-----Original Message-----
From: Jakub Kicinski <kuba@kernel.org>
Sent: Friday, August 05, 2022 12:59 PM
To: Keller, Jacob E <jacob.e.keller@intel.com>
Cc: Jiri Pirko <jiri@resnulli.us>; netdev@vger.kernel.org; davem@davemloft.net;
idosch@nvidia.com; petrm@nvidia.com; pabeni@redhat.com;
edumazet@google.com; mlxsw@nvidia.com; saeedm@nvidia.com;
tariqt@nvidia.com; leon@kernel.org; moshe@nvidia.com
Subject: Re: [patch net-next 2/4] net: devlink: convert reload command to take
implicit devlink->lock

On Fri, 5 Aug 2022 16:21:16 +0000 Keller, Jacob E wrote:

quoted

quoted

Convert reload command to behave the same way as the rest of the
commands and let if be called with devlink->lock held. Remove the
temporary devl_lock taking from drivers. As the DEVLINK_NL_FLAG_NO_LOCK
flag is no longer used, remove it alongside.

Signed-off-by: Jiri Pirko <redacted>

Wasn't reload avoiding the lock done so that drivers could perform other

devlink operations during reload? Or is that no longer a problem with the recent
refactors done to move devlink initialization and such earlier so that its now safe?

Drivers have control over the lock now, they can take the lock
themselves (devl_lock()) and call the "I'm already holding the lock"
functions. So the expectation is that shared paths used by probe and
reload will always run under the lock.

Ah perfect. So essentially, they already take the lock and are calling the "i'm locked" version of the functions, but this now moves the take the lock part into core.

Ok great! That sounds better than assuming drivers will take the lock themselves.

Thanks,
Jake