• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Subscribers 99 subscribers
  • Views 756 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

AM5728: HDQ interface problems

Jonathan Fisher90
Prodigy 50 points

Part Number: AM5728

Hello,

We have a AM57xx-based design and are having trouble with hot-plugging 1-wire devices. We're using the 4.4 Linux kernel from the processor-sdk-linux-3.02.00 branch, with the omap_hdq.c driver for the 1-wire bus master interface.

The issues we're seeing are as follows:

  1. The kernel driver doesn't appear to respond to userspace requests in a timely way (e.g. "cat /sys/bus/w1/devices/w1_bus_master1/w1_master_attempts" can take ~15-30 seconds to actually respond).
  2. The interface only recognizes new devices on boot.
  3. The interface takes on the order of minutes to detect a device removal event
  4. The interface appears to lock up when hot-plugging a new device into the bus.

I enabled DEBUG in the kernel driver and looked at some of the logs, and I noticed that the PRESENCEDETECT bit of the HDQ_CTRL_STATUS register appears to immediately detect the device being removed, but that the Linux 1-wire interface takes several minutes to actually reflect that the device has been removed.

However, when plugging a new device into the bus, the bus appears to lock up--I stop getting HDQ interrupts, and all of the bus scanning calls time out (not via interrupt).

Any help with this would be appreciated!

  • 0
    TI__Guru 65405 points
    Hi Jonathan,

    Could you check whether the issue appears with the last Processor SDK v04.03 with Linux kernel 4.9.69.
    If the issue still appears check device tree weather matching to description at:
    www.kernel.org/.../omap-hdq.txt
    Also please share your kernel configuration settings related to HDQ/1-wire interface.

    BR
    Tsvetolin Shulev
  • 0 in reply to Cvetolin Shulev-XID
    Prodigy 50 points

    Hello,

    Thank you for the response! I am building the 4.9.69 kernel and will report back shortly. In the meantime, here is the relevant portion of our device tree:

    hdqw1w: 1w@480b2000 {
        pinctrl-names = "default";
        pinctrl-0 = <&hdq_pins>;
        status = "okay";
        compatible = "ti,omap3-1w";
        reg = <0x480b2000 0x1000>;
        interrupts = <GIC_SPI 53 IRQ_TYPE_LEVEL_HIGH>;
        ti,hwmods = "hdq1w";
        ti,mode = "1w";
    };

  • 0 in reply to Cvetolin Shulev-XID
    Prodigy 50 points
    Ok, all of the same issues happen with the 4.6.69 kernel (additionally ethernet doesn't appear to work but I imagine that's unrelated).
  • 0 in reply to Jonathan Fisher90
    TI__Mastermind 30836 points
    Jonathan,

    The latest work we've done is for the AM437x device family. We have not validated it, nor can we support it, on AM57x, at this time.

    I will log your request as a Feature Request. Any information that you can provide to help us disposition this request is appreciated.
  • 0 in reply to RonB
    Prodigy 50 points

    OK, we will have to go another way for our product, then.

    In short, we were evaluating the on-chip 1-wire interface for devices on the bus that can appear and disappear at runtime. The problems were easily reproduced on a bench supply with the omap_hdq.c driver installed (with DEBUG enabled for that kernel module) -- just connect and disconnect a device from the bus and the bus seems to stop getting interrupts.

    Thank you,

    Jonathan