Closed Bug 847388 Opened 12 years ago Closed 7 years ago

DHCP broken on Nexus S when WiFi Tethering enabled

Categories

(Firefox OS Graveyard :: Wifi, defect)

ARM
Gonk (Firefox OS)
defect
Not set
major

Tracking

(tracking-b2g:backlog)

RESOLVED WONTFIX
tracking-b2g backlog

People

(Reporter: gerard-majax, Unassigned)

References

Details

Attachments

(3 files, 1 obsolete file)

Basically, DHCP is up and running, but badly configured system, thus: I/dnsmasq ( 4122): started, version 2.51 cachesize 150 I/dnsmasq ( 4122): compile time options: no-IPv6 GNU-getopt no-DBus no-I18N DHCP no-scripts no-TFTP W/dnsmasq ( 4122): warning: no upstream servers configured I/dnsmasq ( 4122): DHCP, IP range 192.168.1.10 -- 192.168.1.30, lease time 1h I/dnsmasq ( 4122): read /etc/hosts - 1 addresses I/dnsmasq ( 4122): using nameserver 212.27.40.241#53 I/dnsmasq ( 4122): using nameserver 212.27.40.240#53 W/dnsmasq ( 4122): DHCP packet received on wl0.1 which has no address W/dnsmasq ( 4122): DHCP packet received on wl0.1 which has no address W/dnsmasq ( 4122): DHCP packet received on wl0.1 which has no address W/dnsmasq ( 4122): DHCP packet received on wl0.1 which has no address W/dnsmasq ( 4122): DHCP packet received on wl0.1 which has no address W/dnsmasq ( 4122): DHCP packet received on wl0.1 which has no address
http://lists.thekelleys.org.uk/pipermail/dnsmasq-discuss/2011q3/005165.html « On 2011-07-22 23:05, Michael Fork wrote: > Is it possible to use dnsmasq to serve DHCP on an interface with no IP Address? The answer is right there: http://tools.ietf.org/html/rfc2131 To save you some time: no, it is not possible. »
Assignee: nobody → lissyx+mozillians
Attached a patch that fixes the issue by using the wl0.1 network interface instead of wlan0 for both DHCP and NAT operations.
Attachment #720698 - Flags: review?(vchang)
This is for real!
Depends on: 847157
Comment on attachment 720698 [details] [diff] [review] Using wl0.1 instead of wlan0 for DHCP and NAT fixes the issue Review of attachment 720698 [details] [diff] [review]: ----------------------------------------------------------------- This patch looks wrong to me because some devices such as unagi and otoro don't have the wl0.1 interface. If the problem is related to name of wifi interface, we should better to detect it automatically.
Attachment #720698 - Flags: review?(vchang)
Blocks: b2g-nexuss
blocking-b2g: --- → backlog
For some reason, the driver and/or netd are bringing up the network on wl0.1 when we expect this to be on wlan0 interface. Let's then use wl0.1 for both handling the DHCP request and for the NAT. This will allow to use WiFi Tethering properly. One consequence of this workaround is that you need to either manually unload and reload the WiFi kernel driver or reboot the device after using WiFi Tethering, otherwise the stack ends in an unexpected state.
Comment on attachment 720698 [details] [diff] [review] Using wl0.1 instead of wlan0 for DHCP and NAT fixes the issue This patch is out-dated.
Attachment #720698 - Attachment is obsolete: true
FYI, the problem of having to reload the driver after disabling tethering can be mitigated by setting the ro.moz.wifi.unloaddriver property to 1. For the wl0.1 naming issue, I could not find anything explicitely explaining what happens: we set the device to be an access point via the iw kernel interface, and everything seems to be done kernel-side.
blocking-b2g: backlog → ---
Unassigning as I'm stopping dogfooding Nexus S.
Assignee: lissyx+mozillians → nobody
For some reason, the driver and/or netd are bringing up the network on wl0.1 when we expect this to be on wlan0 interface. Let's then use wl0.1 for both handling the DHCP request and for the NAT. This will allow to use WiFi Tethering properly. One consequence of this workaround is that you need to either manually unload and reload the WiFi kernel driver or reboot the device after using WiFi Tethering, otherwise the stack ends in an unexpected state.
Firefox OS is not being worked on
Status: NEW → RESOLVED
Closed: 7 years ago
Resolution: --- → WONTFIX
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: