iface_ctx.off check and modem reset high/low time in HL7800 driver

[pj-sunny]

Hi,

For the iface_ctx.off check introduced in the commit here, the error message always shows for me: modem_boot_handler: Err waiting for boot: -11.
It turns out that although the mdm_vgpio_callback_isr() is called within 1ms, it does nothing because the iface_ctx.off is still true.
And this is because the iface_ctx.off will not be set to false until the 10ms MDM_RESET_HIGH_TIME in modem_run() passed.

This seems to be an issue to me because the delay between RESET_IN_N and VGPIO is maximum 1ms according to the HL7800 product technical specification (Table 4-10).
In my opinion, the iface_ctx.off should be set to false right after the gpio_pin_set_dt() instead of after the 10ms MDM_RESET_HIGH_TIME because the driver has requested the modem to run by setting the reset pin high.

But not sure why no one else raises similar question.
Maybe this is just my question.

Also the specification defines the minimum reset low time 100us and has no definition about the reset high time.
Actually I do not understand why there is 50ms MDM_RESET_LOW_TIME with comment /* >20 milliseconds required for reset low */ and 10ms MDM_RESET_HIGH_TIME.
Of course they are necessary for implementation but seem to be a little bit longer than needed.

Maybe there are some reasons for these implementations and hope someone can help to share some insights.
Thank you.

[zafersn]

Hi @pj-sunny

For the iface_ctx.off check introduced in the commit here, the error message always shows for me: modem_boot_handler: Err waiting for boot: -11.

do you have the pm_device enabled?

I had the same issue modem_boot_handler: Err waiting for boot: -11 when I have CONFIG_PM_DEVICE=y enabled

the issue was that forgotten uart enabling.

see shutdown_uart() function in prepare_io_for_reset in modem_reset function. here they shut down the uart but after that.
I think during the modem_run they should've enabled it too, But i need to add by myself and worked.

P.S.
I think there are more conflicts in this driver file. between PM_ENABLE and LOW_POWER_MODE

[pj-sunny]

Hi @zafersn ,

Thank you for the response.
Yes, I do have the CONFIG_PM_DEVICE enabled.
But in my opinion, the CONFIG_PM_DEVICE is not the root cause.

Also the flow to enable the uart, i.e., call the power_on_uart(), is in mdm_vgpio_callback_isr(), which is not executed because the iface_ctx.off is still true.
Before the iface_ctx.off check introduced, actually everything worked correctly on my side.

[zafersn]

@pj-sunny I agree with what you stated there, If anyone wants a quick start, then they basically need to disable CONFIG_PM_DEVICE and then can continue. Just a heads up. Otherwise, they need to consider all the changes you explained in the question very well.

[pj-sunny]

It seems to me that the issue is resolved by PR 66237.
@rerickson1 , would you please help to share some insights about the implementations?
Thank you.

[rerickson1]

here

Hello, yes, that PR should have solve all issues regarding this.