Slashdot Mirror
Firmware Upgrades Creating Doorstops?
Michael Droettboom asks: "I recently purchased a Linksys WMLS11B. Nice enough unit for the money. One gotcha, though: when upgrading the firmware (past its refundable date where I bought it), my machine crashed, turning the WMLS11B into a doorstop along with it. Linksys has offered to replace the unit, but I don't want it if the box is so easily corrupted. I have always been on a strict policy of updating firmware on all my devices, but was wondering if anyone else has had experience with devices so broken as to not recover from a broken upgrade?"
If you're lucky there'll be a jtag-type port on the unit where you can reflash the memory.
:-)
If not, bring on the soldering iron and fix it properly.
If you could be told what you can see or read, then it follows that you could be told what to say or think - BoC
First: don't fix it if it isn't broken. Never upgrade the firmware unless your having problems. Never upgrade the firmware over a wireless link.
Next: Did you know that D-Link has a three year warranty on their stuff? I just found out today while RMA'ing a few 810+ bridges.
The above is not worth reading.
Don't bitch if the problem was cause by your end. How many devices can you begin a flash and then have your computer crash and still use the device?
__________
Love conquers all... except CANCER
This is consumer.. trash.
Most, if not all things made for consumers are made with this lack of quality in mind. Motherboards at least have a replaceable chip (unlike some dell models). Some high-ish quality X86 boards have 2 bioses that keep a known good copy on second chip.
You must be aware that most devives with updateable firmware also suffer this.. problem. Some older graphics cards actually had removable Bioses, as well did scsi cards.
Once device I know that does not have this problem is the stable hardware on teh Compaq Proliant servers. There's extremly stable text mode drivers along with keyboard imput and ide drivers. You update the upgradeable bioses on the hard drive. Something causes problems? Reload from cd and redo the "bios" partition.
Lucent USB wifi modules also do something similar... The driver for the host computer updates "onload" the firmware. A firmware updfate on those are as simple as telling the driver to upload this file instead of this other file.
I was working on my car the other day, changing some things in the engine, but got distracted by a phone call and lost the parts I was changing. Now my car's engine doesn't work. The car dealership has offered to fix it if I tow it down there, but I don't really want a car that can be broken so easily. Shouldn't cars be less difficult to break?
happened to me with my linksys router. I decided well, I've been wanting wifi anyway...
must... stay... awake...
Many vendors say clearly don't upgrade if the router, mobo, or whatever else you have is currently working properly for you. Sometimes new BIOS revisions are specifically created for new hardware revisions that are slipstreamed in, and they don't always properly document that.
Unless you've read in the changelog that it patches a hole/bad code or adds functionality that you need, don't take the chance that it's going to crater. Even if your vendor is nice enough to replace the component out of warranty, you're still going to be inconvenienced waiting.
Get off my launchpad!
This is why I use Gigabyte Motherboards almost exclusively, with their dual-BIOS onboard. If the machine crashes during a flash, all I need to do is reboot and it'll fire up on the backup BIOS and I'm in the running again.
Really though, I wouldn't bitch about the device not being able to recover itself. Adding on self-healing abilities (Basically fixing end user screw ups) adds parts to the device, which adds to the cost. It is a *consumer* level device, remember, so price is the big concern. Stop whining and get a new one - you screwed it up, the company isn't to blame here.
Looking for hardware (Currently need: Large Etch-a-Sketch) Have one? See my journal!
Let's get this straight: as it stands, your device is broken and the manufacturer is offering to replace it with another unit and you're not happy with that situation?
Just what do you think would constitute good customer service on their part if replacing a dead unit with a working one isn't to your satisfaction? Just what do you want out of Linksys? Blood?
If Linksys told you to go take a running jump and were of no help whatsoever I'd understand you having a beef with them but they've done what they should do in this situation and yet you're still not happy.
There's no pleasing some people.
"Accept that some days you are the pigeon, and some days you are the statue." - David Brent, Wernham Hogg
1. What feature did you NEED, when you flashed the firmware.
2. Why didn't you buy hardware that had that feature to begin with.
I gave my original first model Linksys wi-fi 4port dsl router to a friend. In helping him set it up, I flashed it for him since it was done in years and lots of useful changes have been made. So 1 day later, another friend comes over to help him with this computer setup again. He insists to flash the Linksys even though my friend insisted that I had already done it. "Oh it won't hurt to do it again!" And he promptly flashed the firmware from a different (but similarly spelled name) model. Instant brick! My friend, who lost his free linksys, ended up having to shell out at the store for a replacement. Oh well...
Use PC's with multiple ethernet cards, and wireless cards that have real open source drivers even if they are more expensive, and run real operating systems. The vast amount of stuff that has built up around the x86 PC architecture, from boot floppies to boot ethernet ROMs to replacing BIOS chips, simply can't be beat.
I recently noticed that MSI had released a new version of the BIOS for my motherboard, and I decided it can't hurt to upgrade. I flashed the BIOS from Windows (the other option would be a DOS floppy, but my box does not have a floppy drive). There were no errors untill I rebooted into Windows, then everything went wrong. As soon as there would be any significant hdd activity, the computer would just freeze, and it would reboot a few seconds later.
