[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [PATCH v4 15/16] block-dma: properly take MMIO path
- To: Jason Gunthorpe <jgg@xxxxxxxxxx>
- From: Keith Busch <kbusch@xxxxxxxxxx>
- Date: Thu, 28 Aug 2025 13:10:32 -0600
- Cc: Leon Romanovsky <leon@xxxxxxxxxx>, Marek Szyprowski <m.szyprowski@xxxxxxxxxxx>, Abdiel Janulgue <abdiel.janulgue@xxxxxxxxx>, Alexander Potapenko <glider@xxxxxxxxxx>, Alex Gaynor <alex.gaynor@xxxxxxxxx>, Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>, Christoph Hellwig <hch@xxxxxx>, Danilo Krummrich <dakr@xxxxxxxxxx>, iommu@xxxxxxxxxxxxxxx, Jason Wang <jasowang@xxxxxxxxxx>, Jens Axboe <axboe@xxxxxxxxx>, Joerg Roedel <joro@xxxxxxxxxx>, Jonathan Corbet <corbet@xxxxxxx>, Juergen Gross <jgross@xxxxxxxx>, kasan-dev@xxxxxxxxxxxxxxxx, linux-block@xxxxxxxxxxxxxxx, linux-doc@xxxxxxxxxxxxxxx, linux-kernel@xxxxxxxxxxxxxxx, linux-mm@xxxxxxxxx, linux-nvme@xxxxxxxxxxxxxxxxxxx, linuxppc-dev@xxxxxxxxxxxxxxxx, linux-trace-kernel@xxxxxxxxxxxxxxx, Madhavan Srinivasan <maddy@xxxxxxxxxxxxx>, Masami Hiramatsu <mhiramat@xxxxxxxxxx>, Michael Ellerman <mpe@xxxxxxxxxxxxxx>, "Michael S. Tsirkin" <mst@xxxxxxxxxx>, Miguel Ojeda <ojeda@xxxxxxxxxx>, Robin Murphy <robin.murphy@xxxxxxx>, rust-for-linux@xxxxxxxxxxxxxxx, Sagi Grimberg <sagi@xxxxxxxxxxx>, Stefano Stabellini <sstabellini@xxxxxxxxxx>, Steven Rostedt <rostedt@xxxxxxxxxxx>, virtualization@xxxxxxxxxxxxxxx, Will Deacon <will@xxxxxxxxxx>, xen-devel@xxxxxxxxxxxxxxxxxxxx
- Delivery-date: Thu, 28 Aug 2025 19:10:44 +0000
- List-id: Xen developer discussion <xen-devel.lists.xenproject.org>
On Thu, Aug 28, 2025 at 03:41:15PM -0300, Jason Gunthorpe wrote:
> On Thu, Aug 28, 2025 at 11:15:20AM -0600, Keith Busch wrote:
> >
> > I don't think that was ever the case. Metadata is allocated
> > independently of the data payload, usually by the kernel in
> > bio_integrity_prep() just before dispatching the request. The bio may
> > have a p2p data payload, but the integrity metadata is just a kmalloc
> > buf in that path.
>
> Then you should do two dma mapping operations today, that is how the
> API was built. You shouldn't mix P2P and non P2P within a single
> operation right now..
Data and metadata are mapped as separate operations. They're just
different parts of one blk-mq request.
|
