linux-kernel - Re: [PATCH] DMA-API: Change dma_declare_coherent_memory() CPU address to phys_addr

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]

Date:	Thu, 1 May 2014 01:48:17 -0600
From:	Bjorn Helgaas <bhelgaas@...gle.com>
To:	Greg Kroah-Hartman <gregkh@...uxfoundation.org>
Cc:	"linux-pci@...r.kernel.org" <linux-pci@...r.kernel.org>,
	Randy Dunlap <rdunlap@...radead.org>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	"James E.J. Bottomley" <JBottomley@...allels.com>,
	"linux-doc@...r.kernel.org" <linux-doc@...r.kernel.org>
Subject: Re: [PATCH] DMA-API: Change dma_declare_coherent_memory() CPU address
 to phys_addr_t

On Wed, Apr 30, 2014 at 6:07 PM, Greg Kroah-Hartman
<gregkh@...uxfoundation.org> wrote:
> On Wed, Apr 30, 2014 at 02:33:21PM -0600, Bjorn Helgaas wrote:
>> dma_declare_coherent_memory() takes two addresses for a region of memory: a
>> "bus_addr" and a "device_addr".  I think the intent is that "bus_addr" is
>> the physical address a *CPU* would use to access the region, and
>> "device_addr" is the bus address the *device* would use to address the
>> region.
>>
>> Rename "bus_addr" to "phys_addr" and change its type to phys_addr_t.
>
> Isn't this going to cause problems with the callers of this function?
> You aren't changing them...

Sorry, I should have mentioned that in the changelog.

This could cause a problem if it changed "bus_addr" to a smaller type,
e.g., in a config where dma_addr_t was 64 bits but phys_addr_t was
only 32.  In that config (if it exists), this change would truncate
the address if a caller  supplied a 64-bit dma_addr_t.  But most of
the callers in the tree already supply a phys_addr_t (or the
equivalent resource_size_t), and the rest supply 32-bit int values, so
I don't think there's a problem here.

This change could theoretically *fix* a problem in a config with
32-bit dma_addr_t and 64-bit phys_addr_t.  In that case we could
previously have truncated a phys_addr_t, and this change would fix it.
 But I don't know of any issues like this.

Here are the 16 callers in the tree:

4 in arch/arm/mach-imx/mach-imx27_visstrim_m10.c (phys_addr_t)
3 in arch/arm/mach-imx/mach-mx31_3ds.c (phys_addr_t)
1 in arch/arm/mach-imx/mach-pcm037.c (phys_addr_t)
1 in arch/arm/plat-samsung/s5p-dev-mfc.c (phys_addr_t)
1 in arch/sh/drivers/pci/fixups-dreamcast.c (#define constant)
2 in drivers/media/platform/s5p-mfc/s5p_mfc.c (unsigned int from OF property)
1 in drivers/media/platform/soc_camera/sh_mobile_ceu_camera.c (struct
resource.start)
1 in drivers/scsi/NCR_Q720.c (__u32 memory address computed from I/O
port values)
1 in drivers/usb/host/ohci-sm501.c (struct resource.start)
1 in drivers/usb/host/ohci-tmio.c (struct resource.start)

Bjorn
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/