Re: [PATCH net-next v4 1/3] devlink: introduce framework for selftests

[Vikas Gupta]

Hi Jiri,


On Thu, Jul 21, 2022 at 6:27 PM Jiri Pirko <jiri@xxxxxxxxxx> wrote:
>
> Thu, Jul 21, 2022 at 09:21:19AM CEST, vikas.gupta@xxxxxxxxxxxx wrote:
> >Add a framework for running selftests.
> >Framework exposes devlink commands and test suite(s) to the user
> >to execute and query the supported tests by the driver.
> >
> >Below are new entries in devlink_nl_ops
> >devlink_nl_cmd_selftests_list_doit/dumpit: To query the supported
> >selftests by the drivers.
> >devlink_nl_cmd_selftests_run: To execute selftests. Users can
> >provide a test mask for executing group tests or standalone tests.
> >
> >Documentation/networking/devlink/ path is already part of MAINTAINERS &
> >the new files come under this path. Hence no update needed to the
> >MAINTAINERS
> >
> >Signed-off-by: Vikas Gupta <vikas.gupta@xxxxxxxxxxxx>
> >Reviewed-by: Michael Chan <michael.chan@xxxxxxxxxxxx>
> >Reviewed-by: Andy Gospodarek <andrew.gospodarek@xxxxxxxxxxxx>
> >---
> > .../networking/devlink/devlink-selftests.rst  |  38 +++
> > include/net/devlink.h                         |  20 ++
> > include/uapi/linux/devlink.h                  |  29 +++
> > net/core/devlink.c                            | 225 ++++++++++++++++++
> > 4 files changed, 312 insertions(+)
> > create mode 100644 Documentation/networking/devlink/devlink-selftests.rst
> >
> >diff --git a/Documentation/networking/devlink/devlink-selftests.rst b/Documentation/networking/devlink/devlink-selftests.rst
> >new file mode 100644
> >index 000000000000..0e9727895987
> >--- /dev/null
> >+++ b/Documentation/networking/devlink/devlink-selftests.rst
> >@@ -0,0 +1,38 @@
> >+.. SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
> >+
> >+=================
> >+Devlink Selftests
> >+=================
> >+
> >+The ``devlink-selftests`` API allows executing selftests on the device.
> >+
> >+Tests Mask
> >+==========
> >+The ``devlink-selftests`` command should be run with a mask indicating
> >+the tests to be executed.
> >+
> >+Tests Description
> >+=================
> >+The following is a list of tests that drivers may execute.
> >+
> >+.. list-table:: List of tests
> >+   :widths: 5 90
> >+
> >+   * - Name
> >+     - Description
> >+   * - ``DEVLINK_SELFTEST_FLASH``
> >+     - Devices may have the firmware on non-volatile memory on the board, e.g.
> >+       flash. This particular test helps to run a flash selftest on the device.
> >+       Implementation of the test is left to the driver/firmware.
> >+
> >+example usage
> >+-------------
> >+
> >+.. code:: shell
> >+
> >+    # Query selftests supported on the devlink device
> >+    $ devlink dev selftests show DEV
> >+    # Query selftests supported on all devlink devices
> >+    $ devlink dev selftests show
> >+    # Executes selftests on the device
> >+    $ devlink dev selftests run DEV test flash
>
> "test_id" to be consistend with the attr name and outputs. Please see
What is "test_id" referring to in this document? Can you please elaborate ?

> below. Devlink cmdline would accept "test" as well, so you can still use
Are you mentioning the "test" argument in the above devlink command line option?

Thanks,
Vikas
> this.

>
>
> >diff --git a/include/net/devlink.h b/include/net/devlink.h
> >index 88c701b375a2..085d761f1cd3 100644
> >--- a/include/net/devlink.h
> >+++ b/include/net/devlink.h
> >@@ -1509,6 +1509,26 @@ struct devlink_ops {
> >                                   struct devlink_rate *parent,
> >                                   void *priv_child, void *priv_parent,
> >                                   struct netlink_ext_ack *extack);
> >+      /**
> >+       * selftests_check() - queries if selftest is supported
> >+       * @devlink: Devlink instance
>
> Why capital "D"?
>
>
> >+       * @test_id: test index
> >+       * @extack: extack for reporting error messages
> >+       *
> >+       * Return: true if test is supported by the driver
> >+       */
> >+      bool (*selftest_check)(struct devlink *devlink, int test_id,
>
> Why this is an "int". I would be surprised to see a negative value here.
> Have this unsigned please.
