Re: [PATCH] Newbus locking

* Hans Petter Selasky <hselasky_at_c2i.net> wrote:
> I was hoping that the Giant requirement for newbus would go away, but all that 
> has been done is to replace Giant with an sx-lock protecting all access and 
> calls into any device_xxx() method. This is a step backwards.

Gradually decomposing Giant should always be considered a step forward.
For example, I am sure people will have dozens of comments on how I
could have improved locking inside the TTY layer, but it's still better
than what we had.

<nitpick>
Attilio, I'm too stupid to understand Newbus, but I spotted a small typo
in acpi_battery.c: Tollerate -> Tolerate. There's also a small
whitespace regression in the first chunk of the same file.
</nitpick>

-- 
 Ed Schouten <ed_at_80386.nl>
 WWW: http://80386.nl/