Re: [PATCH v6 5/8] usb: typec: cros_ec_ucsi: Add trace events

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Wed, Sep 11, 2024 at 3:36 PM Heikki Krogerus
<heikki.krogerus@xxxxxxxxxxxxxxx> wrote:
>
> On Tue, Sep 10, 2024 at 10:15:24AM +0000, Łukasz Bartosik wrote:
> > Add trace events to ChromeOS UCSI driver to enable debugging.
> >
> > Signed-off-by: Łukasz Bartosik <ukaszb@xxxxxxxxxxxx>
> > ---
> >  MAINTAINERS                                 |  1 +
> >  drivers/usb/typec/ucsi/cros_ec_ucsi.c       |  8 ++
> >  drivers/usb/typec/ucsi/cros_ec_ucsi_trace.h | 92 +++++++++++++++++++++
> >  3 files changed, 101 insertions(+)
> >  create mode 100644 drivers/usb/typec/ucsi/cros_ec_ucsi_trace.h
> >
> > diff --git a/MAINTAINERS b/MAINTAINERS
> > index 8c030ea0b503..d084f32208f0 100644
> > --- a/MAINTAINERS
> > +++ b/MAINTAINERS
> > @@ -5306,6 +5306,7 @@ M:      Łukasz Bartosik <ukaszb@xxxxxxxxxxxx>
> >  L:   chrome-platform@xxxxxxxxxxxxxxx
> >  S:   Maintained
> >  F:   drivers/usb/typec/ucsi/cros_ec_ucsi.c
> > +F:   drivers/usb/typec/ucsi/cros_ec_ucsi_trace.h
> >
> >  CHRONTEL CH7322 CEC DRIVER
> >  M:   Joe Tessler <jrt@xxxxxxxxxx>
> > diff --git a/drivers/usb/typec/ucsi/cros_ec_ucsi.c b/drivers/usb/typec/ucsi/cros_ec_ucsi.c
> > index 422b3b14028c..70185616ec86 100644
> > --- a/drivers/usb/typec/ucsi/cros_ec_ucsi.c
> > +++ b/drivers/usb/typec/ucsi/cros_ec_ucsi.c
> > @@ -16,7 +16,9 @@
> >  #include <linux/slab.h>
> >  #include <linux/wait.h>
> >
> > +#define CREATE_TRACE_POINTS
> >  #include "ucsi.h"
> > +#include "cros_ec_ucsi_trace.h"
> >
> >  /*
> >   * Maximum size in bytes of a UCSI message between AP and EC
> > @@ -62,6 +64,8 @@ static int cros_ucsi_read(struct ucsi *ucsi, unsigned int offset, void *val,
> >               dev_warn(udata->dev, "Failed to send EC message UCSI_PPM_GET: error=%d", ret);
> >               return ret;
> >       }
> > +
> > +     trace_cros_ec_opm_to_ppm_rd(offset, val, val_len);
> >       return 0;
> >  }
> >
> > @@ -96,6 +100,8 @@ static int cros_ucsi_async_control(struct ucsi *ucsi, u64 cmd)
> >               dev_warn(udata->dev, "Failed to send EC message UCSI_PPM_SET: error=%d", ret);
> >               return ret;
> >       }
> > +
> > +     trace_cros_ec_opm_to_ppm_wr(req->offset, &cmd, sizeof(cmd));
> >       return 0;
> >  }
> >
> > @@ -138,6 +144,8 @@ static void cros_ucsi_work(struct work_struct *work)
> >       struct cros_ucsi_data *udata = container_of(work, struct cros_ucsi_data, work);
> >       u32 cci;
> >
> > +     trace_cros_ec_ppm_to_opm(0);
> > +
> >       if (cros_ucsi_read_cci(udata->ucsi, &cci))
> >               return;
> >
> > diff --git a/drivers/usb/typec/ucsi/cros_ec_ucsi_trace.h b/drivers/usb/typec/ucsi/cros_ec_ucsi_trace.h
> > new file mode 100644
> > index 000000000000..b765ef5c8236
> > --- /dev/null
> > +++ b/drivers/usb/typec/ucsi/cros_ec_ucsi_trace.h
> > @@ -0,0 +1,92 @@
> > +/* SPDX-License-Identifier: GPL-2.0 */
> > +
> > +#undef TRACE_SYSTEM
> > +#define TRACE_SYSTEM cros_ec_ucsi
> > +
> > +#if !defined(__CROS_EC_UCSI_TRACE_H) || defined(TRACE_HEADER_MULTI_READ)
> > +#define __CROS_EC_UCSI_TRACE_H
> > +
> > +#include <linux/tracepoint.h>
> > +
> > +#define decode_cmd(cmd)                                                                      \
> > +     __print_symbolic(cmd,                                                           \
> > +             { 0,                            "Unknown command"               },      \
> > +             { UCSI_PPM_RESET,               "PPM_RESET"                     },      \
> > +             { UCSI_CONNECTOR_RESET,         "CONNECTOR_RESET,"              },      \
> > +             { UCSI_ACK_CC_CI,               "ACK_CC_CI"                     },      \
> > +             { UCSI_SET_NOTIFICATION_ENABLE, "SET_NOTIFICATION_ENABLE"       },      \
> > +             { UCSI_GET_CAPABILITY,          "GET_CAPABILITY"                },      \
> > +             { UCSI_GET_CONNECTOR_CAPABILITY, "GET_CONNECTOR_CAPABILITY"     },      \
> > +             { UCSI_SET_UOM,                 "SET_UOM"                       },      \
> > +             { UCSI_SET_UOR,                 "SET_UOR"                       },      \
> > +             { UCSI_SET_PDM,                 "SET_PDM"                       },      \
> > +             { UCSI_SET_PDR,                 "SET_PDR"                       },      \
> > +             { UCSI_GET_ALTERNATE_MODES,     "GET_ALTERNATE_MODES"           },      \
> > +             { UCSI_GET_CAM_SUPPORTED,       "GET_CAM_SUPPORTED"             },      \
> > +             { UCSI_GET_CURRENT_CAM,         "GET_CURRENT_CAM"               },      \
> > +             { UCSI_SET_NEW_CAM,             "SET_NEW_CAM"                   },      \
> > +             { UCSI_GET_PDOS,                "GET_PDOS"                      },      \
> > +             { UCSI_GET_CABLE_PROPERTY,      "GET_CABLE_PROPERTY"            },      \
> > +             { UCSI_GET_CONNECTOR_STATUS,    "GET_CONNECTOR_STATUS"          },      \
> > +             { UCSI_GET_ERROR_STATUS,        "GET_ERROR_STATUS"              })
>
> Couldn't you just export ucsi_cmd_str() ?
>

