[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAG_fn=WeK8q2g0bRna+fFx+ks4HbfoG3Tnw8PpSdmfdH=3+S=A@mail.gmail.com>
Date: Fri, 27 Jun 2025 18:28:14 +0200
From: Alexander Potapenko <glider@...gle.com>
To: Leon Romanovsky <leon@...nel.org>
Cc: Marek Szyprowski <m.szyprowski@...sung.com>, Christoph Hellwig <hch@....de>,
Jonathan Corbet <corbet@....net>, Madhavan Srinivasan <maddy@...ux.ibm.com>,
Michael Ellerman <mpe@...erman.id.au>, Nicholas Piggin <npiggin@...il.com>,
Christophe Leroy <christophe.leroy@...roup.eu>, Robin Murphy <robin.murphy@....com>,
Joerg Roedel <joro@...tes.org>, Will Deacon <will@...nel.org>, "Michael S. Tsirkin" <mst@...hat.com>,
Jason Wang <jasowang@...hat.com>, Xuan Zhuo <xuanzhuo@...ux.alibaba.com>,
Eugenio Pérez <eperezma@...hat.com>,
Marco Elver <elver@...gle.com>, Dmitry Vyukov <dvyukov@...gle.com>,
Masami Hiramatsu <mhiramat@...nel.org>, Mathieu Desnoyers <mathieu.desnoyers@...icios.com>,
Jérôme Glisse <jglisse@...hat.com>,
Andrew Morton <akpm@...ux-foundation.org>, linux-doc@...r.kernel.org,
linux-kernel@...r.kernel.org, linuxppc-dev@...ts.ozlabs.org,
iommu@...ts.linux.dev, virtualization@...ts.linux.dev,
kasan-dev@...glegroups.com, linux-trace-kernel@...r.kernel.org,
linux-mm@...ck.org
Subject: Re: [PATCH 5/8] kmsan: convert kmsan_handle_dma to use physical addresses
On Thu, Jun 26, 2025 at 8:45 PM Leon Romanovsky <leon@...nel.org> wrote:
>
> On Thu, Jun 26, 2025 at 07:43:06PM +0200, Alexander Potapenko wrote:
> > On Wed, Jun 25, 2025 at 3:19 PM Leon Romanovsky <leon@...nel.org> wrote:
> > >
> > > From: Leon Romanovsky <leonro@...dia.com>
Acked-by: Alexander Potapenko <glider@...gle.com>
> >
> > Hi Leon,
> >
> > >
> > > Convert the KMSAN DMA handling function from page-based to physical
> > > address-based interface.
> > >
> > > The refactoring renames kmsan_handle_dma() parameters from accepting
> > > (struct page *page, size_t offset, size_t size) to (phys_addr_t phys,
> > > size_t size).
> >
> > Could you please elaborate a bit why this is needed? Are you fixing
> > some particular issue?
>
> It is soft of the fix and improvement at the same time.
> Improvement:
> It allows direct call to kmsan_handle_dma() without need
> to convert from phys_addr_t to struct page for newly introduced
> dma_map_phys() routine.
>
> Fix:
> It prevents us from executing kmsan for addresses that don't have struct page
> (for example PCI_P2PDMA_MAP_THRU_HOST_BRIDGE pages), which we are doing
> with original code.
>
> dma_map_sg_attrs()
> -> __dma_map_sg_attrs()
> -> dma_direct_map_sg()
> -> PCI_P2PDMA_MAP_THRU_HOST_BRIDGE and nents > 0
> -> kmsan_handle_dma_sg();
> -> kmsan_handle_dma(g_page(item) <---- this is "fake" page.
>
> We are trying to build DMA API that doesn't require struct pages.
Thanks for clarifying that!
> > KMSAN only works on 64-bit systems, do we actually have highmem on any of these?
>
> I don't know, but the original code had this check:
> 344 if (PageHighMem(page))
> 345 return;
>
> Thanks
Ouch, I overlooked that, sorry!
I spent a while trying to understand where this code originated from,
and found the following discussion:
https://lore.kernel.org/all/20200327170647.GA22758@lst.de/
It's still unclear to me whether we actually need this check, because
with my config it doesn't produce any code.
But I think this shouldn't be blocking your patch, I'd rather make a
follow-up fix.
Powered by blists - more mailing lists