From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 2EDB4C6FA90 for ; Thu, 22 Sep 2022 05:58:24 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S229535AbiIVF6W (ORCPT ); Thu, 22 Sep 2022 01:58:22 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:51910 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S229475AbiIVF6U (ORCPT ); Thu, 22 Sep 2022 01:58:20 -0400 Received: from verein.lst.de (verein.lst.de [213.95.11.211]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id B74DDB4428 for ; Wed, 21 Sep 2022 22:58:19 -0700 (PDT) Received: by verein.lst.de (Postfix, from userid 2407) id 593C568AFE; Thu, 22 Sep 2022 07:58:15 +0200 (CEST) Date: Thu, 22 Sep 2022 07:58:15 +0200 From: Christoph Hellwig To: Sagi Grimberg Cc: Rishabh Bhatnagar , linux-kernel@vger.kernel.org, linux-nvme@lists.infradead.org, hch@lst.de, axboe@fb.com, kbusch@kernel.org, mbacco@amazon.com Subject: Re: [PATCH] nvme-pci: Set min align mask before calculating max_hw_sectors Message-ID: <20220922055815.GA27619@lst.de> References: <20220920191932.22797-1-risbhat@amazon.com> <871c0e19-9272-e92b-28dc-a50ae705bba8@grimberg.me> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <871c0e19-9272-e92b-28dc-a50ae705bba8@grimberg.me> User-Agent: Mutt/1.5.17 (2007-11-01) Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Wed, Sep 21, 2022 at 11:10:58AM +0300, Sagi Grimberg wrote: > >> If swiotlb is force enabled dma_max_mapping_size ends up calling >> swiotlb_max_mapping_size which takes into account the min align >> mask for the device. >> Set the min align mask for nvme driver before calling >> dma_max_mapping_size while calculating max hw sectors. > > Does this fix a specific bug? if so it needs a fixes tag so it can > go to stable backports as well. Hmm. Basically this is another fixlet for nvme on swiotlb, which already drove adding the max_mapping size and the align_mask, and it seems we still hadn't covered all corner cases properly. So I think it basically has been broken since day 1, but nvme on swiotlb only started becoming "interesting" with the trusted hypervisor schemes.