I decided to roll back to the previous version of the BIOS, but it turns out upgrading is much easier. Apparently the MSI Windows utility will only allow a newer version of the BIOS to be installed, but not old ones. Since I could not use the fdd method, I had to trick the utility into using the old BIOS file. Everything worked (thankfully) in the end and all I lost was time. But I could have needed a new motherboard...
It is MSI's fault that the new BIOS is extremely unstable, but it would not be their responsability if my motherboard got fried while flashing the BIOS (or after). It was my choice to flash it, even though they recommend against it if everything works properly.
In your case, just get the replacement unit and try not to complain. I think it's more than a fair deal Linksys is willing to give you.
My experience with Linksys routers says "buy from them next time because you know their stuff works." Now you add to that, "they'll replace the hardware even if the screwup wasn't their fault." Sounds pretty positive to me.
I love IBM hardware, its always been good to me, but I've got a related story.
an x225 that was less than 2 weeks old had a motherboard failure, IBM sent a tech out who replaced the motherboard, all nice and good. However the last thing the tech did was flash the bios. Its the last thing the motherboard ever did too. The replacement motherboard was toast. (And believe it or not, it was the last x225 motherboard in the country at the time).
Either way, apparently the flash util erased the bios, but didn't write a new one. Now according to IBM there are procedures in place to stop that from happening and if it does happen, to recover, but none of that worked.
Afaik, when you first burn eeproms, you can mask off a region that can't be reburnt, thus allowing you to have recovery, but I'm guessing this didn't work on this system.
Either way, EIGHT days after the motherboard died, we had a replacement server. I must say though, this has been the only bad experience I've had with IBM support.
I use to have a funny sig, but slash cut it off, and I forgot what the punchline was.
If you can't link to it, then post the information here to this site. If you can't link to it, then just repost it here.
I have a linksys BEFSR81v3 router.
I updated the firmware, and afterwards, Half-Life based games would not work. I could play for up to 5 minutes, then it would halt with an "invalid packet" message of some sort.
I searched around online, and found other people with similiar problems, and forums saying that Linksys was aware of the problem, it had to do with fragmentation of large packets.
The real problem was that their web/FTP did not have any copies of the previous firmware, so I couldn't revert.
I called up tech support. Several times. Usually, the people I was talking to had strong accents that made them hard to understand. On top of that, when I called and asked for the previous firmware, they were basically going to the FTP site themselves, and emailing me a file (that I could have easily downloaded, and was not the correct file).
I started off polite, the first few phone conversations. Several phone conversations (and two days after I first called), this is what my side of the conversation sounded like:
"I have VERSION 3 of the BEFSR81, do you understand? Yes, I have been to your FTP site, and looked all over it. I do not want the firmware for version 2 of the router. If you read its text file, it SPECIFICALLY SAYS IT WILL NOT WORK WITH VERSION 3."
Several hours later, what do I get, but yet another copy of version 2 firmware that won't work on my router.
I don't know about Linksys but there should be an option on how to erase the corrupted firmware. On a D-Link router you just insert a paperclip into the reset hole, take out the power adapter and insert it back again, then take out the paper clip, do this real slow and log onto the router. It won't ask for a username/password and you will only have 1 page to update the firmware, everything would have been erased. These devices are really hard to kill completely.
The way to have safe flashing is to use an EEPROM part that is (1) large enough to hold two copies of the firmware, and (2) has a "bootblock" that cannot be flashed in the field. The bootblock gets control on startup, validates the firmware copies, decides which one to load, and goes from there. This kind of redundency costs money. Cisco will be happy to sell you a router that can safely be flashed, but it will probably cost an order of magnatude more than anything from Linksys. You get what you pay for.
I honestly don't see why firmware upgrades should leave devices as bricks - at the very least, a non-overridable "boot rom" (which normally hands over to the firmware, but can instead run a rescue program in the event of mishap)
Supermicro motherboards had something like this for years. If your BIOS is corrupt you put in a floppy with a new image and hold down some special key combo to activate the emergency flash feature.
Only had to use it once, on a test system, saved me the time of having to find another mb.
This is a flaw with the hardware design, and not with the consumer. Crashes during a firmware flash can happen, and they should be designed around. Don't blame bad design on a user!
Example: All Compaq desktop and server machines have had a feature sinced at least 1990 that allows you to recover from a bad firmware flash. Not too many folks new about it. Basically, you could throw all the dip switches on SWB6 on, and the machine could boot to read firmware from a floppy. The prompt was beeps, but it worked. Problem was, not even a lot of the tech support folks new about it.
Now, the trend is towards having 2 ROM images. Flash goes to new image, if new image can't boot, old image boots.
The point is, this problem is 1) well known, and 2) solutions are well known.
We have a tendancy to excuse these sorts of things, but the moral to the story is to NEVER blame problems that stem from bad design on the end user.
Equipment Exceeds Owners Capabilities
I flashed the bios of my Netgear wireless acess point, and it appeared to have worked sucessfully. However when I tried to access it to look at the security, I couldn't. The damn thing committed suicide.
Check out this page on recovery procedures.