Can firmware cause corrupt downloads in Chrome?

[Tabs]

Hi all,

I have a very strange issue that I'm trying to troubleshoot and I'm wondering if it could somehow be related to my RT-N56U:

For several weeks now I've been getting a lot of downloads that are corrupt even though they appear to finish properly in the browser. I can download the same file with Firefox or IE and there are no issues however.

I've tried reinstalling Chrome completely clean with no extensions or anything - same exact issue.

I'm starting to wonder if the router could be involved somehow - I'm running the latest Padavan currently and I have an Asus PCE-N13 wifi card. I've just been installing the new Padavan versions as they come out over top of what was there. Is this correct or is there some other procedure I should be using? If I wanted to go back to stock Asus firmware to test, how would I go about doing it?

Thanks!

 

[RogerSC]

For Chrome, have you tried setting aside your profile, and letting it create a new one? This seems odd, that one browser would have corrupt downloads, and the others wouldn't, but one way of troubleshooting this is to eliminate your profile as part of the problem.

My user profile is here (Windows 8):

C:\Users\<user name>\AppData\Local\Google\Chrome\User Data\Default

If I were to move this to "Default.old", I believe that Chrome would create a new "Default" that would start clean.

 

[TonyH]

Hi,
You download using WiFi since you mentioned the card?

 

[Tabs]

Hi,
You download using WiFi since you mentioned the card?

Yes, the desktop is WiFi and has worked fine up until this problem started recently. 5/5 "excellent" signal on the meter in the notification area.

 

[Tabs]

For Chrome, have you tried setting aside your profile, and letting it create a new one? This seems odd, that one browser would have corrupt downloads, and the others wouldn't, but one way of troubleshooting this is to eliminate your profile as part of the problem.

My user profile is here (Windows 8):

C:\Users\<user name>\AppData\Local\Google\Chrome\User Data\Default

If I were to move this to "Default.old", I believe that Chrome would create a new "Default" that would start clean.

Yeah when I did the clean install of Chrome I totally wiped everything out and even tried using it without signing into my Google Account - problem was still there.

 

[Shikami]

Anti-virus, storage cable connections, power, overclocking, bad ram, etc. Many of these can cause corruption with data.

Look in Event Viewer (if using Windows) for errors, check cables, and test your ram. That is where I would start.

Other devices are fine correct?

 

[Tabs]

Anti-virus, storage cable connections, power, overclocking, bad ram, etc. Many of these can cause corruption with data.

Look in Event Viewer (if using Windows) for errors, check cables, and test your ram. That is where I would start.

Other devices are fine correct?

Yeah - everything else on the system is fine except for this. I can't imagine bad cables would let everything else work fine. (including the exact same downloads in FF and IE) I did try completely uninstalling my antivirus (MSSE) to test that too - no difference. RAM I'll run Memtest on but again - no other issues, I can do things that push RAM hard like games, audio editing etc all day long without issue.

 

[Guz]

Sounds like you found a bug in Chrome.

You've eliminated the network stack by using other programs, but it stays with Chrome, then it's most likely Chrome. Uninstall it, clean out any remnants by hand, reboot, reinstall.

 

[Shikami]

Yeah - everything else on the system is fine except for this. I can't imagine bad cables would let everything else work fine. (including the exact same downloads in FF and IE) I did try completely uninstalling my antivirus (MSSE) to test that too - no difference. RAM I'll run Memtest on but again - no other issues, I can do things that push RAM hard like games, audio editing etc all day long without issue.

A power plug does not need to be in all the way to pass on electricity either. Basically, due to signals, and also the ability to correct errors (hopefully), you can have "function." The thing about testing, is not all "testers" are the same, nor do they all test very well. They can usually test well enough to find obvious errors.

This is one of the reasons to look in Event Viewer when using Windows. One new motherboard I purchased a bit ago functioned however it would come up with error events in the Viewer as having an issue with the controller. However, to my observation things seemed nominal-that was sent back, obviously!

Also, not all programs and such use the same regions of memory. Therefore, you can have many that work fine but one that does not. Testing memory is a bit difficult too. You may need to run multiple passes with something such as Memtest. Up to 10 before errors will show, due to circumstances that may not happen due to something such as heat affecting the memory banks.

Troubleshooting is not difficult; it is more difficult to not be making assumptions when troubleshooting.

 

[Gingernut]

What OS version are you on and what Ralink driver version are you using for the Asus PCE-N13?

 

[Tabs]

What OS version are you on and what Ralink driver version are you using for the Asus PCE-N13?

Win7 x64 SP1, 5.0.21.0 dated 2/18/2013 on the WiFi card driver. Is there a better one?

Also - if I did want to try stock ASUS firmware on the RT-N56U, how do I go about doing that after having the Padavan installed forever? Is there a clean re-flash procedure at all that ensures I'm starting back like it was out of the box?

 

[Guz]

To go back to Asus,
Set your computer to a static IP (example: 192.168.1.5)

Remove any/all USB devices. Take the configuration back to factory, then reflash with desired Asus firmware. When it's done (only after it's all done and happy and you can connect to it via a browser). Power cycle the router (pull the plug, count to 20, plug back in). When it's rebooted, again set it back to factory. Then configure it how you want.