>
>
> >+                             struct netlink_ext_ack *extack);
> >+      /**
> >+       * selftest_run() - Runs a selftest
> >+       * @devlink: Devlink instance
> >+       * @test_id: test index
> >+       * @extack: extack for reporting error messages
> >+       *
> >+       * Return: Result of the test
> >+       */
> >+      u8 (*selftest_run)(struct devlink *devlink, int test_id,
>
> There too.
>
>
> >+                         struct netlink_ext_ack *extack);
> > };
> >
> > void *devlink_priv(struct devlink *devlink);
> >diff --git a/include/uapi/linux/devlink.h b/include/uapi/linux/devlink.h
> >index b3d40a5d72ff..469846f40e6d 100644
> >--- a/include/uapi/linux/devlink.h
> >+++ b/include/uapi/linux/devlink.h
> >@@ -136,6 +136,9 @@ enum devlink_command {
> >       DEVLINK_CMD_LINECARD_NEW,
> >       DEVLINK_CMD_LINECARD_DEL,
> >
> >+      DEVLINK_CMD_SELFTESTS_LIST,     /* can dump */
>
> The rest of the commands are named "_GET". Please be consistent with
> them.
>
>
> >+      DEVLINK_CMD_SELFTESTS_RUN,
> >+
> >       /* add new commands above here */
> >       __DEVLINK_CMD_MAX,
> >       DEVLINK_CMD_MAX = __DEVLINK_CMD_MAX - 1
> >@@ -276,6 +279,31 @@ enum {
> > #define DEVLINK_SUPPORTED_FLASH_OVERWRITE_SECTIONS \
> >       (_BITUL(__DEVLINK_FLASH_OVERWRITE_MAX_BIT) - 1)
> >
> >+/* Commonly used test cases */
>
> What do you mean by "commonly". Are there some others that are not
> "common"? I don't follow.
>
>
> >+enum devlink_selftest_attr {
> >+      DEVLINK_SELFTEST_ATTR_UNSPEC,
> >+      DEVLINK_SELFTEST_ATTR_FLASH,            /* flag */
> >+
> >+      __DEVLINK_SELFTEST_ATTR_MAX,
> >+      DEVLINK_SELFTEST_ATTR_MAX = __DEVLINK_SELFTEST_ATTR_MAX - 1
> >+};
>
> To be consistent with the attr that caries this:
>
> enum devlink_attr_selftest_test_id {
>         DEVLINK_ATTR_SELFTEST_TEST_ID_UNSPEC,
>         DEVLINK_ATTR_SELFTEST_TEST_ID_FLASH,            /* flag */
>
>         __DEVLINK_ATTR_SELFTEST_TEST_ID_MAX,
>         DEVLINK_ATTR_SELFTEST_TEST_ID_MAX = __DEVLINK_ATTR_SELFTEST_TEST_ID_MAX - 1
>
> >+
> >+enum devlink_selftest_result {
> >+      DEVLINK_SELFTEST_SKIP,
> >+      DEVLINK_SELFTEST_PASS,
> >+      DEVLINK_SELFTEST_FAIL
>
> It is common to have the enum name be root of names of the values.
> Also, be consistent with the attr this value is carried over:
>
> enum devlink_selftest_test_status {
>         DEVLINK_SELFTEST_TEST_STATUS_SKIP,
>         DEVLINK_SELFTEST_TEST_STATUS_PASS,
>         DEVLINK_SELFTEST_TEST_STATUS_FAIL
>
> That way, it is obvious to which enum the value belongs.
>
>
> >+};
> >+
> >+enum devlink_selftest_result_attr {
> >+      DEVLINK_SELFTEST_ATTR_RESULT_UNSPEC,
> >+      DEVLINK_SELFTEST_ATTR_RESULT,           /* nested */
> >+      DEVLINK_SELFTEST_ATTR_TEST_ID,          /* u32, devlink_selftest_attr */
>
> add "enum" ?
>
> >+      DEVLINK_SELFTEST_ATTR_TEST_STATUS,      /* u8, devlink_selftest_result */
>
> add "enum" ?
