matthew/sonic-buildimage
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity
342604a4bb
sonic-buildimage/platform/mellanox/non-upstream-patches/patches/0171-platform-mellanox-mlxreg-lc-Fix-cleanup-on-failure-a.patch
Kebo Liu 27f15d40e1 [Mellanox] Update HW-MGMT package to new version V.7.0030.1011 (#16239) 
- Why I did it
1. Update Mellanox HW-MGMT package to newer version V.7.0030.1011
2. Replace the SONiC PMON Thermal control algorithm with the one inside the HW-MGMT package on all Nvidia platforms
3. Support Spectrum-4 systems

- How I did it
1. Update the HW-MGMT package version number and submodule pointer
2. Remove the thermal control algorithm implementation from Mellanox platform API
3. Revise the patch to HW-MGMT package which will disable HW-MGMT from running on SIMX
4. Update the downstream kernel patch list

Signed-off-by: Kebo Liu <kebol@nvidia.com>
2023-09-21 18:34:07 +08:00

195 lines
7.0 KiB
Diff
Raw Blame History

From 7a45b6a2a14c292e89a09afde8fcc7d1ceb20ebd Mon Sep 17 00:00:00 2001
From: Vadim Pasternak <vadimp@nvidia.com>
Date: Wed, 22 Jun 2022 13:01:15 +0300
Subject: [PATCH backport 5.10 171/182] platform/mellanox: mlxreg-lc: Fix
cleanup on failure and add more verbosity in error flow
Clean client object in case of probing failure.
Prevent running remove routine in case probing failed.
Add error log for each kind of failures during probing.
Fixes: 62f9529b8d5c ("platform/mellanox: mlxreg-lc: Add initial support for Nvidia line card devices")
Signed-off-by: Vadim Pasternak <vadimp@nvidia.com>
---
drivers/platform/mellanox/mlxreg-lc.c | 85 ++++++++++++++++++++-------
1 file changed, 63 insertions(+), 22 deletions(-)
diff --git a/drivers/platform/mellanox/mlxreg-lc.c b/drivers/platform/mellanox/mlxreg-lc.c
index 75c28179dd07..e3caccb1a528 100644
--- a/drivers/platform/mellanox/mlxreg-lc.c
+++ b/drivers/platform/mellanox/mlxreg-lc.c
@@ -569,9 +569,6 @@ static int mlxreg_lc_event_handler(void *handle, enum mlxreg_hotplug_kind kind,
dev_info(mlxreg_lc->dev, "linecard#%d state %d event kind %d action %d\n",
mlxreg_lc->data->slot, mlxreg_lc->state, kind, action);
- if (!(mlxreg_lc->state & MLXREG_LC_INITIALIZED))
- return 0;
-
switch (kind) {
case MLXREG_HOTPLUG_LC_SYNCED:
/*
@@ -725,8 +722,12 @@ mlxreg_lc_config_init(struct mlxreg_lc *mlxreg_lc, void *regmap,
switch (regval) {
case MLXREG_LC_SN4800_C16:
err = mlxreg_lc_sn4800_c16_config_init(mlxreg_lc, regmap, data);
- if (err)
+ if (err) {
+ dev_err(dev, "Failed to config client %s at bus %d at addr 0x%02x\n",
+ data->hpdev.brdinfo->type, data->hpdev.nr,
+ data->hpdev.brdinfo->addr);
return err;
+ }
break;
default:
return -ENODEV;
@@ -739,8 +740,11 @@ mlxreg_lc_config_init(struct mlxreg_lc *mlxreg_lc, void *regmap,
mlxreg_lc->mux = platform_device_register_resndata(dev, "i2c-mux-mlxcpld", data->hpdev.nr,
NULL, 0, mlxreg_lc->mux_data,
sizeof(*mlxreg_lc->mux_data));
- if (IS_ERR(mlxreg_lc->mux))
+ if (IS_ERR(mlxreg_lc->mux)) {
+ dev_err(dev, "Failed to create mux infra for client %s at bus %d at addr 0x%02x\n",
+ data->hpdev.brdinfo->type, data->hpdev.nr, data->hpdev.brdinfo->addr);
return PTR_ERR(mlxreg_lc->mux);
+ }
/* Register IO access driver. */
if (mlxreg_lc->io_data) {
@@ -749,6 +753,9 @@ mlxreg_lc_config_init(struct mlxreg_lc *mlxreg_lc, void *regmap,
platform_device_register_resndata(dev, "mlxreg-io", data->hpdev.nr, NULL, 0,
mlxreg_lc->io_data, sizeof(*mlxreg_lc->io_data));
if (IS_ERR(mlxreg_lc->io_regs)) {
+ dev_err(dev, "Failed to create regio for client %s at bus %d at addr 0x%02x\n",
+ data->hpdev.brdinfo->type, data->hpdev.nr,
+ data->hpdev.brdinfo->addr);
err = PTR_ERR(mlxreg_lc->io_regs);
goto fail_register_io;
}
@@ -762,6 +769,9 @@ mlxreg_lc_config_init(struct mlxreg_lc *mlxreg_lc, void *regmap,
mlxreg_lc->led_data,
sizeof(*mlxreg_lc->led_data));
if (IS_ERR(mlxreg_lc->led)) {
+ dev_err(dev, "Failed to create LED objects for client %s at bus %d at addr 0x%02x\n",