Take your computer back to a DHCP assigned IP address.

You should be good to go after that (emphasis "should").

 

[Tabs]

To go back to Asus,
Set your computer to a static IP (example: 192.168.1.5)

Remove any/all USB devices. Take the configuration back to factory, then reflash with desired Asus firmware. When it's done (only after it's all done and happy and you can connect to it via a browser). Power cycle the router (pull the plug, count to 20, plug back in). When it's rebooted, again set it back to factory. Then configure it how you want.

Take your computer back to a DHCP assigned IP address.

You should be good to go after that (emphasis "should").

Thanks, I'll give it a shot. What's the current recommended stock firmware?

 

[611]

Win7 x64 SP1, 5.0.21.0 dated 2/18/2013 on the WiFi card driver. Is there a better one?

It's the Ralink driver causing your problems - I've had the same problems with the same driver on the same OS. Data corruption was seen in Opera, IE, and TrackChecker, so it's not a bug in Chrome, it's neither related to RT-N56U as I'm using Zyxel Keenetic Giga.

Rolling back the driver (in my case the previous version is 3.2.12.0) immediately solved the problem. It looks like the faulty version of driver was installed by Windows Update in the middle of March as optional update.

 

[Gingernut]

It's the Ralink driver causing your problems - I've had the same problems with the same driver on the same OS. Data corruption was seen in Opera, IE, and TrackChecker, so it's not a bug in Chrome, it's neither related to RT-N56U as I'm using Zyxel Keenetic Giga.

Rolling back the driver (in my case the previous version is 3.2.12.0) immediately solved the problem. It looks like the faulty version of driver was installed by Windows Update in the middle of March as optional update.

Exactly why I asked, this problem drove me nuts for about two days, had to roll back to Ralink's drivers posted on their site, not the ones from Station Drivers or the like, to solve the problem.

From what I've tested the corruption bug only affects the newer Win7 x64 Ralink drivers.

 

[Tabs]

I rolled back the driver on the card and it now shows version 3.2.1.0 (which I believe is the Asus website driver). I looked on the Ralink/MediaTek site for a 3.2.12.0, but I don't see that there. Can someone point me to the proper link?

 

[Tabs]

Nevermind, found it here:

http://www.wireless-driver.com/ralink-rt2860-rt2760-series-windows-driver-ver3220/#down_link

Installed it and I'm going to spend some time testing with it.

If I were to just get a new PCIe Wireless card, what's the best and most trouble free one? I see Intel has a couple - have to imagine those are more robust/stable than something that's using a bunch of other sub-company's parts and drivers like the PCE-N13 is (Pegatron manufactured, Ralink chip and driver etc) right?

 

[Gingernut]

Latest official release made by Ralink can be downloaded here:

http://www.mediatek.com/_en/07_downloads/01_windows.php?sn=500

The Asus PCE-N13 uses a Ralink RT2790 chip.

Ralink was absorbed by Mediatek a while back so downloads are now from them.

AFAIK Intel only make mini-pci wireless cards.

 

[611]

From what I've tested the corruption bug only affects the newer Win7 x64 Ralink drivers.

What a shame for Ralink and Microsoft.

How it happens the faulty driver, newer than the one offered on the manufacturer's site, was distributed via Windows Update?!

And what about TCP/UDP checksums?
Network card driver resides between layers 2 and 3 on OSI model, below OS's IP stack. If the stack checks the packets' checksums, it should request retransmission of corrupted packets, not pass corrupted data to application. So, if everything's done right we would only notice some degradation of performance, not corrupted data.
Maybe Ralink drivers are telling stack they're handling TCP/UDP checksum offload, while actually they're not checking anything?

 

[RogerSC]

What a shame for Ralink and Microsoft.

How it happens the faulty driver, newer than the one offered on the manufacturer's site, was distributed via Windows Update?!

And what about TCP/UDP checksums?
Network card driver resides between layers 2 and 3 on OSI model, below OS's IP stack. If the stack checks the packets' checksums, it should request retransmission of corrupted packets, not pass corrupted data to application. So, if everything's done right we would only notice some degradation of performance, not corrupted data.
Maybe Ralink drivers are telling stack they're handling TCP/UDP checksum offload, while actually they're not checking anything?

The general advice about taking drivers from generic Windows update distributions is to not do that *smile*. I've always had better results by going to the web site of the OEM that made my computers (Dell for this one) or to the web site of the manufacturer of an external network adapter (Linksys for the one that I'm using now) to get drivers that are specifically written for and tested for the configuruation and models that I actually have. I learned this a long time ago when I got a driver from Windows update that caused me to have to use Windows System Restore.

Seeing a new driver on Windows update is a good reminder to go and check to see if there's a driver update for me, but, like I said, I never take drivers from Windows update. Those are the only Windows updates that I "hide" *smile*. I guess that I never have the generic hardware configurations that these drivers apply to, and don't want to take the chance of wasting time with a driver that's not for my specific hardware.