On Wed May 24, 2023 at 6:40 PM EEST, A. Sverdlin wrote:
> From: Alexander Sverdlin <alexander.sverdlin@xxxxxxxxxxx>
>
> Underlying I2C bus drivers not always support longer transfers and
> imx-lpi2c for instance doesn't. The fix is symmetric to previous patch
> which fixed the read direction.
>
Cc: stable@xxxxxxxxxxxxxxx # v5.20+
> Fixes: bbc23a07b072 ("tpm: Add tpm_tis_i2c backend for tpm_tis_core")
> Tested-by: Michael Haener <michael.haener@xxxxxxxxxxx>
> Signed-off-by: Alexander Sverdlin <alexander.sverdlin@xxxxxxxxxxx>
> ---
> drivers/char/tpm/tpm_tis_i2c.c | 22 +++++++++++++++-------
> 1 file changed, 15 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/char/tpm/tpm_tis_i2c.c b/drivers/char/tpm/tpm_tis_i2c.c
> index 106fd20d94e4..82fda488e98b 100644
> --- a/drivers/char/tpm/tpm_tis_i2c.c
> +++ b/drivers/char/tpm/tpm_tis_i2c.c
> @@ -230,19 +230,27 @@ static int tpm_tis_i2c_write_bytes(struct tpm_tis_data *data, u32 addr, u16 len,
> struct i2c_msg msg = { .addr = phy->i2c_client->addr };
> u8 reg = tpm_tis_i2c_address_to_register(addr);
> int ret;
> + u16 wrote = 0;
>
> if (len > TPM_BUFSIZE - 1)
> return -EIO;
>
> - /* write register and data in one go */
> phy->io_buf[0] = reg;
> - memcpy(phy->io_buf + sizeof(reg), value, len);
> -
> - msg.len = sizeof(reg) + len;
> msg.buf = phy->io_buf;
> - ret = tpm_tis_i2c_retry_transfer_until_ack(data, &msg);
> - if (ret < 0)
> - return ret;
> + while (wrote < len) {
> + /* write register and data in one go */
> + msg.len = sizeof(reg) + len - wrote;
> + if (msg.len > I2C_SMBUS_BLOCK_MAX)
> + msg.len = I2C_SMBUS_BLOCK_MAX;
> +
> + memcpy(phy->io_buf + sizeof(reg), value + wrote,
> + msg.len - sizeof(reg));
> +
> + ret = tpm_tis_i2c_retry_transfer_until_ack(data, &msg);
> + if (ret < 0)
> + return ret;
> + wrote += msg.len - sizeof(reg);
> + }
>
> return 0;
> }
> --
> 2.40.1
Reviewed-by: Jarkko Sakkinen <jarkko@xxxxxxxxxx>
BR, Jarkko
