[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <b59f9861afd658008fbc4f58b75a995bfe00d6ae.1575932654.git.robin.murphy@arm.com>
Date: Tue, 10 Dec 2019 13:24:31 +0000
From: Robin Murphy <robin.murphy@....com>
To: lee.jones@...aro.org
Cc: linux-kernel@...r.kernel.org, heiko@...ech.de, smoch@....de,
linux.amoon@...il.com, linux-rockchip@...ts.infradead.org
Subject: [PATCH 2/4] mfd: rk808: Always register syscore ops
Registering the syscore shutdown notifier even when it's a
no-op for the given RK8xx variant should be harmless, and
saves a lot of bother in handling unregistering on probe
failure or module removal, which has been woefully lacking.
Signed-off-by: Robin Murphy <robin.murphy@....com>
---
drivers/mfd/rk808.c | 5 ++++-
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/drivers/mfd/rk808.c b/drivers/mfd/rk808.c
index f2f2f98552a0..387105830736 100644
--- a/drivers/mfd/rk808.c
+++ b/drivers/mfd/rk808.c
@@ -623,7 +623,6 @@ static int rk808_probe(struct i2c_client *client,
nr_pre_init_regs = ARRAY_SIZE(rk817_pre_init_reg);
cells = rk817s;
nr_cells = ARRAY_SIZE(rk817s);
- register_syscore_ops(&rk808_syscore_ops);
break;
default:
dev_err(&client->dev, "Unsupported RK8XX ID %lu\n",
@@ -667,6 +666,7 @@ static int rk808_probe(struct i2c_client *client,
}
rk808_i2c_client = client;
+ register_syscore_ops(&rk808_syscore_ops);
ret = devm_mfd_add_devices(&client->dev, PLATFORM_DEVID_NONE,
cells, nr_cells, NULL, 0,
@@ -684,6 +684,7 @@ static int rk808_probe(struct i2c_client *client,
return 0;
err_irq:
+ unregister_syscore_ops(&rk808_syscore_ops);
regmap_del_irq_chip(client->irq, rk808->irq_data);
return ret;
}
@@ -694,6 +695,8 @@ static int rk808_remove(struct i2c_client *client)
regmap_del_irq_chip(client->irq, rk808->irq_data);
+ unregister_syscore_ops(&rk808_syscore_ops);
+
/**
* pm_power_off may points to a function from another module.
* Check if the pointer is set by us and only then overwrite it.
--
2.17.1
Powered by blists - more mailing lists