linux-kernel - Re: [PATCH] tpm_tis: Fix undefined behavior in tpm_tis_spi_transfer

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]

Message-ID: <d24qnjxzidsbi6oidx7fk5i7txinsxj5rthnzx4n2ylbblejtx@26n4ecbdgyqc>
Date: Thu, 18 Sep 2025 10:55:30 +0200
From: Stefano Garzarella <sgarzare@...hat.com>
To: Gunnar Kudrjavets <gunnarku@...zon.com>
Cc: peterhuewe@....de, jarkko@...nel.org, jgg@...pe.ca, 
	kyarlagadda@...dia.com, linux-integrity@...r.kernel.org, linux-kernel@...r.kernel.org, 
	Justinien Bouron <jbouron@...zon.com>
Subject: Re: [PATCH] tpm_tis: Fix undefined behavior in
 tpm_tis_spi_transfer_half()

On Wed, Sep 17, 2025 at 03:29:56PM +0000, Gunnar Kudrjavets wrote:
>When len is 0, the while loop in tpm_tis_spi_transfer_half() never
>executes, leaving ret uninitialized. This will lead to undefined
>behavior when the function returns.
>
>The issue was introduced when tpm_tis_spi_transfer() was refactored
>to call tpm_tis_spi_transfer_half() or tpm_tis_spi_transfer_full().
>While ret is properly initialized in tpm_tis_spi_transfer_full(), it
>was missed in tpm_tis_spi_transfer_half().
>
>Initialize ret to 0 at the beginning of the function to ensure
>defined behavior in all code paths.
>
>Found by GCC 14.2.0 static analyzer with -fanalyzer.
>
>Fixes: a86a42ac2bd6 ("tpm_tis_spi: Add hardware wait polling")
>Signed-off-by: Gunnar Kudrjavets <gunnarku@...zon.com>
>Reviewed-by: Justinien Bouron <jbouron@...zon.com>
>---
> drivers/char/tpm/tpm_tis_spi_main.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
>diff --git a/drivers/char/tpm/tpm_tis_spi_main.c b/drivers/char/tpm/tpm_tis_spi_main.c
>index 61b42c83ced8..1b6d79662ca1 100644
>--- a/drivers/char/tpm/tpm_tis_spi_main.c
>+++ b/drivers/char/tpm/tpm_tis_spi_main.c
>@@ -85,7 +85,7 @@ static int tpm_tis_spi_transfer_half(struct tpm_tis_data *data,	u32 addr,
> 	struct spi_transfer spi_xfer[3];
> 	struct spi_message m;
> 	u8 transfer_len;
>-	int ret;
>+	int ret = 0;
>
> 	while (len) {
> 		transfer_len = min_t(u16, len, MAX_SPI_FRAMESIZE);
>

As an alternative, we can move the declaration in the while block where 
`ret` is always initialized, and we can return 0 at the end of the 
function since all the errors have an early return, I mean something 
like this:

diff --git a/drivers/char/tpm/tpm_tis_spi_main.c b/drivers/char/tpm/tpm_tis_spi_main.c
index 61b42c83ced8..c78c0a2d0541 100644
--- a/drivers/char/tpm/tpm_tis_spi_main.c
+++ b/drivers/char/tpm/tpm_tis_spi_main.c
@@ -85,9 +85,10 @@ static int tpm_tis_spi_transfer_half(struct tpm_tis_data *data,            u32 addr,
         struct spi_transfer spi_xfer[3];
         struct spi_message m;
         u8 transfer_len;
-       int ret;

         while (len) {
+               int ret;
+
                 transfer_len = min_t(u16, len, MAX_SPI_FRAMESIZE);

                 spi_message_init(&m);
@@ -135,7 +136,7 @@ static int tpm_tis_spi_transfer_half(struct tpm_tis_data *data,           u32 addr,
                 len -= transfer_len;
         }

-       return ret;
+       return 0;
  }

  static int tpm_tis_spi_transfer_full(struct tpm_tis_data *data, u32 addr,

BTW, I don't have a strong opinion on that, your patch also LGTM:

Reviewed-by: Stefano Garzarella <sgarzare@...hat.com>

Thanks,
Stefano