On 17 July 2017 at 05:53, Rafa=C5=82 Mi=C5=82ecki [off-list ref] wrote:

On 16 July 2017 at 13:21, Ian Molton [off-list ref] wrote:

quoted

I've been doing some cleanups in the broadcome brcmfmac driver, trying t=

o

quoted

understand how it works.

I think this makes the driver a *lot* more readable.

Of note:

* Gets rid of the arbitrary and completely 1:1 mapping of
 struct brcmf_{core,chip}_priv/struct brcmf_{core,chip} structures
 which add unreadability whilst offering no real seperation.

* The patch titled HACK - stabilise the value of ->sbwad in use
 highlights an issue that is either bizarre undocumented behaviour or a
 genuine bug - without datasheets, I dont know. Essentially the value of=

 sbwad

quoted

 is taken as the base address in several functions, even though it is su=

bject

quoted

 to change should other IO occur that moves the window offset

Obviously this is a first cut at this and needs substantial cleanup itse=

lf,

quoted

however I wanted to get some idea of wether I should continue working on=

 this.

I looked at 4 random patches and none got any description. Not to
mention their chaotic subjects. In this state I can't even review it.
If you want to have some change accepted, you've to convince us it's
needed. Work on cleaning your patches and resend them. You also need
to signed off your changes.

--
Rafa=C5=82

As someone who is interested in any bug fixes to this driver (Device
is used on Raspberry Pi3/0W and we have a number of issues reported
which we are actively investigating), it would be very useful to more
clearly split out any actual fixes vs simply tidying up (Yes, I agree
the driver is mostly incomprehensible). Perhaps asking the
list/maintainers for comments on any located issues/bugs fixes would
be a useful starting point, along with ensuring the description gives
a good explanation of what the suspect issue is.

James