee914000 r8:ee81d400 r7:00000038 > > r6:eeafd400 r5:eeafd464 > > [   21.430296]  r4:80e08488 > > [   21.432852] [<8018cfc4>] (handle_irq_event_percpu) from > > [<8018d098>] (handle_irq_event+0x48/0x6c) > > [   21.441736]  r6:eeafd464 r5:eeafd464 r4:eeafd400 > > [   21.446374] [<8018d050>] (handle_irq_event) from [<8019146c>] > > (handle_fasteoi_irq+0xc4/0x138) > > [   21.454912]  r7:00000038 r6:eeafd464 r5:80e10a60 r4:eeafd400 > > [   21.460589] [<801913a8>] (handle_fasteoi_irq) from [<8018bd9c>] > > (generic_handle_irq+0x28/0x38) > > [   21.469214]  r7:00000038 r6:80d92ae4 r5:00000000 r4:00000000 > > [   21.474893] [<8018bd74>] (generic_handle_irq) from [<8018c48c>] > > (__handle_domain_irq+0xb0/0xc4) > > [   21.483611] [<8018c3dc>] (__handle_domain_irq) from [<80102330>] > > (gic_handle_irq+0x68/0xa0) > > [   21.491978]  r9:ee914000 r8:f4001100 r7:80e5c6bc r6:ee915f00 > > r5:80e08c3c r4:f4000100 > > [   21.499738] [<801022c8>] (gic_handle_irq) from [<801019f0>] > > (__irq_svc+0x70/0xb0) > > [   21.507233] Exception stack(0xee915f00 to 0xee915f48) > > [   21.512303] 5f00: 00000001 00000004 00000000 ee926c00 ee9270a8 > > ee926c00 ecc45e00 ee9270a8 > > [   21.520498] 5f20: 80ec23b0 ee914000 00000000 ee915f64 ee915f20 > > ee915f50 8017c7c0 809b78ac > > [   21.528687] 5f40: 20070013 ffffffff > > [   21.532193]  r9:ee914000 r8:80ec23b0 r7:ee915f34 r6:ffffffff > > r5:20070013 r4:809b78ac > > [   21.539959] [<809b7874>] (_raw_spin_unlock_irq) from > > [<8014e98c>] > > (task_work_run+0x90/0xb8) > > [   21.548321]  r5:ee926c00 r4:ecc45e00 > > [   21.551913] [<8014e8fc>] (task_work_run) from [<8010da3c>] > > (do_work_pending+0xc8/0xd0) > > [   21.559848]  r9:ee914000 r8:801011c4 r7:ee915fb0 r6:ffffe000 > > r5:00000004 r4:801011c4 > > [   21.567608] [<8010d974>] (do_work_pending) from [<80101034>] > > (slow_work_pending+0xc/0x20) > > [   21.575797] Exception stack(0xee915fb0 to 0xee915ff8) > > [   21.580864] 5fa0:                                     00000000 > > 020c34e8 46059f00 00000000 > > [   21.589059] 5fc0: 00000002 76f133a4 020df680 00000006 76e6e168 > > 00000035 7ef81778 00000035 > > [   21.597253] 5fe0: 00000006 7ef816a0 76c75d67 76c77e86 20070030 > > 00000039 > > [   21.603883]  r7:00000006 r6:020df680 r5:76f133a4 r4:00000002 > > > > Am Donnerstag, den 14.06.2018, 22:02 +0800 schrieb Robin Gong: > > > > > > The legacy sdma driver has below limitations or drawbacks: > > >   1. Hardcode the max BDs number as "PAGE_SIZE / sizeof(*)", and > > > alloc > > >      one page size for one channel regardless of only few BDs > > > needed > > >      most time. But in few cases, the max PAGE_SIZE maybe not > > > enough. > > >   2. One SDMA channel can't stop immediatley once channel > > > disabled > > > which > > >      means SDMA interrupt may come in after this channel > > > terminated.There > > >      are some patches for this corner case such as commit > > > "2746e2c389f9", > > >      but not cover non-cyclic. > > > > > > The common virt-dma overcomes the above limitations. It can alloc > > > bd > > > dynamically and free bd once this tx transfer done. No memory > > > wasted or > > > maximum limititation here, only depends on how many memory can be > > > requested > > > from kernel. For No.2, such issue can be workaround by checking > > > if > > > there > > > is available descript("sdmac->desc") now once the unwanted > > > interrupt > > > coming. At last the common virt-dma is easier for sdma driver > > > maintain. > > > > > > Change from v3: > > >   1. add two uart patches which impacted by this patchset. > > >   2. unlock 'vc.lock' before cyclic dma callback and lock again > > > after > > >      it because some driver such as uart will call > > > dmaengine_tx_status > > >      which will acquire 'vc.lock' again and dead lock comes out. > > >   3. remove 'Revert commit' stuff since that patch is not wrong > > > and > > >      combine two patch into one patch as Sascha's comment. > > > > > > Change from v2: > > >   1. include Sascha's patch to make the main patch easier to > > > review. > > >      Thanks Sacha. > > >   2. remove useless 'desc'/'chan' in struct sdma_channe. > > > > > > Change from v1: > > >   1. split v1 patch into 5 patches. > > >   2. remove some unnecessary condition check. > > >   3. remove unnecessary 'pending' list. > > > > > > Robin Gong (6): > > >   tty: serial: imx: correct dma cookie status > > >   dmaengine: imx-sdma: add virt-dma support > > >   dmaengine: imx-sdma: remove useless 'lock' and 'enabled' in > > > 'struct > > >     sdma_channel' > > >   dmaengine: imx-sdma: remove the maximum limitation for bd > > > numbers > > >   dmaengine: imx-sdma: add sdma_transfer_init to decrease code > > > overlap > > >   tty: serial: imx: split all dma setup operations out of > > > 'port.lock' > > >     protector > > > > > > Sascha Hauer (1): > > >   dmaengine: imx-sdma: factor out a struct sdma_desc from struct > > >     sdma_channel > > > > > >  drivers/dma/Kconfig      |   1 + > > >  drivers/dma/imx-sdma.c   | 394 +++++++++++++++++++++++++++------ > > > -------------- > > >  drivers/tty/serial/imx.c |  99 ++++++------ > > >  3 files changed, 282 insertions(+), 212 deletions(-)