Hi Ilpo,
On 2/27/2024 21:15, Ilpo Järvinen wrote:
Hi Shyam & Armin,IMO, this patch is not required, reason being:
Shyam, please take a look at the question below.
On Tue, 27 Feb 2024, Armin Wolf wrote:
The length of the policy buffer is not validated before accessing it,
which means that multiple out-of-bounds memory accesses can occur.
This is especially bad since userspace can load policy binaries over
debugfs.
- the debugfs patch gets created only when CONFIG_AMD_PMF_DEBUG is
enabled.
- Sideload of policy binaries is only supported with a valid signing
key. (think like this can be tested & verified within AMD environment)
- Also, in amd_pmf_get_pb_data() there are boundary conditions that
are being checked. Is that not sufficient enough?
Not sure if I get your question clearly. Can you elaborate a bit more+ if (dev->policy_sz < POLICY_COOKIE_LEN + sizeof(length))This starts to feel like adding a struct for the header(?) would be better
+ return -EINVAL;
+
cookie = *(u32 *)(dev->policy_buf + POLICY_COOKIE_OFFSET);
length = *(u32 *)(dev->policy_buf + POLICY_COOKIE_LEN);
course of action here as then one could compare against sizeof(*header)
and avoid all those casts (IMO, just access the header fields directly
w/o the local variables).
on the struct you are envisioning?
but IHMO, we actually don't need a struct - as all that we would need
is to make sure the signing cookie is part of the policy binary and
leave the rest of the error handling to ASP/TEE modules (we can rely
on the feedback from those modules).
Shyam, do you think a struct makes sense here? There's some header inYes, the policy binary on a whole has multiple sections within it and
this policy, right?
there are multiple headers (like signing, OEM header, etc).
But that might be not real interest to the PMF driver. The only thing
the driver has to make sure is that the policy binary going into ASP
(AMD Secure Processor) is with in the limits and has a valid signing
cookie. So this part is already taken care in the current code.
ah! it has mix of both int and u32 :-)
There are more thing to address here...
1) amd_pmf_start_policy_engine() function returns -EINVAL & res that is
TA_PMF_* which inconsistent in type of the return value
Armin, would you like to amend this in your current series? or else I
will submit a change for this in my next series.
Thanks,
Shyam
2) Once 1) is fixed, the caller shadowing the return code can be fixed as
well:
ret = amd_pmf_start_policy_engine(dev);
if (ret)
return -EINVAL;