>
> The same note as above:
> enum devlink_attr_selftest_result {
>         DEVLINK_ATTR_SELFTEST_RESULT_UNSPEC,
>         DEVLINK_ATTR_SELFTEST_RESULT,             /* nested */
>         DEVLINK_ATTR_SELFTEST_RESULT_TEST_ID,     /* u32, enum devlink_selftest_attr */
>         DEVLINK_ATTR_SELFTEST_RESULT_TEST_STATUS, /* u8, devlink_selftest_result */
>
>
>
>
> >+
> >+      __DEVLINK_SELFTEST_ATTR_RES_MAX,
> >+      DEVLINK_SELFTEST_ATTR_RES_MAX = __DEVLINK_SELFTEST_ATTR_RES_MAX - 1
> >+};
> >+
> > /**
> >  * enum devlink_trap_action - Packet trap action.
> >  * @DEVLINK_TRAP_ACTION_DROP: Packet is dropped by the device and a copy is not
> >@@ -576,6 +604,7 @@ enum devlink_attr {
> >       DEVLINK_ATTR_LINECARD_TYPE,             /* string */
> >       DEVLINK_ATTR_LINECARD_SUPPORTED_TYPES,  /* nested */
> >
> >+      DEVLINK_ATTR_SELFTESTS_INFO,            /* nested */
> >       /* add new attributes above here, update the policy in devlink.c */
> >
> >       __DEVLINK_ATTR_MAX,
> >diff --git a/net/core/devlink.c b/net/core/devlink.c
> >index a9776ea923ae..ef9439f2502f 100644
> >--- a/net/core/devlink.c
> >+++ b/net/core/devlink.c
> >@@ -198,6 +198,10 @@ static const struct nla_policy devlink_function_nl_policy[DEVLINK_PORT_FUNCTION_
> >                                DEVLINK_PORT_FN_STATE_ACTIVE),
> > };
> >
> >+static const struct nla_policy devlink_selftest_nl_policy[DEVLINK_SELFTEST_ATTR_MAX + 1] = {
> >+      [DEVLINK_SELFTEST_ATTR_FLASH] = { .type = NLA_FLAG },
> >+};
> >+
> > static DEFINE_XARRAY_FLAGS(devlinks, XA_FLAGS_ALLOC);
> > #define DEVLINK_REGISTERED XA_MARK_1
> >
> >@@ -4791,6 +4795,215 @@ static int devlink_nl_cmd_flash_update(struct sk_buff *skb,
> >       return ret;
> > }
> >
> >+static int
> >+devlink_nl_selftests_fill(struct sk_buff *msg, struct devlink *devlink,
> >+                        u32 portid, u32 seq, int flags,
> >+                        struct netlink_ext_ack *extack)
> >+{
> >+      struct nlattr *selftests_list;
> >+      void *hdr;
> >+      int err;
> >+      int i;
> >+
> >+      hdr = genlmsg_put(msg, portid, seq, &devlink_nl_family, flags,
> >+                        DEVLINK_CMD_SELFTESTS_LIST);
> >+      if (!hdr)
> >+              return -EMSGSIZE;
> >+
> >+      err = -EMSGSIZE;
> >+      if (devlink_nl_put_handle(msg, devlink))
> >+              goto err_cancel_msg;
> >+
> >+      selftests_list = nla_nest_start(msg, DEVLINK_ATTR_SELFTESTS_INFO);
> >+      if (!selftests_list)
> >+              goto err_cancel_msg;
> >+
> >+      for (i = 1; i < DEVLINK_SELFTEST_ATTR_MAX + 1; i++) {
>
> **)
> It is a bit odd to see "1" here. Maybe "DEVLINK_SELFTEST_ATTR_UNSPEC + 1"
> would be more obvious for the reader.
>
> also:
> i < DEVLINK_SELFTEST_ATTR_MAX + 1
> would be rather nicer to be:
> i <= DEVLINK_SELFTEST_ATTR_MAX
>
>
> >+              if (devlink->ops->selftest_check(devlink, i, extack)) {
> >+                      err = nla_put_flag(msg, i);
> >+                      if (err)
> >+                              goto err_cancel_msg;
> >+              }
> >+      }
> >+
> >+      nla_nest_end(msg, selftests_list);
> >+
>
> No need for this empty line.
>
>
> >+      genlmsg_end(msg, hdr);
> >+
>
> No need for this empty line.
