Opened 2 years ago

Closed 2 years ago

#21376 closed defect (wontfix)

BCM53115 phy driver quite incomplete

Reported by: ehem+openwrt@… Owned by: developers
Priority: normal Milestone:
Component: kernel Version: Trunk
Keywords: BCM53115 phy kernel switch Cc:


The b53 driver for Broadcom switch chips is missing a lot of features. It appears the driver can reset the "MIB" counters for the whole switch, but cannot reset them for individual ports. This setting is also misnamed, "reset_mib" for the whole switch, rather than "reset_mibs".

The BCM53115 switch is documented to have support for double-tagging, spanning tree, QoS, and ingress/egress rate control. Yet none of these is implemented in OpenWRT's kernel.

I'll concede I don't know whether Broadcom has properly documented these functions, but I certainly hope they have. These at least shouldn't require a binary blob.

Attachments (0)

Change History (3)

comment:1 Changed 2 years ago by nbd

  • Resolution set to not_a_bug
  • Status changed from new to closed

Feel free to submit patches, but keeping this kind of bug report open is not particularly useful to anybody.

comment:2 Changed 2 years ago by ehem+openwrt@…

  • Resolution not_a_bug deleted
  • Status changed from closed to reopened

My hope was if nothing else this report would serve as a gathering point. Generally the expected situation for this circumstance is to mark "wontfix" and leave it open waiting for someone to take action.

Producing patches isn't helped by both Broadcom and OpenWRT lacking documentation.

I'm primarily interested in double-tagging, but am having difficulty finding any documentation (or even any implementation!). The information on generic single-tagging is reasonably out there. What documentation generated the original RoboSwitch module?

I've got no timeframe, but who knows, maybe patches might appear though I've got weird aims compared to OpenWRT (I really hope Debian targets an official "stable" kernel next release).

comment:3 Changed 2 years ago by nbd

  • Resolution set to wontfix
  • Status changed from reopened to closed

It's already hard enough as it is to keep track of open bug reports, since we have thousands of open tickets already. Please don't insist on making the problem worse by reopening a ticket that just *might* be useful to somebody someday.

It is my experience that people who can actually help with implementing such patches generally do not go around looking for open tickets to see if somebody might want them to implement something specific.

And if by chance somebody does, to a search engine a closed ticket is just as good as an open one, so please leave this one closed.

Add Comment

Modify Ticket

as closed .
The resolution will be deleted. Next status will be 'reopened'.

E-mail address and user name can be saved in the Preferences.

Note: See TracTickets for help on using tickets.