On Fri, Jul 13, 2018 at 11:06 AM, Mimi Zohar <zohar@xxxxxxxxxxxxxxxxxx> wrote:
> Some systems are memory constrained but they need to load very large
> firmwares. The firmware subsystem allows drivers to request this
> firmware be loaded from the filesystem, but this requires that the
> entire firmware be loaded into kernel memory first before it's provided
> to the driver. This can lead to a situation where we map the firmware
> twice, once to load the firmware into kernel memory and once to copy the
> firmware into the final resting place.
>
> To resolve this problem, commit a098ecd2fa7d ("firmware: support loading
> into a pre-allocated buffer") introduced request_firmware_into_buf() API
> that allows drivers to request firmware be loaded directly into a
> pre-allocated buffer.
>
> Do devices using pre-allocated memory run the risk of the firmware being
> accessible to the device prior to the completion of IMA's signature
> verification any more than when using two buffers? (Refer to mailing list
> discussion[1]).
>
> Only on systems with an IOMMU can the access be prevented. As long as
> the signature verification completes prior to the DMA map is performed,
> the device can not access the buffer. This implies that the same buffer
> can not be re-used. Can we ensure the buffer has not been DMA mapped
> before using the pre-allocated buffer?
>
> [1] https://lkml.org/lkml/2018/7/10/56
>
> Signed-off-by: Mimi Zohar <zohar@xxxxxxxxxxxxxxxxxx>
> Cc: Luis R. Rodriguez <mcgrof@xxxxxxxx>
> Cc: Stephen Boyd <sboyd@xxxxxxxxxx>
> Cc: Bjorn Andersson <bjorn.andersson@xxxxxxxxxx>
> Cc: Ard Biesheuvel <ard.biesheuvel@xxxxxxxxxx>
I can't decide if it's worth adding the link (maybe using the
lkml.kernel.org url[1]) directly in the code or not.
Either way:
Reviewed-by: Kees Cook <keescook@xxxxxxxxxxxx>
-Kees
[1] https://lkml.kernel.org/r/CAKv+Gu-knHeBRGqo+2pb3X9cCjwovEykoXUf=DZyP7aJpoS60A@xxxxxxxxxxxxxx
--
Kees Cook
Pixel Security
