> -----邮件原件----- > 发件人: Marek Vasut <[email protected]> > 发送时间: 2025年3月4日 5:05 > 收件人: Alice Guo (OSS) <[email protected]>; Tom Rini > <[email protected]>; Stefano Babic <[email protected]>; Fabio Estevam > <[email protected]>; dl-uboot-imx <[email protected]>; Lukasz > Majewski <[email protected]>; Sean Anderson <[email protected]>; Simon > Glass <[email protected]>; Alper Nebi Yasak <[email protected]>; > Alice Guo <[email protected]> > 抄送: [email protected]; Viorel Suman <[email protected]>; Ye Li > <[email protected]> > 主题: Re: 回复: [PATCH v6 02/20] firmware: scmi: smt: Interrupt > communication enable > > On 3/3/25 3:51 AM, Alice Guo (OSS) wrote: > > Hi, > > >>> @@ -48,6 +58,9 @@ int scmi_dt_get_smt_buffer(struct udevice *dev, > >>> struct > >> scmi_smt *smt) > >>> if (!smt->buf) > >>> return -ENOMEM; > >>> > >>> + if (IS_ENABLED(CONFIG_SCMI_TRANSPORT_SMT_INTR)) > >>> + scmi_smt_enable_intr(smt, true); > >> Would it be possible to check DT /firmware/scmi node for compatible > >> == "arm,scmi" and presence of "mboxes" property , and based on that , > >> determine that this is mailbox based SCMI ? > > > > Hi Marek, > > > > Do you mean that replace > IS_ENABLED(CONFIG_SCMI_TRANSPORT_SMT_INTR) by checking "arm,scmi" > and "mboxes" in device tree? > If that is possible, yes.
I do not think that Bit[0] of channel flags should be set to 1 only when the SCMI node has the compatible "arm,scmi" and the property "mboxes". Whether to set it depends on the implementation of the SCMI callee-side.
