From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from smtp.kernel.org (aws-us-west-2-korg-mail-1.web.codeaurora.org [10.30.226.201]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 2BE6910FC for ; Wed, 7 Jun 2023 20:25:25 +0000 (UTC) Received: by smtp.kernel.org (Postfix) with ESMTPSA id 99AA3C4339B; Wed, 7 Jun 2023 20:25:24 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=linuxfoundation.org; s=korg; t=1686169525; bh=qtzyEzDQ4ttmkO3L/rqrL/ykfQSxPIOGbQRfRH8HjBM=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=Dk2dZfcsWjkGQ1QtBWfsyXz6xRQHSW3T2bbhAtrvKkl/KGrR/d0nAwMxGpNoOoazd lcBhC17uc/EjnhS8Omh/5wVvoyRYZg+8RtFlsp6nn0vr1waqHKnWHlNgTgUfSHk/5a 2QECCR5R9dQTfCg4DlQb0qaz/GftWfNncBpMJxTs= From: Greg Kroah-Hartman To: stable@vger.kernel.org Cc: Greg Kroah-Hartman , patches@lists.linux.dev, Keith Busch , Adrian Huang , Christoph Hellwig , Sasha Levin , Jiwei Sun Subject: [PATCH 6.3 114/286] nvme-pci: clamp max_hw_sectors based on DMA optimized limitation Date: Wed, 7 Jun 2023 22:13:33 +0200 Message-ID: <20230607200926.805511051@linuxfoundation.org> X-Mailer: git-send-email 2.41.0 In-Reply-To: <20230607200922.978677727@linuxfoundation.org> References: <20230607200922.978677727@linuxfoundation.org> User-Agent: quilt/0.67 Precedence: bulk X-Mailing-List: patches@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit From: Adrian Huang [ Upstream commit 3710e2b056cb92ad816e4d79fa54a6a5b6ad8cbd ] When running the fio test on a 448-core AMD server + a NVME disk, a soft lockup or a hard lockup call trace is shown: [soft lockup] watchdog: BUG: soft lockup - CPU#126 stuck for 23s! [swapper/126:0] RIP: 0010:_raw_spin_unlock_irqrestore+0x21/0x50 ... Call Trace: fq_flush_timeout+0x7d/0xd0 ? __pfx_fq_flush_timeout+0x10/0x10 call_timer_fn+0x2e/0x150 run_timer_softirq+0x48a/0x560 ? __pfx_fq_flush_timeout+0x10/0x10 ? clockevents_program_event+0xaf/0x130 __do_softirq+0xf1/0x335 irq_exit_rcu+0x9f/0xd0 sysvec_apic_timer_interrupt+0xb4/0xd0 asm_sysvec_apic_timer_interrupt+0x1f/0x30 ... Obvisouly, fq_flush_timeout spends over 20 seconds. Here is ftrace log: | fq_flush_timeout() { | fq_ring_free() { | put_pages_list() { 0.170 us | free_unref_page_list(); 0.810 us | } | free_iova_fast() { | free_iova() { * 85622.66 us | _raw_spin_lock_irqsave(); 2.860 us | remove_iova(); 0.600 us | _raw_spin_unlock_irqrestore(); 0.470 us | lock_info_report(); 2.420 us | free_iova_mem.part.0(); * 85638.27 us | } * 85638.84 us | } | put_pages_list() { 0.230 us | free_unref_page_list(); 0.470 us | } ... ... $ 31017069 us | } Most of cores are under lock contention for acquiring iova_rbtree_lock due to the iova flush queue mechanism. [hard lockup] NMI watchdog: Watchdog detected hard LOCKUP on cpu 351 RIP: 0010:native_queued_spin_lock_slowpath+0x2d8/0x330 Call Trace: _raw_spin_lock_irqsave+0x4f/0x60 free_iova+0x27/0xd0 free_iova_fast+0x4d/0x1d0 fq_ring_free+0x9b/0x150 iommu_dma_free_iova+0xb4/0x2e0 __iommu_dma_unmap+0x10b/0x140 iommu_dma_unmap_sg+0x90/0x110 dma_unmap_sg_attrs+0x4a/0x50 nvme_unmap_data+0x5d/0x120 [nvme] nvme_pci_complete_batch+0x77/0xc0 [nvme] nvme_irq+0x2ee/0x350 [nvme] ? __pfx_nvme_pci_complete_batch+0x10/0x10 [nvme] __handle_irq_event_percpu+0x53/0x1a0 handle_irq_event_percpu+0x19/0x60 handle_irq_event+0x3d/0x60 handle_edge_irq+0xb3/0x210 __common_interrupt+0x7f/0x150 common_interrupt+0xc5/0xf0 asm_common_interrupt+0x2b/0x40 ... ftrace shows fq_ring_free spends over 10 seconds [1]. Again, most of cores are under lock contention for acquiring iova_rbtree_lock due to the iova flush queue mechanism. [Root Cause] The root cause is that the max_hw_sectors_kb of nvme disk (mdts=10) is 4096kb, which streaming DMA mappings cannot benefit from the scalable IOVA mechanism introduced by the commit 9257b4a206fc ("iommu/iova: introduce per-cpu caching to iova allocation") if the length is greater than 128kb. To fix the lock contention issue, clamp max_hw_sectors based on DMA optimized limitation in order to leverage scalable IOVA mechanism. Note: The issue does not happen with another NVME disk (mdts = 5 and max_hw_sectors_kb = 128) [1] https://gist.github.com/AdrianHuang/bf8ec7338204837631fbdaed25d19cc4 Suggested-by: Keith Busch Reported-and-tested-by: Jiwei Sun Signed-off-by: Adrian Huang Reviewed-by: Keith Busch Signed-off-by: Christoph Hellwig Signed-off-by: Sasha Levin --- drivers/nvme/host/pci.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/nvme/host/pci.c b/drivers/nvme/host/pci.c index a7772c0194d5a..a389f1ea0b151 100644 --- a/drivers/nvme/host/pci.c +++ b/drivers/nvme/host/pci.c @@ -2960,7 +2960,7 @@ static struct nvme_dev *nvme_pci_alloc_dev(struct pci_dev *pdev, * over a single page. */ dev->ctrl.max_hw_sectors = min_t(u32, - NVME_MAX_KB_SZ << 1, dma_max_mapping_size(&pdev->dev) >> 9); + NVME_MAX_KB_SZ << 1, dma_opt_mapping_size(&pdev->dev) >> 9); dev->ctrl.max_segments = NVME_MAX_SEGS; /* -- 2.39.2