On Wed, 9 Sept 2026 at 18:14, Jan Kiszka <[email protected]> wrote:
>
> On 09.09.26 16:35, Jan Kiszka wrote:
> > On 09.09.26 13:46, Jan Kiszka wrote:
> >> On 09.09.26 12:04, Ilias Apalodimas wrote:
> >>> On Wed, 9 Sept 2026 at 12:12, Jan Kiszka <[email protected]> wrote:
> >>>>
> >>>> On 09.09.26 09:50, Ilias Apalodimas wrote:
> >>>>> Hi Jan
> >>>>>
> >>>>> [...]
> >>>>>
> >>>>>> --- a/drivers/mmc/mmc.c
> >>>>>> +++ b/drivers/mmc/mmc.c
> >>>>>> @@ -27,6 +27,7 @@
> >>>>>> #include <linux/list.h>
> >>>>>> #include <linux/printk.h>
> >>>>>> #include <div64.h>
> >>>>>> +#include <tee/optee.h>
> >>>>>> #include "mmc_private.h"
> >>>>>>
> >>>>>> #define DEFAULT_CMD6_TIMEOUT_MS 500
> >>>>>> @@ -3168,6 +3169,9 @@ int mmc_init(struct mmc *mmc)
> >>>>>> mmc->cfg->name);
> >>>>>> }
> >>>>>>
> >>>>>> + if (CONFIG_IS_ENABLED(OPTEE) && mmc->capacity_rpmb > 0)
> >>>>>> + optee_rpmb_available();
> >>>>>
> >>>>> With this we'll end up calling the optee bind methods again for
> >>>>> discovered devices calling bind_service_list(). I haven't tested this
> >>>>> locally yet, but is there any chance this ends up binding the devices
> >>>>> that depend on an RPMB twice?
> >>>>
> >>>> I don't think we will have an issue here: Either OP-TEE isn't ready yet,
> >>>> or RPMB wasn't yet when we called it first. So I do not see yet who
> >>>> actual successful listing could be done twice.
> >>>
> >>> The question is what happens if OP-TEE & the RPMB is ready and you
> >>> issue an mmc rescan. That will force optee_rpmb_available() to re-run
> >>> and rediscover all the devices no?
> >>
> >> Hmm, let me check...
> >>
> >
> > It's a problem. Is there a helper/service to check the binding state, or
> > should we track that via a static flag in optee_rpmb_available?
> >
>
> Didn't find anything, would go with something like this:
I am not sure if I am correct here but looking at the call path
optee_bind_services -> bind_service_list -> device_bind_driver_to_node
-> device_bind_with_driver_data.
The latter sets DM_FLAG_BOUND. I don't know the DM code too well
though to be 100% the flag is safe to use.
There's also DM_FLAG_ACTIVATED
Cheers
/Ilias
>
>
> diff --git a/drivers/tee/optee/rpmb.c b/drivers/tee/optee/rpmb.c
> index 7ae6045511b..bcad14a7251 100644
> --- a/drivers/tee/optee/rpmb.c
> +++ b/drivers/tee/optee/rpmb.c
> @@ -194,9 +194,14 @@ void optee_suppl_rpmb_release(struct udevice *dev)
>
> void optee_rpmb_available(void)
> {
> + static bool rpmb_services_bound;
> struct udevice *dev;
> struct uclass *uc;
> u32 tee_sess;
> + int ret;
> +
> + if (rpmb_services_bound)
> + return;
>
> uclass_id_foreach_dev(UCLASS_TEE, dev, uc) {
> if (strcmp(dev->driver->name, "optee") != 0 ||
> @@ -206,8 +211,13 @@ void optee_rpmb_available(void)
> if (optee_open_enum_session(dev, &tee_sess) != 0)
> continue;
>
> - optee_bind_services(dev, tee_sess, PTA_CMD_GET_DEVICES_RPMB);
> + ret = optee_bind_services(dev, tee_sess,
> PTA_CMD_GET_DEVICES_RPMB);
>
> tee_close_session(dev, tee_sess);
> +
> + if (!ret) {
> + rpmb_services_bound = true;
> + break;
> + }
> }
> }
>
> Jan
>
> --
> Siemens AG, Foundational Technologies
> Linux Expert Center
