Merge branch 'sh_eth-fix-reboot-crash'

Geert Uytterhoeven says:

====================
sh_eth: Fix reboot crash

This patch fixes a regression v5.11-rc1, where rebooting while a sh_eth
device is not opened will cause a crash.

Changes compared to v1:
  - Export mdiobb_{read,write}(),
  - Call mdiobb_{read,write}() now they are exported,
  - Use mii_bus.parent to avoid bb_info.dev copy,
  - Drop RFC state.

Alternatively, mdio-bitbang could provide Runtime PM-aware wrappers
itself, and use them either manually (through a new parameter to
alloc_mdio_bitbang(), or a new alloc_mdio_bitbang_*() function), or
automatically (e.g. if pm_runtime_enabled() returns true).  Note that
the latter requires a "struct device *" parameter to operate on.
Currently there are only two drivers that call alloc_mdio_bitbang() and
use Runtime PM: the Renesas sh_eth and ravb drivers.  This series fixes
the former, while the latter is not affected (it keeps the device
powered all the time between driver probe and driver unbind, and
changing that seems to be non-trivial).
====================

Link: https://lore.kernel.org/r/20210118150656.796584-1-geert+renesas@glider.be
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

1 files changed, 3 insertions, 0 deletions

diff --git a/include/linux/mdio-bitbang.h b/include/linux/mdio-bitbang.h
index 5d71e8a8500f..aca4dc037b70 100644
--- a/include/linux/mdio-bitbang.h
+++ b/include/linux/mdio-bitbang.h
@@ -35,6 +35,9 @@ struct mdiobb_ctrl {
 	const struct mdiobb_ops *ops;
 };
 
+int mdiobb_read(struct mii_bus *bus, int phy, int reg);
+int mdiobb_write(struct mii_bus *bus, int phy, int reg, u16 val);
+
 /* The returned bus is not yet registered with the phy layer. */
 struct mii_bus *alloc_mdio_bitbang(struct mdiobb_ctrl *ctrl);