TRIM update hosed my Windows 7 install

[idata]

Just did the firmware update and it hosed my Windows 7 installation. The updater showed a successful firware update. Initially the computer booted just fine, but once I was within Windows it installed some drivers and asked for a reboot. That's when the trouble started. Now the drive won't boot Windows 7 anymore. I don't know if it's a Dell problem or Intel problem. The Dell BIOS claims a SMART error. I have a Dell XPS 8000.

[idata]

Hi ambizytl,

There is no connection win x64. That is not the issue. Your own configuration proves that. If you notice and you have to read the statement carefully. All Intel did was acknowledge that they can re-created the problem. What they did was re-create the problem based off the what people have posted, but i dare say that they know what the root cause of the problem. Re-creating the problem may be very easy. Understanding why it is a problem is what the engineers are probably stumbling at. They may very well have a fix for recovering the "bricked" drives, but the problem they are facing is "how do i not effect those that upgraded and are NOT having problems" And the second thing is recovering the "bricked" drive may not stop the drive from become unusable again if there is some external factor causing the drive to become unusable. Though the TRIM command is a standard spec, the implementation of TRIM by the various vendors IS NOT STANDARD. There is no standard spec that says how TRIM should be implemented in the WIN OS environment. Same for Linux and Unix also.

The person that wrote the update IS NOT an engineer. He is probably a PR guy. Basically the statement says the firmware works and doesn't work for some x64 configurations. The only other revealing information is that they have recreated the failure, but here is the KEY, it does not say that they understand why the drives are bricked. The excuse about the update tool is bogus. The update tool is run in FREEDOS, not a windows environment. NONE of the poster's SSD's were bricked before starting Windows, so obviously, the update tool did not corrupt the firmware.

Thank goodness i'm using a MAC and can recover the speed of my SSD by doing a secure erase, then restore. Windows users are going to tied at the hip between MS, Intel, and their chipset vendor and it only takes 1 out of the three to do something that the other is not expecting for the drive to become corrupt.

[idata]

Creating a fix might be a difficult problem since Intel now essentially needs two types of firmware: one for fixing bricked drives, the other one for updating drives that work while still preserving all data stored on it.

Personally, I doubt that they will be able to offer a firmware which can repair the bricked drives without the need of a complete erase. It looks like the firmware is randomly damaging sectors on those drives (maybe each time a TRIM command is issued?), something which cannot be recovered at a later stage even if the root of the problem gets fixed.

Also, even on the drives that do work, there could be some hidden damage such as invalid SMART data. E.g., my drive shows far too high values for the "Host Writes" SMART attribute, and I'm not sure if this can be attributed to the new firmware. Thus, Intel might be forced to only offer an update which requires erasing the entire drive since any data on it can't be trusted as long as the bad firmware has been flashed beforehand.

As all firmware updates up to now can be applied while keeping all stored data, Intel is probably evaluating if they can offer a firmware which at least saves the data on working drives.

There is almost certainly no "external factor" causing the drive to become unusable. TRIM support works fine in Windows 7 with Indilinx drives. This means that MS is not to blame. Essentially, TRIM tells the drive the sectors which are marked as deleted by the OS. This is easy to implement by the OS supplier but difficult for the SSD manufacturer since the latter has to do all the dirty low-level work. It's Intel's fault only, it's their firmware which is doing all the mess. My guess is that a firmware algorithm - the one which calculates the sector adresses which are being cleared once a TRIM command has been issued - is messing up under some special conditions. Thus, the wrong sectors are being cleared.

I have no idea what this sentence is supposed to mean: "No related issues have been reported by users who have successfully upgraded to 02HA firmware via the firmware upgrade tool". I mean, at first, pretty much everyone upgraded to 02HA via the firmware update tool version 1.3, so why mention it anyway? Then again, of course, those who successfully upgraded did not experience "related issues". After all, if they had experienced related issues, they would not belong anymore to those who successfully upgraded, wouldn't they? At least as long as a "successful upgrade" means that the SSD survives not only the flashing process but also the next reboot.

That's called circular logic: "Those who had no problems had no problems". Yeah, I guessed so.

dbm wrote:

NONE of the poster's SSD's were bricked before starting Windows, so obviously, the update tool did not corrupt the firmware.

This isn't obvious in any way. There are no data written to the SSD before you start Windows. Once you did, Windows will do lots of small writes to different places, especially if your swapfile is turned on. If there is an error inside the firmware code which handles writes, it won't surface until some data have actually been written to the SSD. If the error is within the firmware code which handles TRIM commands, the SSD won't be damaged until the OS actually issued the command, which can routinely happen at bootup (e.g., some temporary files might be created and then deleted, which triggers a TRIM command).

I don't know exactly, but I'd guess that concerning those SSDs which didn't survive the reboot, trying to install any other OS with TRIM support without having run Windows 7 once after the firmware update would have resulted in the very same problems.

[idata]

Just wondering

Those who failed/bricked their drive, how did they go about changing between AHCI/RAID to IDE and back?

I installed Windows 7 x64 in IDE mode (didn't realise BIOS was set this way) and then changed to SATA/AHCI mode by editing the registry.

When I attempted to boot Windows 7 in AHCI mode before I made the registry change it would just blue screen and restart immediately on loading Windows.

After changing the registry, it booted into Windows and loaded the appropriate driver.

I then restarted as it asked and all was well.

After another re-boot, Windows 7 could not find the boot loader so I had to use the Windows 7 disc and repair the start up/boot loader.

All is well now and things are much faster in AHCI mode than IDE.

So my question is...

Have people changed the IDE mode to install the firmware, and then forgotten to change back before re-booting into Windows?

Just a thought

Edit: I haven't flashed to the new firmware as I got the SSD a couple of days ago, after the firmware was taken down.

[idata]

Hi all,

I couldn't wait when the firmware update was out, so I upgraded it. Everything went well. Just for the curious:

1. I had a 160 GB G2 SSD with an (fresh, 2 days) Win7 x64 install already on it, plus the Intel Chipset drivers and the Matrix Storage Manager.

2. I had Win7 installed in AHCI mode and never changed it since then in the BIOS.

3. When I tried to flash first, the update tool wouldn't detect my SSD at all. Got me a bad stomach

4. When I switched my SSD to SATA-Port 1 and set it to HDD0 in the BIOS (and disconnected all other drives besides one optical as well, but I did that simultaneously, so i can't tell which consequence a single action had) it detected the SSD.

5. Flash went well. Same with the reboot. No problems since then.

6. SMART shows everything's fine, besides the "writes" seem to be >400 GB for a drive being 2 weeks in action > a little bit too much considering a 15 Terabyte durability limit for those SSDs?!

[idata]

1.Your drivers from what I know do not pass the trim command since intel matrix storage manager doesnt support it (only the generic windows one do).

2.Your writes are possibly down becasue the "garbage" are still in your drive. You have to run the -now unavailable- SSD toolbox to clean the drive from the old stuff, and "downgrade" your drivers to windows generic ones.

3.There is a possibility if you revert to windows generic AHCI drivers or run the SSD toolbox optimization (which both send the trim command), to get the same nasty problem everyone mentions.

*If you remember the problem starts after the first reboot after the firmware, where windows installs specific software to match the firmware upgrade.