Slow Bios after flashing Marvell 2.2.0.1125 - how to restore original Bios FULLY

[Lsdmeasap]

Remove ALL USB Devices, printers, hubs, ect and set BIOS option Legacy USB Storage (Or legacy USB detect) to disabled (Integrated Peripherals), that may help with the memory checking speed, but it may not as it may just be slow because while that's running the board is also trying to load the Marvell options which are failing.  

I only said the Marvell needs to be in AHCI when you update the AHCI because that's how you should do it, it has nothing to do with Intel AHCI and cannot affect that at all.   Marvell is it's own chip, Intel AHCI is inside the Gigabyte BIOS, totally separate from the Marvell chip and thus the Intel AHCI cannot be touched by the Marvell flasher.   You are just reading too deep into all of this   Anything can slow anything else down on the boards booting procedure, but the failed flash isn't because it messed up anything else other than the Marvell chip's firmware itself.

Just use your computer as you normally would, with as little down time as possible so you don't have to worry about booting up so often.  Then wait until we can find out more, or if no one has any ideas I'll try to brick mine for you and then see if I can fix it.  I don't use the Marvell controller anyway, nor that board for anything other than testing and helping others on stuff like this, so I have no problem trying to purposefully brick that controller to try and help you!

Sit back and wait, let's see if we can get some help and advice on this from others before moving any further.

[CuteBirdy]

Remove ALL USB Devices, printers, hubs, ect and set BIOS option Legacy USB Storage (Or legacy USB detect) to disabled (Integrated Peripherals), that may help with the memory checking speed, but it may not as it may just be slow because while that's running the board is also trying to load the Marvell options which are failing.  

Those things I did before, but didn't make memory checking any faster. I remember reading you advising someone to take off their scanner, and set Legacy off, and it helped him, but not my case, probably because (like you said) the board's trying to load Marvell stuff that are failing.

I only said the Marvell needs to be in AHCI when you update the AHCI because that's how you should do it, it has nothing to do with Intel AHCI and cannot affect that at all.   Marvell is it's own chip, Intel AHCI is inside the Gigabyte BIOS, totally separate from the Marvell chip and thus the Intel AHCI cannot be touched by the Marvell flasher.   You are just reading too deep into all of this    Anything can slow anything else down on the boards booting procedure, but the failed flash isn't because it messed up anything else other than the Marvell chip's firmware itself.

Hmm ok, but if that's the case, why when I revert to an older BIOS, like F7 or F2, I can't get the BIOS to boot when a hard drive is connected  under AHCI mode? It just freezes after the "wait a few seconds" line, and nothing happens. But, maybe this is not a problem with the Intel AHCI, but a conflict with the Marvell things that were loading incorrectly during the memory checking stage... who knows.

Just use your computer as you normally would, with as little down time as possible so you don't have to worry about booting up so often.  Then wait until we can find out more, or if no one has any ideas I'll try to brick mine for you and then see if I can fix it.  I don't use the Marvell controller anyway, nor that board for anything other than testing and helping others on stuff like this, so I have no problem trying to purposefully brick that controller to try and help you!

Sit back and wait, let's see if we can get some help and advice on this from others before moving any further.

I'm lucky to have someone as helpful as you attending to my problem, risking your own board to help others thanks so much. OK I'll sit back and relax, computer is running fine right now except for the slow memory checking stage, ironically the amount of time I've spent on this problem probably exceeds the total extra time that the bootup takes for the life of the board. Going to bed now, will keep my eye on this thread when you or others chime in!

[F5BJR]

*
Station drivers firmwares are not for BIOS but for PCI-e Sata Host Adpater card with 88SE9128 as this :

http://www.ioisata.com/products/Host-Adapter/S3-PCIE1XG202.htm

Pierre

[Lsdmeasap]

Thanks F5BJR, I thought that as well, however I have seen many users including Gigabyte board users flash that firmware successfully.  Or at least they say so, maybe the ones I see mention this are lying out their rear's?

Do you know how this user can recover from the incorrect flash?  Or is he going to have to send his board in so they can reflash it via eprom flashing, or just give up use of the Marvell controller and deal with the other issues this has caused?

[CuteBirdy]

I'm happy to give up use of the Marvell controller, I just want to get normal booting times again.

Problem is right now, even if I disable the Marvell ports in BIOS and don't use the Marvell, I get the problems.

In addition to the slow memory checking stage I mentioned earlier, I notice now the "quick boot" function you can enable under BIOS, might not be working properly anymore: after I do a hard reset or turn off/turn on the computer, the BIOS will scan memory again, even though I have "quick boot" on. Isn't quick boot supposed to skip the memory scan after the first successful scan? If I do a soft reset however (control alt delete), the quick boot seems to work, because BIOS doesn't say "checking memory" or to that effect, so it's a little faster. But still slow due to the freezing.

