Re: [bug report] platform/x86/amd/pmf: Add support for PMF-TA interaction

[Shyam Sundar S K <Shyam-sundar.S-k@amd.com>]

Hi Dan,

I was OOO. Apologies for the long delay in responding back on this
thread.

On 4/23/2026 12:39, Dan Carpenter wrote:
> [You don't often get email from error27@gmail.com. Learn why this is important at https://aka.ms/LearnAboutSenderIdentification ]
> 
> Hello Shyam Sundar S K,
> 
> Commit ae82cef7d9c5 ("platform/x86/amd/pmf: Add support for PMF-TA
> interaction") from Dec 12, 2023 (linux-next), leads to the following
> Smatch static checker warning:
> 
> drivers/platform/x86/amd/pmf/tee-if.c:261 amd_pmf_invoke_cmd_enact() warn: set error code if 'arg.ret != 0'
> drivers/platform/x86/amd/pmf/tee-if.c:307 amd_pmf_invoke_cmd_init() warn: set error code if 'arg.ret != 0'
> 
> drivers/platform/x86/amd/pmf/tee-if.c
>     234 int amd_pmf_invoke_cmd_enact(struct amd_pmf_dev *dev)
>     235 {
>     236         struct ta_pmf_shared_memory *ta_sm = NULL;
>     237         struct ta_pmf_enact_result *out = NULL;
>     238         struct ta_pmf_enact_table *in = NULL;
>     239         struct tee_param param[MAX_TEE_PARAM];
>     240         struct tee_ioctl_invoke_arg arg;
>     241         int ret = 0;
>     242
>     243         if (!dev->tee_ctx)
>     244                 return -ENODEV;
>     245
>     246         memset(dev->shbuf, 0, dev->policy_sz);
>     247         ta_sm = dev->shbuf;
>     248         out = &ta_sm->pmf_output.policy_apply_table;
>     249         in = &ta_sm->pmf_input.enact_table;
>     250
>     251         memset(ta_sm, 0, sizeof(*ta_sm));
>     252         ta_sm->command_id = TA_PMF_COMMAND_POLICY_BUILDER_ENACT_POLICIES;
>     253         ta_sm->if_version = PMF_TA_IF_VERSION_MAJOR;
>     254
>     255         amd_pmf_populate_ta_inputs(dev, in);
>     256         amd_pmf_prepare_args(dev, TA_PMF_COMMAND_POLICY_BUILDER_ENACT_POLICIES, &arg, param);
>     257
>     258         ret = tee_client_invoke_func(dev->tee_ctx, &arg, param);
>     259         if (ret < 0 || arg.ret != 0) {
>     260                 dev_err(dev->dev, "TEE enact cmd failed. err: %x, ret:%d\n", arg.ret, ret);
> --> 261                 return ret;
> 
> return ret ?: -EINVAL?  (not sure what kind of error codes are stored in
> arg.ret).

Yes.

return ret ?: -EINVAL is the right thing to do in this case.

But I quickly ran the smatch and could not catch the problem what you
is being reported. Can you let me know how to trigger this warning?

Thanks,
Shyam

> 
>     262         }
>     263
>     264         if (ta_sm->pmf_result == TA_PMF_TYPE_SUCCESS && out->actions_count) {
>     265                 amd_pmf_dump_ta_inputs(dev, in);
>     266                 dev_dbg(dev->dev, "action count:%u result:%x\n", out->actions_count,
>     267                         ta_sm->pmf_result);
>     268                 amd_pmf_apply_policies(dev, out);
>     269         }
>     270
>     271         return 0;
>     272 }
> 
> This email is a free service from the Smatch-CI project [smatch.sf.net].
> 
> regards,
> dan carpenter