Hi,
> -----Original Message-----
> From: Clément Péron <peron.clem@xxxxxxxxx>
> Sent: Thursday, October 18, 2018 12:32 PM
> To: Ulf Hansson <ulf.hansson@xxxxxxxxxx>; Avri Altman
> <Avri.Altman@xxxxxxx>; linux-mmc@xxxxxxxxxxxxxxx
> Cc: Nikita Maslov <wkernelteam@xxxxxxxxx>; Clément Péron
> <peron.clem@xxxxxxxxx>
> Subject: [PATCH v2] mmc-utils: RPMB fails with status 0x0001 on system eMMC
> chips
>
> From: Nikita Maslov <wkernelteam@xxxxxxxxx>
>
> Use MMC_IOC_MULTI_CMD for RPMB access
>
> On some systems which use MMC as a main storage device
> it is possible that RPMB commands are mixed with
> generic MMC access commands which invalidates RPMB.
> This patch uses MMC_IOC_MULTI_CMD.
I was hoping that this patch would be part of a series that rectifies
some more of mmc-utils rpmb deficiencies: Allowing multiple frame writes,
and adding Authenticated Device Configuration {Read/Write}.
But if you don't have time to attend all 3, fixing just this one is fine too.
>
> Signed-off-by: Clément Péron <peron.clem@xxxxxxxxx>
> ---
>
> Hi,
>
> this patch was firstly proposed by Nikita, as some modification was requested
> and Nikita didn't answer, I propose this new version.
>
> I didn't add the Nikita signature has it wasn't signed on the first patch and
> I don't think I'm allowed to do that.
>
> I still have some trouble to access the RPMB partition with controller that
> have the auto-cmd12 feature.
>
> So consider this patch as untested, I will be able to perform test on other
> controller by the end of the next week.
I can help you test the write/write key operations if you don't have a key
To your device
>
> Thanks,
> Clement
>
> mmc.h | 10 +++++++
> mmc_cmds.c | 84 ++++++++++++++++++++++++++----------------------------
> 2 files changed, 50 insertions(+), 44 deletions(-)
>
> diff --git a/mmc.h b/mmc.h
> index 285c1f1..e5d92a0 100644
> --- a/mmc.h
> +++ b/mmc.h
> @@ -194,3 +194,13 @@
>
> #define MMC_RSP_R1
> (MMC_RSP_PRESENT|MMC_RSP_CRC|MMC_RSP_OPCODE)
> #define MMC_RSP_R1B
> (MMC_RSP_PRESENT|MMC_RSP_CRC|MMC_RSP_OPCODE|MMC_RS
> P_BUSY)
> +
> +static inline void set_single_cmd(struct mmc_ioc_cmd *ioc, __u32 opcode,
> + int write_flag)
> +{
> + ioc->opcode = opcode;
> + ioc->write_flag = write_flag;
> + ioc->arg = 0x0;
> + ioc->blksz = 512;
> + ioc->flags = MMC_RSP_SPI_R1 | MMC_RSP_R1 | MMC_CMD_ADTC;
> +}
Any particular reason why it is not part of mmc_cmds.c ?
> diff --git a/mmc_cmds.c b/mmc_cmds.c
> index 44623fe..5544e4e 100644
> --- a/mmc_cmds.c
> +++ b/mmc_cmds.c
> @@ -1825,6 +1825,8 @@ int do_sanitize(int nargs, char **argv)
> ret;
> \
> })
>
> +#define RMPB_MULTI_CMD_MAX_CMDS 3
> +
> enum rpmb_op_type {
> MMC_RPMB_WRITE_KEY = 0x01,
> MMC_RPMB_READ_CNT = 0x02,
> @@ -1864,19 +1866,20 @@ static int do_rpmb_op(int fd,
> int err;
> u_int16_t rpmb_type;
>
> - struct mmc_ioc_cmd ioc = {
> - .arg = 0x0,
> - .blksz = 512,
> - .blocks = 1,
> - .write_flag = 1,
> - .opcode = MMC_WRITE_MULTIPLE_BLOCK,
> - .flags = MMC_RSP_SPI_R1 | MMC_RSP_R1 |
> MMC_CMD_ADTC,
> - .data_ptr = (uintptr_t)frame_in
> - };
> + struct mmc_ioc_multi_cmd *mioc;
If <linux/mmc/ioctl.h> does not includes MMC_IOC_MULTI_CMD,
Which is the case sometimes - this will not compile.