>
>
> >+      return 0;
> >+
> >+err_cancel_msg:
> >+      genlmsg_cancel(msg, hdr);
> >+      return err;
> >+}
> >+
> >+static int devlink_nl_cmd_selftests_list_doit(struct sk_buff *skb,
> >+                                            struct genl_info *info)
> >+{
> >+      struct devlink *devlink = info->user_ptr[0];
> >+      struct sk_buff *msg;
> >+      int err;
> >+
> >+      if (!devlink->ops->selftest_check)
> >+              return -EOPNOTSUPP;
> >+
> >+      msg = nlmsg_new(NLMSG_DEFAULT_SIZE, GFP_KERNEL);
> >+      if (!msg)
> >+              return -ENOMEM;
> >+
> >+      err = devlink_nl_selftests_fill(msg, devlink, info->snd_portid,
> >+                                      info->snd_seq, 0, info->extack);
> >+      if (err) {
> >+              nlmsg_free(msg);
> >+              return err;
> >+      }
> >+
> >+      return genlmsg_reply(msg, info);
> >+}
> >+
> >+static int devlink_nl_cmd_selftests_list_dumpit(struct sk_buff *msg,
> >+                                              struct netlink_callback *cb)
> >+{
> >+      struct devlink *devlink;
> >+      int start = cb->args[0];
> >+      unsigned long index;
> >+      int idx = 0;
> >+      int err = 0;
> >+
> >+      mutex_lock(&devlink_mutex);
> >+      xa_for_each_marked(&devlinks, index, devlink, DEVLINK_REGISTERED) {
> >+              if (!devlink_try_get(devlink))
> >+                      continue;
> >+
> >+              if (!net_eq(devlink_net(devlink), sock_net(msg->sk)))
> >+                      goto retry;
> >+
> >+              if (idx < start || !devlink->ops->selftest_check)
> >+                      goto inc;
> >+
> >+              mutex_lock(&devlink->lock);
> >+              err = devlink_nl_selftests_fill(msg, devlink,
> >+                                              NETLINK_CB(cb->skb).portid,
> >+                                              cb->nlh->nlmsg_seq, NLM_F_MULTI,
> >+                                              cb->extack);
> >+              mutex_unlock(&devlink->lock);
> >+              if (err) {
> >+                      devlink_put(devlink);
> >+                      break;
> >+              }
> >+inc:
> >+              idx++;
> >+retry:
> >+              devlink_put(devlink);
> >+      }
> >+      mutex_unlock(&devlink_mutex);
> >+
> >+      if (err != -EMSGSIZE)
> >+              return err;
> >+
> >+      cb->args[0] = idx;
> >+      return msg->len;
> >+}
> >+
> >+static int devlink_selftest_result_put(struct sk_buff *skb, int test_id,
>
> unsigned.
>
> >+                                     u8 result)
>
> Please be consistend and call this "test_status"
>
>
> >+{
> >+      struct nlattr *result_attr;
> >+
> >+      result_attr = nla_nest_start(skb, DEVLINK_SELFTEST_ATTR_RESULT);
> >+      if (!result_attr)
> >+              return -EMSGSIZE;
> >+
> >+      if (nla_put_u32(skb, DEVLINK_SELFTEST_ATTR_TEST_ID, test_id) ||
> >+          nla_put_u8(skb, DEVLINK_SELFTEST_ATTR_TEST_STATUS, result))
> >+              goto nla_put_failure;
> >+
> >+      nla_nest_end(skb, result_attr);
> >+
>
> No need for this empty line.
>
>
> >+      return 0;
> >+
> >+nla_put_failure:
> >+      nla_nest_cancel(skb, result_attr);
> >+      return -EMSGSIZE;
> >+}
> >+
> >+static int devlink_nl_cmd_selftests_run(struct sk_buff *skb,
> >+                                      struct genl_info *info)
> >+{
> >+      struct nlattr *tb[DEVLINK_SELFTEST_ATTR_MAX + 1];
> >+      struct devlink *devlink = info->user_ptr[0];
> >+      struct nlattr *attrs, *tests_info;
> >+      struct sk_buff *msg;
> >+      void *hdr;
> >+      int err;
> >+      int i;
> >+
> >+      if (!devlink->ops->selftest_run)
> >+              return -EOPNOTSUPP;
> >+
> >+      if (!info->attrs[DEVLINK_ATTR_SELFTESTS_INFO])
>
> Fill extack message here please.
