Not even with the patch it seems to work, there's still something odd (apart from the other issues it might have). I don't think we will ever see VLAN Support on the MR33, mainly because nobody will work on this anymore as better hardware becomes available.
What's some of the better hardware here, in your opinion?
Newer 802.11ax compatible devices that slowly start to become available. Don't get me wrong, I would love to finally get my MR33 out of the shelf and actually use it, but so far it's simply not possible and soon there will be the point where it gets outdated so it's no longer worth for most people to focus on it, so it gets more and more unlikely that someone will implement it (even though I was hoping that new driver would help here in this case).
any news on this ? 
There is work being done in "pre-master" to explore the ipqess and qca8k drivers and DSA system in general.
One commit on a staging tree (which can be though of as a developer's public "play space" for exploration, not as in confirmed that "this goes to master") that calls out the MR33 is
ipq40xx: convert Cisco Meraki MR33 to DSA
This patch converts the Cisco Meraki MR33 to use ipqess + qca8k's DSA.
This patch alone is not sufficient to build an image for the MR33 with these drivers.
I've used some of the patches on a different IPQ4019-based device with mixed results. I can't say that it is "ready for prime time" yet. There is a lot of motivation to get the drivers working smoothly (and submitted to upstream Linux, I believe). There is also the need to decide how to integrate then into UCI (swconfig is no longer the way that a DSA switch is managed) and implement that code in both the base system and in LuCI.
1 Like
thanks for the update
1 Like
Is there any update on this? (sorry for asking, i do not have a clue where to look for it)
Still WIP, literally just got an EA8300 on the bench and poking at it.
2 Likes
Thanks for the every fast response. Good to hear there is some progress. :).
@chunkeey I saw that you did some work on the MR33 recently (6 days ago if I can believe git), anything new in regards to this?
@jeff Not entirely sure how knowledge gained from the EA8300 would help here, as far as I understand the MR33 issue has to do with the special single port setup the device has. Maybe you could elaborate on that a little.
I’m not sure I can help much more as the DTS is very different than the EA8300, especially in the bitmasks for the ports.
I’d look into the (patched) driver, which seems to be different than in the EA8300 as well.
I haven’t checked the status of the DSA drivers recently, but early this year they didn’t support VLANs.
Hi, any news about this?.. i'd be nice working with vlans on MR33.
2 Likes
I wish I had found this thread earlier, could have saved many hours. I have submitted an update to the MR33 page to show this.
Firmware: OpenWrt 19.07.4 r11208-ce6496d796 / LuCI openwrt-19.07 branch git-20.310.50568-7b1b82d
Kernel: 4.14.195
At the very least the documentation on https://openwrt.org/toh/meraki/mr33 needs to be updated with VLANs listed in the "Unsupported Features" section.
1 Like
Does anyone have any new information about the issue with VLANs, especially with the MR33, or if there was any progress made in the last few months?
The changelog for 19.07.4 mentions a fix for VLANs on this target, is it related somehow?
ipq40xx: fix ethernet vlan double tagging (+193)
I'm on MR33 & stable OpenWrt v19.07.6 Changelog
I did a quick test adding eth0.x but could not get it to work, it receives some packages on the interface but does not send anything out. I have not tested with the trunk.
Looking at the git code changes, there are no specific changes for mr33 & ipq40xx besides some i2c that I'm not sure if they are related.
As AP is the main WiFi for my home network, I did not want to test more, but I have another one that is pending for boot loader downgrade and OpenWRT liberation, and I will have spare to do more tests if it's needed and play with it.
Cheers,
I wanted to use VLANs to have seperate networks for IoT, guest, etc. Are you doing this somehow in your network?
How are you downgrading the bootloader?
Thanks for the feedback!
Hi Sergio,
I do love the speeds of MR33, and the lack of VLAN is a bit downside but I can't rewrite or contribute to the code yet 
and I'm grateful to all developers that made OpenWRT possible on MR33 because it's still very reliable and does support multi SSID without VLANs.
For example, I run three SSID's WiD, KIDs, DEV on a flat network, KIDs is auto-down/up on schedule when it's time for a rest. I do not run any wifi IoT yet only a wired LAN bridge but have a few MR16,18,24 that do supports VLANs and pair well with my router OpenWRT/MX60 and most of my VLAN ports are directly connected to a MacMini as VM host and to a secondary AP for my garage(workshop).
For MR33 with multi SSID, it should be possible to divide the SSID across the three radio interfaces and add isolation with firewall rules, but it's a trade-off. Actually right now I do not use radio0 - Qualcomm Atheros QCA9887 802.11bgnac because when I did measure the wifi speeds it was performing better without it. But have not tested any isolation.
How are you downgrading the bootloader?
Most of the details are here
Cheers,
For VLAN to not work in MR33 is this driver specific or hardware config or both?
Looking here (mind this may be hardware-specific) ESS-SWITCH config and the MR33 config they look completely different.
The "IPQ4xx Ethernet Analysis" article ess is different and on different addresses:
ess-switch@c000000 {
compatible = "qcom,ess-switch";
reg = <0xc000000 0x80000>; /* 512KB */
switch_access_mode = "local bus";
resets = <&gcc ESS_RESET>, <&gcc ESS_MAC1_CLK_DIS>, \
<&gcc ESS_MAC2_CLK_DIS>, <&gcc ESS_MAC3_CLK_DIS>, \
<&gcc ESS_MAC4_CLK_DIS>, <&gcc ESS_MAC5_CLK_DIS>;
reset-names = "ess_rst","ess_mac1_clk_dis", \
"ess_mac2_clk_dis","ess_mac3_clk_dis", \
"ess_mac4_clk_dis", "ess_mac5_clk_dis";
clocks = <&gcc GCC_ESS_CLK>;
clock-names = "ess_clk";
switch_cpu_bmp = <0x1>; /* cpu port bitmap */
switch_lan_bmp = <0x1e>; /* lan port bitmap */
switch_wan_bmp = <0x20>; /* wan port bitmap */
};
MR33:
ess-switch@c000000 {
switch_mac_mode = <0x3>; /* mac mode for RGMII RMII */
switch_lan_bmp = <0x0>; /* lan port bitmap */
switch_wan_bmp = <0x10>; /* wan port bitmap */
};
I guess ess-switch config is not the same and depends on the PHY chip used in MR33 AR8035?
Edit: Disregard the ESS above it's for ADM6996L this is switch controller It is not single-port "dumb" PHY
Best regards,
If there is anything i can check / test / verify please let me know. I have 2 unlocked MR33 here to test on.
You could check what the current state is. Is incoming vlan tagged traffic working? Is outgoing vlan tagged traffic working?