Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
MIME-Version: 1.0
References: <20190611125904.1013-1-cai@lca.pw> <CAK8P3a1rK79aj38H0i9vnzeycv6YZ0iUhBFz4giAFc7COTnmWQ@mail.gmail.com>
In-Reply-To: <CAK8P3a1rK79aj38H0i9vnzeycv6YZ0iUhBFz4giAFc7COTnmWQ@mail.gmail.com>
From:   Bartosz Szczepanek <bsz@semihalf.com>
Date:   Thu, 13 Jun 2019 13:41:21 +0200
Message-ID: <CABLO=+kZnpBm8W9MmSkSf=18R5fLMFe65+_YWw1-of46B+B1dA@mail.gmail.com>
Subject: Re: [PATCH -next] efi/tpm: fix a compilation warning
To:     Arnd Bergmann <arnd@arndb.de>
Cc:     Qian Cai <cai@lca.pw>, Ard Biesheuvel <ard.biesheuvel@linaro.org>,
        Matthew Garrett <mjg59@google.com>,
        linux-efi <linux-efi@vger.kernel.org>,
        Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Thu, Jun 13, 2019 at 10:55 AM Arnd Bergmann <arnd@arndb.de> wrote:
>
> - efi.tpm_final_log is a physical address that gets passed into
>   memremap() to return a pointer
> - tpm2_calc_event_log_size() takes a pointer argument and
>   dereferences it.

Where does it? It's passed with some added offset to
__calc_tpm2_event_size, which does the remapping part. That's why
physical address is used here.

> My best guess is that we should pass the output of memremap()
> here rather than the input:
>
> --- a/drivers/firmware/efi/tpm.c
> +++ b/drivers/firmware/efi/tpm.c
> @@ -75,7 +75,7 @@ int __init efi_tpm_eventlog_init(void)
>                 goto out;
>         }
>
> -       tbl_size = tpm2_calc_event_log_size(efi.tpm_final_log
> +       tbl_size = tpm2_calc_event_log_size(final_tbl
>                                             + sizeof(final_tbl->version)
>                                             + sizeof(final_tbl->nr_events),
>                                             final_tbl->nr_events,
>
> No idea if that is actually what was intended here, but it makes
> the code look more plausible.

Passing final_tbl will lead to failure, as it will be remapped for
second time. Cast is needed here. Changing the type from void* to
unsigned long or phys_addr_t (and casting later) would also do the
job. I'm fine with both options.