Hi Andy,
On Fri, Apr 3, 2020 at 6:05 PM Andy Shevchenko andriy.shevchenko@linux.intel.com wrote:
On Fri, Apr 03, 2020 at 02:58:08AM -0700, Bin Meng wrote:
Currently the driver gets ns16550 base address in the driver probe() routine, which may potentially break any ns16550 wrapper driver that does additional initialization before calling ns16550_serial_probe().
Things are complicated that we need consider ns16550 devices on both simple-bus and PCI bus. To fix the issue we move the base address assignment for simple-bus ns16550 device back to the ofdata_to_platdata(), and assign base address for PCI ns16550 device in ns16550_serial_probe().
This is still not perfect. Ideally if any PCI bus based ns16550 wrapper driver tries to access plat->base before calling probe(), it is subject to break.
Thank you. I have tested it and it fixes my case.
Tested-by: Andy Shevchenko andriy.shevchenko@linux.intel.com
Thanks for testing!
Fixes: 720f9e1fdb0c9 ("serial: ns16550: Move PCI access from ofdata_to_platdata() to probe()")
I believe this should be one line.
I thought putting it in one line violates the 80 characters per column rule. And checkpatch.pl does not complain about this, so I think it's fine :)
Reported-by: Andy Shevchenko andriy.shevchenko@linux.intel.com Signed-off-by: Bin Meng bmeng.cn@gmail.com
drivers/serial/ns16550.c | 51 +++++++++++++++++++++++------------------------- 1 file changed, 24 insertions(+), 27 deletions(-)
Regards, Bin