> + struct mmc_ioc_cmd *ioc;
>
> if (!frame_in || !frame_out || !out_cnt)
> return -EINVAL;
>
> + /* prepare arguments for MMC_IOC_MUTLI_CMD ioctl */
> + mioc = (struct mmc_ioc_multi_cmd *)
> + malloc(sizeof (struct mmc_ioc_multi_cmd) +
> + RMPB_MULTI_CMD_MAX_CMDS * sizeof (struct
> mmc_ioc_cmd));
> + if (!mioc) {
> + return -ENOMEM;
> + }
> +
> rpmb_type = be16toh(frame_in->req_resp);
>
> switch(rpmb_type) {
> @@ -1887,33 +1890,27 @@ static int do_rpmb_op(int fd,
> goto out;
> }
>
> + mioc->num_of_cmds = 3;
> +
> /* Write request */
> - ioc.write_flag |= (1<<31);
> - err = ioctl(fd, MMC_IOC_CMD, &ioc);
> - if (err < 0) {
> - err = -errno;
> - goto out;
> - }
> + ioc = &mioc->cmds[0];
> + set_single_cmd(ioc, MMC_WRITE_MULTIPLE_BLOCK, (1 << 31)
> | 1);
> + ioc->blocks = 1;
Can this be part of set_single_cmd?
> + mmc_ioc_cmd_set_data((*ioc), frame_in);
>
> /* Result request */
> + ioc = &mioc->cmds[1];
> memset(frame_out, 0, sizeof(*frame_out));
> frame_out->req_resp = htobe16(MMC_RPMB_READ_RESP);
> - ioc.write_flag = 1;
> - ioc.data_ptr = (uintptr_t)frame_out;
> - err = ioctl(fd, MMC_IOC_CMD, &ioc);
> - if (err < 0) {
> - err = -errno;
> - goto out;
> - }
> + set_single_cmd(ioc, MMC_WRITE_MULTIPLE_BLOCK, 1);
> + ioc->blocks = 1;
> + mmc_ioc_cmd_set_data((*ioc), frame_out);
>
> /* Get response */
> - ioc.write_flag = 0;
> - ioc.opcode = MMC_READ_MULTIPLE_BLOCK;
> - err = ioctl(fd, MMC_IOC_CMD, &ioc);
> - if (err < 0) {
> - err = -errno;
> - goto out;
> - }
> + ioc = &mioc->cmds[2];
> + set_single_cmd(ioc, MMC_READ_MULTIPLE_BLOCK, 0);
> + ioc->blocks = 1;
> + mmc_ioc_cmd_set_data((*ioc), frame_out);
You are using the same frame for both the result request and the response -
Expects troubles...
>
> break;
> case MMC_RPMB_READ_CNT:
> @@ -1924,23 +1921,19 @@ static int do_rpmb_op(int fd,
> /* fall through */
>
> case MMC_RPMB_READ:
> - /* Request */
> - err = ioctl(fd, MMC_IOC_CMD, &ioc);
> - if (err < 0) {
> - err = -errno;
> - goto out;
> - }
> + mioc->num_of_cmds = 2;
> +
> + /* Read request */
> + ioc = &mioc->cmds[0];
> + set_single_cmd(ioc, MMC_WRITE_MULTIPLE_BLOCK, 1);
> + ioc->blocks = 1;
> + mmc_ioc_cmd_set_data((*ioc), frame_in);
>
> /* Get response */
> - ioc.write_flag = 0;
> - ioc.opcode = MMC_READ_MULTIPLE_BLOCK;
> - ioc.blocks = out_cnt;
> - ioc.data_ptr = (uintptr_t)frame_out;
> - err = ioctl(fd, MMC_IOC_CMD, &ioc);
> - if (err < 0) {
> - err = -errno;
> - goto out;
> - }
> + ioc = &mioc->cmds[1];
> + set_single_cmd(ioc, MMC_READ_MULTIPLE_BLOCK, 0);
> + ioc->blocks = out_cnt;
> + mmc_ioc_cmd_set_data((*ioc), frame_out);
>
> break;
> default:
> @@ -1948,7 +1941,10 @@ static int do_rpmb_op(int fd,
> goto out;
> }
>
> + err = ioctl(fd, MMC_IOC_MULTI_CMD, mioc);
> +
> out:
> + free(mioc);
> return err;
> }
Thanks,
Avri
>
> --
> 2.17.1
