[PATCH v3 2/5] remoteproc: imx_rproc: Add support for System Manager API

[Mathieu Poirier]

On Wed, Jul 09, 2025 at 05:02:31PM +0800, Peng Fan wrote:
> Hi Mathieu,
> 
> On Tue, Jul 08, 2025 at 10:32:34AM -0600, Mathieu Poirier wrote:
> >Good day,
> 
> Thanks, good day.
> 
> >
> >On Wed, Jun 25, 2025 at 10:23:28AM +0800, Peng Fan (OSS) wrote:
> >> From: Peng Fan <peng.fan at nxp.com>
> >> 
> >> i.MX95 features a Cortex-M33 core, six Cortex-A55 cores, and
> >> one Cortex-M7 core. The System Control Management Interface(SCMI)
> >> firmware runs on the M33 core. The i.MX95 SCMI firmware named System
> >> Manager(SM) includes vendor extension protocols, Logical Machine
> >> Management(LMM) protocol and CPU protocol and etc.
> >> 
> >> There are three cases for M7:
> >>  (1) M7 in a separate Logical Machine(LM) that Linux can't control it.
> >>  (2) M7 in a separate Logical Machine that Linux can control it using
> >>      LMM protocol
> >>  (3) M7 runs in same Logical Machine as A55, so Linux can control it
> >>      using CPU protocol
> >> 
> >> So extend the driver to using LMM and CPU protocol to manage the M7 core.
> >>  - Add IMX_RPROC_SM to indicate the remote core runs on a SoC that
> >>    has System Manager.
> >>  - Compare linux LM ID(got using scmi_imx_lmm_info) and M7 LM ID(the ID
> >>    is fixed as 1 in SM firmware if M7 is in a seprate LM),
> >>    if Linux LM ID equals M7 LM ID(linux and M7 in same LM), use CPU
> >>    protocol to start/stop. Otherwise, use LMM protocol to start/stop.
> >>    Whether using CPU or LMM protocol to start/stop, the M7 status
> >>    detection could use CPU protocol to detect started or not. So
> >>    in imx_rproc_detect_mode, use scmi_imx_cpu_started to check the
> >>    status of M7.
> >>  - For above case 1 and 2, Use SCMI_IMX_LMM_POWER_ON to detect whether
> >>    the M7 LM is under control of A55 LM.
> >>
> >
> >Thanks for the context, it really helps.
> 
> Glad that helps.
> 
> >
> 
> [....]
> 
> >> @@ -592,6 +637,38 @@ static int imx_rproc_prepare(struct rproc *rproc)
> >>  		rproc_add_carveout(rproc, mem);
> >>  	}
> >>  
> >> +	switch (dcfg->method) {
> >> +	case IMX_RPROC_SM:
> >> +		if (!(priv->flags & IMX_RPROC_FLAGS_SM_LMM_OP))
> >> +			break;
> >> +		/*
> >> +		 * Power on the Logical Machine to make sure TCM is available.
> >> +		 * Also serve as permission check. If in different Logical
> >> +		 * Machine, and linux has permission to handle the Logical
> >> +		 * Machine, set IMX_RPROC_FLAGS_SM_LMM_AVAIL.
> >> +		 */
> >> +		ret = scmi_imx_lmm_operation(dcfg->lmid, SCMI_IMX_LMM_POWER_ON, 0);
> >> +		if (ret == 0) {
> >> +			dev_info(priv->dev, "lmm(%d) powered on\n", dcfg->lmid);
> >> +			priv->flags |= IMX_RPROC_FLAGS_SM_LMM_AVAIL;
> >
> >This is set all the time imx_rproc_prepare() is called - isn't there a way to
> >set it once at initialisation time?
> 
> Yeah. Moving this to probe path should be ok, such as in imx_rproc_detect_mode.
> The drawback is M7 logical machine will be left in powered up state if
> moving to probe path before user starts M7 LM. Leaving in here means M7 logcal
> machine will be only powered up when user does
> "echo start > /xx/remoteproc-y/state", but needs to set
> IMX_RPROC_FLAGS_SM_LMM_AVAIL flags each time do preparing.
> 
> If you prefer moving this logic to probe, I could give a try to move
> to imx_rproc_detect_mode which is probe path.
> 
> How do you think?

Just leave it where it is.

Mathieu

> 
> >
> >> +		} else if (ret == -EACCES) {
> >> +			dev_info(priv->dev, "lmm(%d) not under Linux Control\n", dcfg->lmid);
> >> +			/*
> >> +			 * If remote cores boots up in detached mode, continue;
> >> +			 * else linux has no permission, return -EACCES.
> >> +			 */
> >> +			if (priv->rproc->state != RPROC_DETACHED)
> >> +				return -EACCES;
> >> +		} else if (ret) {
> >> +			dev_err(priv->dev, "Failed to power on lmm(%d): %d\n", ret, dcfg->lmid);
> >> +			return ret;
> >> +		}
> >> +
> >> +		break;
> >> +	default:
> >> +		break;
> >> +	};
> >> +
> >
> >Please put this in a function and get rid of the switch{}.
> 
> ok. Fix in v4
> 
> >
> >>  	return  0;
> >>  }
> >>  
> >> @@ -911,13 +988,41 @@ static int imx_rproc_detect_mode(struct imx_rproc *priv)
> >>  	struct regmap_config config = { .name = "imx-rproc" };
> >>  	const struct imx_rproc_dcfg *dcfg = priv->dcfg;
> >>  	struct device *dev = priv->dev;
> >> +	struct scmi_imx_lmm_info info;
> >>  	struct regmap *regmap;
> >>  	struct arm_smccc_res res;
> >> +	bool started = false;
> >>  	int ret;
> >>  	u32 val;
> >>  	u8 pt;
> >>  
> >>  	switch (dcfg->method) {
> >> +	case IMX_RPROC_SM:
> >> +		/* Get current Linux Logical Machine ID */
> >> +		ret = scmi_imx_lmm_info(LMM_ID_DISCOVER, &info);
> >> +		if (ret) {
> >> +			dev_err(dev, "Failed to get current LMM ID err: %d\n", ret);
> >> +			return ret;
> >> +		}
> >> +
> >> +		/*
> >> +		 * Check whether remote processor is in same Logical Machine as Linux.
> >> +		 * If no, need use Logical Machine API to manage remote processor, and
> >> +		 * set IMX_RPROC_FLAGS_SM_LMM_OP.
> >> +		 * If yes, use CPU protocol API to manage remote processor.
> >> +		 */
> >> +		if (dcfg->lmid != info.lmid) {
> >> +			priv->flags |= IMX_RPROC_FLAGS_SM_LMM_OP;
> >> +			dev_info(dev, "Using LMM Protocol OPS\n");
> >> +		} else {
> >> +			dev_info(dev, "Using CPU Protocol OPS\n");
> >> +		}
> >> +
> >> +		scmi_imx_cpu_started(dcfg->cpuid, &started);
> >
> >Function scmi_imx_cpu_started() returns a value that is not taken into account.
> >
> >> +		if (started)
> 
> I will update to "if (started || ret)" in v4, with ret assigned the return
> value of scmi_imx_cpu_started.
> 
> Thanks,
> Peng