I'm afraid sending the board in for warranty, won't be feasible can I do the "eprom flash" myself?

The Marvell firmware doesn't say it's meant for the PCI-E device, nor is it mentioned under the station-drivers page, and the firmware doesn't look for the correct hardware before doing the flash.... it is a dangerous tool!

[F5BJR]

*
If you have a GA-X58A-UD5 with 8SE9128 as me you can flash

but only the embedded firmware ( little eeprom 8 pins )

In the BIOS you have 2 * options for running the Marvell port :

1 - using this firmware
2 - using the normal BIOS firmware

3 - also the possibility to copy the firmware in BIOS to the 88SE9128 firmware

**

Another solution but risky is to modify the BIOS with CBROM32 !!

Pierre

-> for restore original

- reflash your motherboard BIOS and use BIOS option to copy Marvell firmware part in the little eeprom if BIOS update not automactly


 

[Lsdmeasap]

Pierre,

Thank you for your replies and help on this!!

He has already tried reflashing the BIOS, and tried the Force Marvell Firmwrare update from within the BIOS, it does not correct the issue.   The X58A-UD7 does have the 9128 Marvell controller, same as the UD5 you mention, that is why I said I've seen heard users say they've flashed these board successfully, and now you agree and say it is possible as well.

So did he flash incorrectly, or if not why did it fail for him?  He said he used the AHCI go.bat file, but the Marvell controller was in IDE mode.  Do you think that was the cause of the problem?  If so, do you think if he flashes it again, with Marvell AHCI mode enabled, it will correct it?

I can modify a BIOS for him just fine with Cbrom, I've done it many times for others and myself, but the issue is he can't even get Force BIOS option to fix it with a normal un-modified BIOS.   He says he did that, and it does update the firmware like it should, but does not correct the issue.   So I was thinking that when he flashed AHCI firmware while in IDE mode that caused the issue, do you think a reflash in AHCI mode will correct it?

Also, what did you mean by this comment?  " but only the embedded firmware ( little eeprom 8 pins ) "

Can we flash the firmware offered at Station-Drivers to motherboard onboard Marvell controller SE9182 or not?  You said no before, and now yes, but then you mentioned embedded, so I'm confused if you mean yes or no for flashing the firmware from there or not.  These boards DO have the exact same Marvell controller as your UD5 model.

I am willing to flash my board to test various version for him, but I need to know for sure if it should be possible or not, and you said you do with your UD5, so it sounds like it!   Although, I'm not sure I'll be able to brick mine like he did, if it's supposed to work then it will probably work fine for me and I wont have any problems, so wont be able to further help this user? 

Hopefully if he reflashes in AHCI mode as I'll mention below it will correct the issue


@ CuteBirdy,

Have you tried reflashing that same firmware yet, but in AHCI mode for the Marvell controller?  How about flashing the next older version Firmware instead?   I've seen many users have issues with various versions, so maybe it's just the version you tried that's causing the issues?   This is definitely worth checking, sorry I forgot to think of that earlier!   If you do flash an older version, be sure to set AHCI mode for the Marvell when you flash the AHCI firmware.

What does the controller currently do itself, is it failing?   I mean does it work properly in AHCI mode, and IDE?  What about RAID?

Do you have other memory to try with, to rule out your current memory as a possible cause of the slow memory check issue?   If not, I know you said you removed all USB Devices, are you sure about that?  I meant USB Keyboard and mouse too, remove all USB device and see if that stops the memory check slow down.

[CuteBirdy]

Hmm, no I haven't tried reflashing the Marvell AHCI firmware with Marvell AHCI enabled in BIOS yet, because I'm so afraid of it making the motherboard even slower. But if you insist, I can try it later in a few hours.

As for flashing older firmwares, I have tried F7 (the latest stable BIOS) and F2 (the first BIOS for the motherboard), both yielded exactly the same results - namely - no noticeable memory freeze, but when I enable Intel AHCI, the whole BIOS freezes to a total stop as it's scanning for a AHCI hard drive I connected.

To summarise, we have 2 outcomes:
- Flash BIOS F9+ (what I'm using now): Memory check for 6gb (3x2gb) 1600mhz RAM takes an unreasonably long time, up to 16 seconds, and "quick boot" feature doesn't work properly. But having hard drive connected to Intel AHCI port (when enabled) has no problem.
- Flash BIOS F2 or F7: Memory check no problem. But hard drive connected to Intel AHCI port (when enabled) causes BIOS to freeze.

