Increase broadcast and unicast table sizes

[puddly]

• The broadcast table size increase should help with NETWORK_BUSY errors. This is the table EmberZNet uses to determine if there are too many broadcasts happening at once.
• The unicast table should help enqueue more packets.

[puddly]

For ease of use, here are firmwares for some common sticks:

• skyconnect_ncp-uart-hw_7.4.2.0.gbl.zip
• yellow_ncp-uart-hw_7.4.2.0.gbl.zip
• sonoff-e_ncp-uart-hw_7.4.2.0.gbl.zip

GitHub does not allow .gbl files so I have zipped them. Please unzip them and flash them with an appropriate tool:

• SkyConnect: https://skyconnect.home-assistant.io/firmware-update/
• Sonoff: https://darkxst.github.io/silabs-firmware-builder/
• Yellow: not possible just yet without manually flashing using universal-silabs-flasher.

[puddly]

@jclendineng Let's move firmware discussions over here. If your firmware is showing up as 7.4.0.0, it didn't flash. It'll show up as Firmware: 7.4.2.0 build 0.  

[puddly]

Feedback has been positive so far. Merging.

[evelant]

Per home-assistant/core#86411 this is a massive improvement but seems to have introduced an issue with group behavior. To summarize from that thread:

Random members of groups don't respond to group commands. Which members of the group fail to respond is random with each command sent. This never happened prior to this build. Groups addressed via a binding on an inovelli blue switch always respond in unison. Only commands to groups from the coordinator since this build exhibit the random failure. I'm not sure how the changes made in this build could have introduced this but I am certain this issue was introduced in this build.

[jclendineng]

I agree; after installing the latest build it started getting wonky again. Now that could also just be coincidence. On Jul 4, 2024, at 10:04 AM, Andrew M ***@***.***> wrote: Per home-assistant/core#86411 this is a massive improvement but seems to have introduced an issue with group behavior. To summarize from that thread: Random members of groups don't respond to group commands. Which members of the group fail to respond is random with each command sent. This never happened prior to this build. Groups addressed via a binding on an inovelli blue switch always respond in unison. Only commands to groups from the coordinator since this build exhibit the random failure. I'm not sure how the changes made in this build could have introduced this but I am certain this issue was introduced in this build. —Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you were mentioned.Message ID: ***@***.***>