Mon, May 06, 2019 at 12:54:00PM CEST, mo...@mellanox.com wrote:
>
>
>On 5/5/2019 6:52 PM, Jiri Pirko wrote:
>> Sun, May 05, 2019 at 02:33:33AM CEST, sae...@mellanox.com wrote:
>>> From: Moshe Shemesh <mo...@mellanox.com>
>>>
>>> Add support of dump callback for mlx5 FW fatal reporter.
>>> The FW fatal dump use cr-dump functionality to gather cr-space data for
>>> debug. The cr-dump uses vsc interface which is valid even if the FW
>>> command interface is not functional, which is the case in most FW fatal
>>> errors.
>>> The cr-dump is stored as a memory region snapshot to ease read by
>>> address.
>>>
>>> Command example and output:
>>> $ devlink health dump show pci/0000:82:00.0 reporter fw_fatal
>>> devlink_region_name: cr-space snapshot_id: 1
>>>
>>> $ devlink region read pci/0000:82:00.0/cr-space snapshot 1 address 983064
>>> length 8
>>> 00000000000f0018 e1 03 00 00 fb ae a9 3f
>>>
>>> Signed-off-by: Moshe Shemesh <mo...@mellanox.com>
>>> Signed-off-by: Saeed Mahameed <sae...@mellanox.com>
>>> ---
>>> .../net/ethernet/mellanox/mlx5/core/health.c | 39 +++++++++++++++++++
>>> 1 file changed, 39 insertions(+)
>>>
>>> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/health.c
>>> b/drivers/net/ethernet/mellanox/mlx5/core/health.c
>>> index e64f0e32cd67..5271c88ef64c 100644
>>> --- a/drivers/net/ethernet/mellanox/mlx5/core/health.c
>>> +++ b/drivers/net/ethernet/mellanox/mlx5/core/health.c
>>> @@ -547,9 +547,48 @@ mlx5_fw_fatal_reporter_recover(struct
>>> devlink_health_reporter *reporter,
>>> return mlx5_health_care(dev);
>>> }
>>>
>>> +static int
>>> +mlx5_fw_fatal_reporter_dump(struct devlink_health_reporter *reporter,
>>> + struct devlink_fmsg *fmsg, void *priv_ctx)
>>> +{
>>> + struct mlx5_core_dev *dev = devlink_health_reporter_priv(reporter);
>>> + char crdump_region[20];
>>> + u32 snapshot_id;
>>> + int err;
>>> +
>>> + if (!mlx5_core_is_pf(dev)) {
>>> + mlx5_core_err(dev, "Only PF is permitted run FW fatal dump\n");
>>> + return -EPERM;
>>> + }
>>> +
>>> + err = mlx5_crdump_collect(dev, crdump_region, &snapshot_id);
>>> + if (err)
>>> + return err;
>>> +
>>> + if (priv_ctx) {
>>> + struct mlx5_fw_reporter_ctx *fw_reporter_ctx = priv_ctx;
>>> +
>>> + err = mlx5_fw_reporter_ctx_pairs_put(fmsg, fw_reporter_ctx);
>>> + if (err)
>>> + return err;
>>> + }
>>> +
>>> + err = devlink_fmsg_string_pair_put(fmsg, "devlink_region_name",
>>> + crdump_region);
>>
>> Oh come on. You cannot be serious :/ Please do proper linkage to region
>> and snapshot in devlink core.
>>
>
>Not sure I understand what you mean, as I wrote in the commit message,
>the region snapshot added value here, is that user can read data by offset.
If there is a region/snapshot affiliated with a reporter, it should be
linked together in devlink core. Not in the driver by exposing
arbitrary strings...
>
>>
>>
>>> + if (err)
>>> + return err;
>>> +
>>> + err = devlink_fmsg_u32_pair_put(fmsg, "snapshot_id", snapshot_id);
>>> + if (err)
>>> + return err;
>>> +
>>> + return 0;
>>> +}
>>> +
>>> static const struct devlink_health_reporter_ops mlx5_fw_fatal_reporter_ops
>>> = {
>>> .name = "fw_fatal",
>>> .recover = mlx5_fw_fatal_reporter_recover,
>>> + .dump = mlx5_fw_fatal_reporter_dump,
>>> };
>>>
>>> #define MLX5_REPORTER_FW_GRACEFUL_PERIOD 1200000
>>> --
>>> 2.20.1
>>>
