Right now, it doesn't work if I turn on these two functions at same time. The dnsmasq doesn't response client's dns request. Not sure if dnsmasq can receive dhcp request from different network interfaces simultaneously. Also, we need to modify net_worker.js to deal with the case when tethering is already enabled.

So I checked when USB Tethering is enabled and sys.usb.config is set to rndis,adb When USB Mass Storage is enabled, sys.usb.config is set to mass_storage,adb I took a look at init.qcom.usb.rc and I don't see any mode that just has rndis,mass_storage,adb so the question becomes is this possible?

FWIW. The UI on my Android sgs2 doesn't permit USB tethering and mass storage to be active concurrently

The wifi/usb hotspot support dhcp server function by using dnsmasq native daemon. When set dhcp server's ip range, it should always in the same subnet with interface ip. For instance, if we have wifi/3G interfaces with ip address set to 192.168.0.1/192.168.1.1, the IP range assigns to dnsmasq should be 192.168.0.2 ~ 192.168.0.254 / 192.168.1.2 ~ 192.168.1.254.

Comment on attachment 764688 [details] [diff] [review] Patch v1.0 Review of attachment 764688 [details] [diff] [review]: ----------------------------------------------------------------- One small comment. r=me with it answered. ::: dom/system/gonk/net_worker.js @@ +330,2 @@ > if (gCurrentCallback) { > + gCurrentCallback(isError(code), {code: code, reason: gReason.join(" ")}); Is there a strong reason to join the reason array? It seems like the only use would be as happy with an array as with a string.

There might be multiple provisional responses from netd for a request. For instance, if we request a tethering interface list, the netd might reply the interface name by several responses, and end with 2xx final response. We use array to gather these information. gecko netd interface list -----------> <--------------------- 1xx rmnet0 <--------------------- 1xx wlan0 <--------------------- 2xx rndis0

Nominate for leo+, since we don't have any warning notification in Gaia when users try to turn on wifi and usb hotspot simultaneously.

No new reason (or customer feedback) suggesting we should block on this in 1.1 but not for 1.0.1.

please apply this patch to Master git for LEO.

Push to birch .... https://hg.mozilla.org/projects/birch/rev/7345a2261e42

This patch is for b2g18 ... I have verified it on buri manually.

(In reply to leo.bugzilla.gecko from comment #9) > please apply this patch to Master git for LEO. Please provide a justification for blocking, as this is not regression in 1.1 and isn't milestoned.

(In reply to Alex Keybl [:akeybl] from comment #13) Please provide a justification for blocking, as this > is not regression in 1.1 and isn't milestoned. We decided not appling this patch. Thanks.