soc: fsl: qe: Check of ioremap return value

BugLink: https://bugs.launchpad.net/bugs/1964361
[ Upstream commit a222fd8541394b36b13c89d1698d9530afd59a9c ]

As the possible failure of the ioremap(), the par_io could be NULL.
Therefore it should be better to check it and return error in order to
guarantee the success of the initiation.
But, I also notice that all the caller like mpc85xx_qe_par_io_init() in
`arch/powerpc/platforms/85xx/common.c` don't check the return value of
the par_io_init().
Actually, par_io_init() needs to check to handle the potential error.
I will submit another patch to fix that.
Anyway, par_io_init() itsely should be fixed.

Fixes: 7aa1aa6ecec2 ("QE: Move QE from arch/powerpc to drivers/soc")
Signed-off-by: Jiasheng Jiang <jiasheng@iscas.ac.cn>
Signed-off-by: Li Yang <leoyang.li@nxp.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Paolo Pisati <paolo.pisati@canonical.com>

diff --git a/drivers/soc/fsl/qe/qe_io.c b/drivers/soc/fsl/qe/qe_io.c
index e277c827bdf336f4789ff438a8413b235fc1ca1e..a5e2d0e5ab511fae5f929a8434fb499f0ea6d53a 100644 (file)
--- a/drivers/soc/fsl/qe/qe_io.c
+++ b/drivers/soc/fsl/qe/qe_io.c
@@ -35,6 +35,8 @@ int par_io_init(struct device_node *np)
        if (ret)
                return ret;
        par_io = ioremap(res.start, resource_size(&res));
+       if (!par_io)
+               return -ENOMEM;
 
        if (!of_property_read_u32(np, "num-ports", &num_ports))
                num_par_io_ports = num_ports;