+ data->hpdev.brdinfo->type, data->hpdev.nr,
+ data->hpdev.brdinfo->addr);
err = PTR_ERR(mlxreg_lc->led);
goto fail_register_led;
}
@@ -818,7 +828,8 @@ static int mlxreg_lc_probe(struct platform_device *pdev)
if (!data->hpdev.adapter) {
dev_err(&pdev->dev, "Failed to get adapter for bus %d\n",
data->hpdev.nr);
- return -EFAULT;
+ err = -EFAULT;
+ goto i2c_get_adapter_fail;
}
/* Create device at the top of line card I2C tree.*/
@@ -827,32 +838,40 @@ static int mlxreg_lc_probe(struct platform_device *pdev)
if (IS_ERR(data->hpdev.client)) {
dev_err(&pdev->dev, "Failed to create client %s at bus %d at addr 0x%02x\n",
data->hpdev.brdinfo->type, data->hpdev.nr, data->hpdev.brdinfo->addr);
-
- i2c_put_adapter(data->hpdev.adapter);
- data->hpdev.adapter = NULL;
- return PTR_ERR(data->hpdev.client);
+ err = PTR_ERR(data->hpdev.client);
+ goto i2c_new_device_fail;
}
regmap = devm_regmap_init_i2c(data->hpdev.client,
&mlxreg_lc_regmap_conf);
if (IS_ERR(regmap)) {
+ dev_err(&pdev->dev, "Failed to create regmap for client %s at bus %d at addr 0x%02x\n",
+ data->hpdev.brdinfo->type, data->hpdev.nr, data->hpdev.brdinfo->addr);
err = PTR_ERR(regmap);
- goto mlxreg_lc_probe_fail;
+ goto devm_regmap_init_i2c_fail;
}
/* Set default registers. */
for (i = 0; i < mlxreg_lc_regmap_conf.num_reg_defaults; i++) {
err = regmap_write(regmap, mlxreg_lc_regmap_default[i].reg,
mlxreg_lc_regmap_default[i].def);
- if (err)
- goto mlxreg_lc_probe_fail;
+ if (err) {
+ dev_err(&pdev->dev, "Failed to set default regmap %d for client %s at bus %d at addr 0x%02x\n",
+ i, data->hpdev.brdinfo->type, data->hpdev.nr,
+ data->hpdev.brdinfo->addr);
+ goto regmap_write_fail;
+ }
}
/* Sync registers with hardware. */
regcache_mark_dirty(regmap);
err = regcache_sync(regmap);
- if (err)
- goto mlxreg_lc_probe_fail;
+ if (err) {
+ dev_err(&pdev->dev, "Failed to sync regmap for client %s at bus %d at addr 0x%02x\n",
+ data->hpdev.brdinfo->type, data->hpdev.nr, data->hpdev.brdinfo->addr);
+ err = PTR_ERR(regmap);
+ goto regcache_sync_fail;
+ }
par_pdata = data->hpdev.brdinfo->platform_data;
mlxreg_lc->par_regmap = par_pdata->regmap;
@@ -863,12 +882,27 @@ static int mlxreg_lc_probe(struct platform_device *pdev)
/* Configure line card. */
err = mlxreg_lc_config_init(mlxreg_lc, regmap, data);
if (err)
- goto mlxreg_lc_probe_fail;
+ goto mlxreg_lc_config_init_fail;
return err;
-mlxreg_lc_probe_fail:
+mlxreg_lc_config_init_fail:
+regcache_sync_fail:
+regmap_write_fail:
+devm_regmap_init_i2c_fail:
+ if (data->hpdev.client) {
+ i2c_unregister_device(data->hpdev.client);
+ data->hpdev.client = NULL;
+ }
+i2c_new_device_fail:
i2c_put_adapter(data->hpdev.adapter);
+ data->hpdev.adapter = NULL;
+i2c_get_adapter_fail:
+ /* Clear event notification callback and handle. */
+ if (data->notifier) {
+ data->notifier->user_handler = NULL;
+ data->notifier->handle = NULL;
+ }
return err;
}
@@ -877,11 +911,18 @@ static int mlxreg_lc_remove(struct platform_device *pdev)
struct mlxreg_core_data *data = dev_get_platdata(&pdev->dev);
struct mlxreg_lc *mlxreg_lc = platform_get_drvdata(pdev);
- /* Clear event notification callback. */
- if (data->notifier) {
- data->notifier->user_handler = NULL;
- data->notifier->handle = NULL;
- }
+ /*
+ * Probing and removing are invoked by hotplug events raised on line card insertion and
+ * removing. If probing procedure fails all data is cleared. However, hotplug event still
+ * will be raised on line card removing and activate removing procedure. In this case there
+ * is nothing to remove.
+ */
+ if (!data->notifier || !data->notifier->handle)
+ return 0;
+
+ /* Clear event notification callback and handle. */
+ data->notifier->user_handler = NULL;
+ data->notifier->handle = NULL;
/* Destroy static I2C device feeding by main power. */
mlxreg_lc_destroy_static_devices(mlxreg_lc, mlxreg_lc->main_devs,
--
2.20.1
Reference in New Issue View Git Blame Copy Permalink