I ran into an issue when using exported ucsi_cmd_str(). Trace-cmd
report reports [FAILED TO PARSE].
For example
trace-cmd record -e cros_ec_ucsi
...
trace-cmd report
     kworker/0:3-1891  [000]    59.432234: cros_ec_ppm_to_opm:   notification
     kworker/0:3-1891  [000]    59.432971: cros_ec_opm_to_ppm_rd:
[FAILED TO PARSE] cmd=4 offset=4 length=4 msg=ARRAY[04, 00, 00, 20]

I believe this issue applies also to some UCSI trace events.

With decode_cmd from the above it works fine
For example
trace-cmd record -e cros_ec_ucsi
...
trace-cmd report
     kworker/0:1-10    [000]    49.248920: cros_ec_ppm_to_opm:   notification
     kworker/0:1-10    [000]    49.249857: cros_ec_opm_to_ppm_rd:
(CCI) : 04 00 00 20
     kworker/0:1-10    [000]    49.253230: cros_ec_opm_to_ppm_wr:
(CTRL) GET_CONNECTOR_STATUS: 12 00 02 00 00 00 00 00

> > +#define decode_offset(offset)                                        \
> > +     __print_symbolic(offset,                                \
> > +             { UCSI_VERSION,         "VER"           },      \
> > +             { UCSI_CCI,             "CCI"           },      \
> > +             { UCSI_CONTROL,         "CTRL"          },      \
> > +             { UCSI_MESSAGE_IN,      "MSG_IN"        },      \
> > +             { UCSI_MESSAGE_OUT,     "MSG_OUT"       },      \
> > +             { UCSIv2_MESSAGE_OUT,   "MSG_OUTv2"     })
> > +
> > +DECLARE_EVENT_CLASS(cros_ec_opm_to_ppm,
> > +     TP_PROTO(u16 offset, const void *value, size_t length),
> > +     TP_ARGS(offset, value, length),
> > +     TP_STRUCT__entry(
> > +             __field(u8, cmd)
> > +             __field(u16, offset)
> > +             __field(size_t, length)
> > +             __dynamic_array(char, msg, length)
> > +     ),
> > +     TP_fast_assign(
> > +             __entry->cmd = *((u64 *) value + 3);
> > +             __entry->offset = offset;
> > +             __entry->length = length;
> > +             memcpy(__get_dynamic_array(msg), value, length);
> > +     ),
> > +     TP_printk("(%s) %s: %s",
> > +             decode_offset(__entry->offset),
> > +             __entry->offset == UCSI_CONTROL ?
> > +             decode_cmd(__entry->cmd) : "",
> > +             __print_hex(__get_dynamic_array(msg), __entry->length))
> > +);
> > +
> > +DEFINE_EVENT(cros_ec_opm_to_ppm, cros_ec_opm_to_ppm_rd,
> > +     TP_PROTO(u16 offset, const void *value, size_t length),
> > +     TP_ARGS(offset, value, length)
> > +);
> > +
> > +DEFINE_EVENT(cros_ec_opm_to_ppm, cros_ec_opm_to_ppm_wr,
> > +     TP_PROTO(u16 offset, const void *value, size_t length),
> > +     TP_ARGS(offset, value, length)
> > +);
> > +
> > +TRACE_EVENT(cros_ec_ppm_to_opm,
> > +     TP_PROTO(int x),
> > +     TP_ARGS(x),
> > +     TP_STRUCT__entry(__array(char, x, 0)),
> > +     TP_fast_assign((void)x),
> > +     TP_printk("notification%s", "")
> > +);
>
> This does not look cros_ec specific. Could you check if this could be
> made part of the ucsi core tracepoints?
>

Good point. I will look into it.

> I can also look into this more carefully, and throw ideas if you like.
> Let me know.
>

I would definitely like to hear your ideas on the topic :).

Thanks,
Lukasz

> Br,



>
> --
> heikki





[Index of Archives]     [Linux Media]     [Linux Input]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]     [Old Linux USB Devel Archive]

  Powered by Linux