>
>
> >+              return -EINVAL;
> >+
> >+      attrs = info->attrs[DEVLINK_ATTR_SELFTESTS_INFO];
> >+
> >+      err = nla_parse_nested(tb, DEVLINK_SELFTEST_ATTR_MAX, attrs,
> >+                             devlink_selftest_nl_policy, info->extack);
> >+      if (err < 0)
> >+              return err;
> >+
> >+      msg = nlmsg_new(NLMSG_DEFAULT_SIZE, GFP_KERNEL);
> >+      if (!msg)
> >+              return -ENOMEM;
> >+
> >+      err = -EMSGSIZE;
> >+      hdr = genlmsg_put(msg, info->snd_portid, info->snd_seq,
> >+                        &devlink_nl_family, 0, DEVLINK_CMD_SELFTESTS_RUN);
> >+      if (!hdr)
> >+              goto free_msg;
> >+
> >+      if (devlink_nl_put_handle(msg, devlink))
> >+              goto genlmsg_cancel;
> >+
> >+      tests_info = nla_nest_start(msg, DEVLINK_ATTR_SELFTESTS_INFO);
> >+      if (!tests_info)
> >+              goto genlmsg_cancel;
> >+
> >+      for (i = 1; i < DEVLINK_SELFTEST_ATTR_MAX + 1; i++) {
>
> Same notes to the iteration as above. **
>
>
> >+              u8 res = DEVLINK_SELFTEST_SKIP;
>
> u8 test_status;
>
>
> >+
> >+              if (nla_get_flag(tb[i])) {
> >+                      if (devlink->ops->selftest_check &&
>
> No need to test in every iteration. I think it is safe to assume
> that driver that does not fill selftest_check() does not support
> selftests at all, so please move to the beginning of this function
> alongside selftest_run() check:
>
>         if (!devlink->ops->selftest_run || !devlink->ops->selftest_check)
>                 return -EOPNOTSUPP;
>
> >+                          !devlink->ops->selftest_check(devlink, i,
> >+                                                        info->extack)) {
> >+                              err = devlink_selftest_result_put(msg, i, res);
>
> Just do devlink_selftest_result_put(msg, i, .._SKIP); here and avoid
> initializing "res" at the beginning.
>
>
> >+                              if (err)
> >+                                      goto selftests_list_nest_cancel;
> >+                              continue;
> >+                      }
> >+
> >+                      res = devlink->ops->selftest_run(devlink, i,
> >+                                                       info->extack);
> >+                      err = devlink_selftest_result_put(msg, i, res);
> >+                      if (err)
> >+                              goto selftests_list_nest_cancel;
> >+              }
> >+      }
> >+
> >+      nla_nest_end(msg, tests_info);
> >+
>
> No need for this empty line.
>
>
> >+      genlmsg_end(msg, hdr);
> >+
>
> No need for this empty line.
>
>
> >+      return genlmsg_reply(msg, info);
> >+
> >+selftests_list_nest_cancel:
> >+      nla_nest_cancel(msg, tests_info);
> >+genlmsg_cancel:
> >+      genlmsg_cancel(msg, hdr);
> >+free_msg:
> >+      nlmsg_free(msg);
> >+      return err;
> >+}
> >+
> > static const struct devlink_param devlink_param_generic[] = {
> >       {
> >               .id = DEVLINK_PARAM_GENERIC_ID_INT_ERR_RESET,
> >@@ -8997,6 +9210,7 @@ static const struct nla_policy devlink_nl_policy[DEVLINK_ATTR_MAX + 1] = {
> >       [DEVLINK_ATTR_RATE_PARENT_NODE_NAME] = { .type = NLA_NUL_STRING },
> >       [DEVLINK_ATTR_LINECARD_INDEX] = { .type = NLA_U32 },
> >       [DEVLINK_ATTR_LINECARD_TYPE] = { .type = NLA_NUL_STRING },
> >+      [DEVLINK_ATTR_SELFTESTS_INFO] = { .type = NLA_NESTED },
> > };
> >
> > static const struct genl_small_ops devlink_nl_ops[] = {
> >@@ -9356,6 +9570,17 @@ static const struct genl_small_ops devlink_nl_ops[] = {
> >               .doit = devlink_nl_cmd_trap_policer_set_doit,
> >               .flags = GENL_ADMIN_PERM,
> >       },
> >+      {
> >+              .cmd = DEVLINK_CMD_SELFTESTS_LIST,
> >+              .doit = devlink_nl_cmd_selftests_list_doit,
> >+              .dumpit = devlink_nl_cmd_selftests_list_dumpit
> >+              /* can be retrieved by unprivileged users */
> >+      },
> >+      {
> >+              .cmd = DEVLINK_CMD_SELFTESTS_RUN,
> >+              .doit = devlink_nl_cmd_selftests_run,
> >+              .flags = GENL_ADMIN_PERM,
> >+      },
> > };
> >
> > static struct genl_family devlink_nl_family __ro_after_init = {
> >--
> >2.31.1
> >
>
>

Attachment: smime.p7s
Description: S/MIME Cryptographic Signature