On 2019-11-19 9:09 AM, Uwe Kleine-König wrote:

On Tue, Nov 19, 2019 at 12:28:16AM +0200, Leonard Crestez wrote:

quoted

The clock drivers on imx8m series are registered as platform devices and
this opens the possibility of reloading the driver at runtime:

This doesn't actually work: clocks are never removed and attempting to
bind again results in registration errors and a crash.

Fix this by explicitly suppressing bind attrs like several other
drivers.

Signed-off-by: Leonard Crestez <redacted>

---
No cc: stable because because there are likely many other opportunities
to crash the system by echoing random stuff in sysfs as root.

  drivers/clk/imx/clk-imx8mm.c | 1 +
  drivers/clk/imx/clk-imx8mn.c | 1 +
  drivers/clk/imx/clk-imx8mq.c | 1 +
  3 files changed, 3 insertions(+)

diff --git a/drivers/clk/imx/clk-imx8mm.c b/drivers/clk/imx/clk-imx8mm.c
index 9246e89bb5fd..3cb75ad4270d 100644
--- a/drivers/clk/imx/clk-imx8mm.c
+++ b/drivers/clk/imx/clk-imx8mm.c

@@ -619,9 +619,10 @@ MODULE_DEVICE_TABLE(of, imx8mm_clk_of_match);
  
  static struct platform_driver imx8mm_clk_driver = {
  	.probe = imx8mm_clocks_probe,
  	.driver = {
  		.name = "imx8mm-ccm",
+		.suppress_bind_attrs = true,

Maybe add a comment similar to the motivation in the commit log here?
(And of course in the other files, too.)

Is it really useful to say "disable feature X because it doesn't work" 
right before disabling the feature?

--
Regards,
Leonard

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel