linux-kernel - Re: [PATCH] dma-mapping: treat dev->bus_dma

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <33ba915ee84839286c69d048b15758a911c02844.camel@suse.de>
Date:   Thu, 14 Nov 2019 10:47:00 +0100
From:   Nicolas Saenz Julienne <nsaenzjulienne@...e.de>
To:     Robin Murphy <robin.murphy@....com>,
        Dave Hansen <dave.hansen@...ux.intel.com>,
        Andy Lutomirski <luto@...nel.org>,
        Peter Zijlstra <peterz@...radead.org>,
        Bjorn Helgaas <bhelgaas@...gle.com>,
        Lorenzo Pieralisi <lorenzo.pieralisi@....com>,
        Hanjun Guo <guohanjun@...wei.com>,
        Sudeep Holla <sudeep.holla@....com>,
        Jens Axboe <axboe@...nel.dk>, Joerg Roedel <joro@...tes.org>,
        Rob Herring <robh+dt@...nel.org>,
        Frank Rowand <frowand.list@...il.com>,
        Christoph Hellwig <hch@....de>,
        Marek Szyprowski <m.szyprowski@...sung.com>
Cc:     Benjamin Herrenschmidt <benh@...nel.crashing.org>,
        linux-pci@...r.kernel.org, linux-kernel@...r.kernel.org,
        linux-ide@...r.kernel.org, Paul Mackerras <paulus@...ba.org>,
        "H. Peter Anvin" <hpa@...or.com>,
        Paul Burton <paulburton@...nel.org>,
        Michael Ellerman <mpe@...erman.id.au>, x86@...nel.org,
        phil@...pberrypi.org, linux-acpi@...r.kernel.org,
        Ingo Molnar <mingo@...hat.com>,
        James Hogan <jhogan@...nel.org>, Len Brown <lenb@...nel.org>,
        devicetree@...r.kernel.org, Borislav Petkov <bp@...en8.de>,
        Thomas Gleixner <tglx@...utronix.de>,
        linux-arm-kernel@...ts.infradead.org,
        "Rafael J. Wysocki" <rjw@...ysocki.net>,
        linux-mips@...r.kernel.org, Ralf Baechle <ralf@...ux-mips.org>,
        iommu@...ts.linux-foundation.org, linuxppc-dev@...ts.ozlabs.org
Subject: Re: [PATCH] dma-mapping: treat dev->bus_dma_mask as a DMA limit

On Wed, 2019-11-13 at 20:34 +0000, Robin Murphy wrote:
> On 13/11/2019 4:13 pm, Nicolas Saenz Julienne wrote:
> > Using a mask to represent bus DMA constraints has a set of limitations.
> > The biggest one being it can only hold a power of two (minus one). The
> > DMA mapping code is already aware of this and treats dev->bus_dma_mask
> > as a limit. This quirk is already used by some architectures although
> > still rare.
> > 
> > With the introduction of the Raspberry Pi 4 we've found a new contender
> > for the use of bus DMA limits, as its PCIe bus can only address the
> > lower 3GB of memory (of a total of 4GB). This is impossible to represent
> > with a mask. To make things worse the device-tree code rounds non power
> > of two bus DMA limits to the next power of two, which is unacceptable in
> > this case.
> > 
> > In the light of this, rename dev->bus_dma_mask to dev->bus_dma_limit all
> > over the tree and treat it as such. Note that dev->bus_dma_limit is
> > meant to contain the higher accesible DMA address.
> 
> Neat, you win a "why didn't I do it that way in the first place?" :)

:)

> Looking at it without all the history of previous attempts, this looks 
> entirely reasonable, and definitely a step in the right direction.
> 
> [...]
> > diff --git a/drivers/acpi/arm64/iort.c b/drivers/acpi/arm64/iort.c
> > index 5a7551d060f2..f18827cf96df 100644
> > --- a/drivers/acpi/arm64/iort.c
> > +++ b/drivers/acpi/arm64/iort.c
> > @@ -1097,7 +1097,7 @@ void iort_dma_setup(struct device *dev, u64 *dma_addr,
> > u64 *dma_size)
> >   		 * Limit coherent and dma mask based on size
> >   		 * retrieved from firmware.
> >   		 */
> > -		dev->bus_dma_mask = mask;
> > +		dev->bus_dma_limit = mask;
> 
> Although this preserves the existing behaviour, as in of_dma_configure() 
> we can do better here since we have the original address range to hand. 
> I think it's worth keeping the ACPI and OF paths in sync for minor 
> tweaks like this, rather than letting them diverge unnecessarily.

I figure you mean something like this:

@@ -1085,19 +1085,15 @@ void iort_dma_setup(struct device *dev, u64 *dma_addr,
u64 *dma_size)
        }

        if (!ret) {
-               msb = fls64(dmaaddr + size - 1);
-               /*
-                * Round-up to the power-of-two mask or set
-                * the mask to the whole 64-bit address space
-                * in case the DMA region covers the full
-                * memory window.
-                */
-               mask = msb == 64 ? U64_MAX : (1ULL << msb) - 1;
+               /* Round-up to the power-of-two */
+               end = dmaddr + size - 1;
+               mask = DMA_BIT_MASK(ilog2(end) + 1);
+
                /*
                 * Limit coherent and dma mask based on size
                 * retrieved from firmware.
                 */
-               dev->bus_dma_limit = mask;
+               dev->bus_dma_limit = end;
                dev->coherent_dma_mask = mask;
                *dev->dma_mask = mask;
        }

> Otherwise, the rest looks OK to me - in principle we could store it as 
> an exclusive limit such that we could then streamline the min_not_zero() 
> tests to just min(mask, limit - 1), but that's probably too clever for 
> its own good.

Yes, that was my first intuition and in a perfect world I'd prefer it like
that. But as you say, it's probably going to cause more trouble than anything.

Regards,
Nicolas


Download attachment "signature.asc" of type "application/pgp-signature" (489 bytes)