Return-path: Received: from wolverine02.qualcomm.com ([199.106.114.251]:63813 "EHLO wolverine02.qualcomm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752424AbdESJCT (ORCPT ); Fri, 19 May 2017 05:02:19 -0400 From: Kalle Valo To: Sven Eckelmann CC: "ath10k@lists.infradead.org" , Adrian Chadd , "linux-wireless@vger.kernel.org" , "simon.wunderlich@openmesh.com" Subject: Re: [PATCH] [ath10k] go back to using dma_alloc_coherent() for firmware scratch memory. Date: Fri, 19 May 2017 09:02:08 +0000 Message-ID: <8760gxw776.fsf@kamboji.qca.qualcomm.com> (sfid-20170519_110242_113178_37685E15) References: <20170501214327.5621-1-adrian@freebsd.org> <2107639.cIofNFgAHq@bentobox> In-Reply-To: <2107639.cIofNFgAHq@bentobox> (Sven Eckelmann's message of "Wed, 17 May 2017 12:06:01 +0200") Content-Type: text/plain; charset="iso-8859-1" MIME-Version: 1.0 Sender: linux-wireless-owner@vger.kernel.org List-ID: Sven Eckelmann writes: > On Montag, 1. Mai 2017 14:43:27 CEST Adrian Chadd wrote: >> This reverts b057886524be060021e3cfad0ba8458c850330cd in 2015 >> which converted this allocation from dma_map_coherent() to >> kzalloc() / dma_map_single(). >>=20 >> The current problem manifests when using later model NICs with larger >> (>700KiB) scratch spaces in memory. Although the kzalloc call >> succeeds, the software IOMMU TLB code (via dma_map_single()) panics >> because it can't find 700KiB of linear physmem bounce buffers for DMA. >> Now, this is a bit of a silly failure mode for the dma map API, >> but it's what we currently have to play with. > [....] >>=20 >> Signed-off-by: Adrian Chadd >> --- >> drivers/net/wireless/ath/ath10k/wmi.c | 35 ++++++++++------------------= ------- >> 1 file changed, 10 insertions(+), 25 deletions(-) > > Thanks for investigating this. This partial revert fixes following crash = for > me with QCA99X0 on amd64 with SWIOMMU: > > [ 9.167281] DMA: Out of SW-IOMMU space for 689816 bytes at device = 0000:02:00.0 > [ 9.174719] Kernel panic - not syncing: DMA: Random memory could b= e DMA read I recently got a Compex WLE1216V5-20 board and this patch also fixed a similar crash for me on my HP x86_64 test laptop: [ 840.491019] do_IRQ: 3.82 No irq handler for vector [ 840.965049] ath10k_pci 0000:02:00.0: pci irq msi oper_irq_mode 2 irq_mod= e 0 reset_mode 0 [ 841.094985] ath10k_pci 0000:02:00.0: qca9984/qca9994 hw1.0 target 0x0100= 0000 chip_id 0x00000000 sub 168c:cafe [ 841.095096] ath10k_pci 0000:02:00.0: kconfig debug 1 debugfs 1 tracing 1= dfs 1 testmode 1 [ 841.095603] ath10k_pci 0000:02:00.0: firmware ver 10.4-3.4-00082 api 5 f= eatures no-p2p,mfp,peer-flow-ctrl,btcoex-param,allows-mesh-bcast crc32 f301= de65 [ 842.518071] ath10k_pci 0000:02:00.0: board_file api 2 bmi_id 0:1 crc32 7= 51efba1 [ 845.220186] ath10k_pci 0000:02:00.0: swiotlb buffer is full (sz: 1172264= bytes) [ 845.220478] ath10k_pci 0000:02:00.0: DMA: Out of SW-IOMMU space for 1172= 264 bytes [ 845.220506] Kernel panic - not syncing: DMA: Random memory could be DMA = accessed#012[ 845.220506]=20 [ 845.220532] CPU: 0 PID: 116 Comm: kworker/u16:4 Tainted: G E = 4.11.0-wt-ath+ #336 [ 845.220554] Hardware name: Hewlett-Packard HP ProBook 6540b/1722, BIOS 6= 8CDD Ver. F.04 01/27/2010 [ 845.220600] Workqueue: ath10k_aux_wq ath10k_wmi_event_service_ready_work= [ath10k_core] [ 845.220626] Call Trace: [ 845.220656] dump_stack+0x85/0xc2 [ 845.220684] panic+0xda/0x228 [ 845.220719] swiotlb_full+0x88/0xa0 [ 845.220746] swiotlb_map_page+0x214/0x280 [ 845.220785] ath10k_wmi_event_service_ready_work+0x3f2/0x7d0 [ath10k_cor= e] [ 845.220812] ? swiotlb_map_sg_attrs+0x140/0x140 [ 845.220858] process_one_work+0x23b/0x7d0 [ 845.221093] worker_thread+0x4e/0x4a0 [ 845.221126] kthread+0x117/0x150 [ 845.221149] ? process_one_work+0x7d0/0x7d0 [ 845.221175] ? kthread_create_on_node+0x40/0x40 [ 845.221205] ret_from_fork+0x31/0x40 [ 845.221314] Kernel Offset: 0xb000000 from 0xffffffff81000000 (relocation= range: 0xffffffff80000000-0xffffffffbfffffff) --=20 Kalle Valo=