306191 - Firefox 1.0.6 Does not properly handle firmware upgrading procedures on a Netgear MR814v2 Firmware Rev. 5.00

Status: RESOLVED WONTFIX

(Firefox :: General, defect)

Description

[KhanReaper]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.2; en-US; rv:1.8b3) Gecko/20050712 Firefox/1.0+
Build Identifier: 

While upgrading the firmware on an old Netgear MR814v2 (then firmware version
5.00) to 5.30, Gentoo's Firefox could not properly communicate with the router's
HTML interface, resulting in a potential firmware upgrade disaster. I have
Javascript enabled on the browser, so it is not as if I tried to prevent Firefox
from responding to what the router requested. In any case, I booted another
machine (this one) into Windows and used the 64-bit version of I.E. to upload
the firmware and the procedure went perfectly.

While in Linux, I tried to upgrade the firmware multiple times, and after the
file appeared to upload but the router never upgraded the data, I decided to use
ngrep, tcpdump, iptraf, and a few other utilities just to see what was and was
not sent.

The new firmware (5.30) is less than a megabyte in size, and I saw that it was
sent perfectly to the router. After it was sent, Firefox just hung there with
some message in its bottom status bar claiming that it was waiting for a
response of some kind.

In any case, Firefox just hangs when logged-in to the router
(http://192.168.0.1) and after uploading the new firmware. It just does not
respond to whatever requests the router sends to it or the javascript used on
the router.

I regret now that I did not copy the tcpdump and ngrep output so it could be
analyzed by the Mozilla team.

Reproducible: Always

Steps to Reproduce:
1. Connect to a Netgear MR814v2 router (firmware 5.0) and go to the firmware
upgrade page.
2. Attempt to upgrade the firmware with version 5.0 firmware
(http://kbserver.netgear.com/release_notes/D102358.asp)
3. Wait

Actual Results:  
The firmware actually is uploaded via a file-upload form request, but Firefox
just hangs after it uploads, never to respond again, leaving the user completely
uncertain whether the upgrade was successful or failed.

Expected Results:  
The firmware should upload immediately, and the user should be redirected to a
progress page that illustrates the progress of the firmware migration.
Immediately thereafter, the user should be prompted about any post-upgrade steps
that need to be taken.

This bug may be hard to replicate, but it is bad and needs further examination,
because I dare not imagine the confusion and anger that a faulty-upgrade would
cause. Interestingly, it is not as if Firefox has always had errors when
upgrading the firmware on an MR814v2 router; in fact, I have upgraded firmware
on them several times without error.

Comment 1

[Michael (GoodWill)]

I had this tested with my friends MR814. It worked.

If I had to guess the html handling capabilities on the router are limitted at best.

I would mark this a WONTFIX.

Comment 2

[Michael (GoodWill)]

The best guess here is that router firmware is not standards compliant. In fact
all the netgear series based on the same chipset as MR814 have been reported as
having all kinds of horrific problems. 

Comment 3

[KhanReaper]

(In reply to comment #1 and #2)
While it certainly would not surprise me if the router uses some incompliant
form of HTML, HTTP, and Javascript, I am still slightly doubtful that such a
thing could have caused this entirely due to a recollection of what I was doing
with the browser earlier in the day.

Long before I upgraded the router's firmware, I ran a beta of Deer Park alpha 2
under the same user account, and I believe that I set some Javascript option to
disable common annoyances. I have to admit that the phrase common annoyances is
ambiguous, so I would like to know exactly what functionality differences it
entails. Would this option have altered behavior that could be saved (i.e., made
persistent) in ~/.mozilla/* and affected an older version of Firefox (1.0.6) if
it executed sometime thereafter? I only ask this, because I would assume that
this common annoyances feature would modify a few settings in about:config that
would be made persistent in user.js or something to that effect.

Is there then any potential in this for the common annoyances feature to have
affected the firmware upgrade? None of this should excuse bad coding work on the
part of Netgear, yet the common annoyances feature ought not to affect web
browsing adversely.

I would appreciate some commentary on this.