You suggested flashing the "next older" Bios, you mean F6? If F2 and F7 don't work, doesn't it suggest it will be the same for all BIOSes in between?

Something just occurred to me. BIOS F9+ is a bigger bios than normal. While BIOS F2 to F7 are normal size. Could this be why I am getting 2 different  alternative problems, a sign of what the problem is? Could the Marvell firmware be writing in the wrong sections, because it thought the BIOS size was normal, when in fact it was bigger? (At the time I flashed the Marvell, I was using F9d)

You ask me how is my controller currently. Which controller? Intel or Marvell? Intel - it fails only when I'm on BIOS F2 or F7, it is fine when I'm on BIOS F9+. Marvell - it is fine on any of the BIOS, maybe because it hasn't been touched by the new firmware yet.

As for other memory to test with - I don't have. But I'll tell you what I've done. I have 3 sticks of memory (triple channel). I checked with memtest86, no problem. Also I have tried switching the sticks around the slots, as well as leaving only 1 stick plugged in at a time, for each stick. The problem still exists. It cannot be that all 3 sticks of memory have become faulty, all at one time. And, before I installed the Marvell firmware, the memory all worked fine. So I really don't think it's the memory problem. Although worth noting, the memory freeze duration depends on the number of sticks plugged in, so for example when only 1 stick is plugged in (2gb) the freeze isn't as long, but still noticeable.

Regarding USB devices - yes, everything USB was removed, still had the slowdown. My keyboard is a USB keyboard but I have a USB to PS2 converter plugged into the PS2 slot (I did this so I can use the "power on" button on the keyboard to turn the computer on). So, I believe the keyboard is seen by the computer as a PS2 device. Ive been using the keyboard this way for years, with no problems.

[Lsdmeasap]

I meant flashing older Marvell firmwares, you know, not the latest version.

Motherboard BIOS, always use the latest

As for the BIOS sizes, no that is normal, and has nothing to do with the firmwares.   The only reason you may be seeing differences between versions is just how each version differs in doing things.   Just keep on the latest BIOS.  

And I would go ahead and try flashing in that same Marvell firmware you used before, the AHCI one, but this time with the Marvell controller in AHCI mode.   If that does not help, try an older version of the Marvell firmware.

When I asked about controller, yes, I'd like to know about both.   Is Intel AHCI bootable with hard drives attached, and is Marvell ports bootable with hard drives attached, in AHCI mode, IDE mode, and RAID mode (you'll need two disks and a RAID created to test that, so probably not feasible).

Ohh, sorry, I reply as I read.   So, BOTH Intel and Marvell contrllers function fine?  In AHCI or IDE mode?  Really?  If so, then there is nothing wrong with yoru Marvell controller at all!    And the flash probably went fine, you can verify the firmware version in windows with Marvell Tray installed, of from the Marvell RAID BIOS, set RAID on Intel and AHCI on Marvell, reboot and go back to the Marvell and set raid or enter the RAID ROM for it in the BIOS, version will also be shown there.

If both controllers are working properly, as it sounds like you are saying, then there is no controller or firmware issues.  The memory issue would could be anything, and since you don't have other memory to test with you wont know if it's due to your memory until you can test other memory.  The firmware cannot be incorrectly flashed if the Marvell controller is working properly, so it sounds like we are both maybe wasting time thinking about this as it sounds like it works from what you just said!

You are right, it's not very ideal that all of your memory went bad at once, but it would be nice if you could borrow a single stick from someone for testing.   You are using ONLY the white slots correct?  If you are testing 1, 2, or 3 sticks you must use the white slots only.   Just saying, I know you probably are because otherwise it's usually not bootable in the blue slots unless the white ones are already full.

Hope Pierre will reply in depth to my questions above, so I can consider flashing to test this for you!  It's not that I wont or don't want to test, I'd just rather not get out the board, remove my current board, setup the other, ect all for nothing if you have already properly flashed the firmware and it's installed as expected, or would rather not do all that if there is no way it can flash on these boards (IE you never had a firmware failure on flash because updater does not work on these boards, thus fail is impossible)   Hard to say because he said you can't do it, then next reply he said you can on 9128 UD5 boards which is the same chip and PCB as UD7 boards.   So hope he will confirm yes or no, and you confirm my thoughts on your firmware flash actually taking, because it sounds a lot like it did properly now from all your comments tonight.

Please install Marvell Tray and then look in these sections and see what firmware you have installed, and tell us if any of it looks like the updated version you tried flashing, or confirm within the BIOS RAID module itself in the Integrated Peripherals section

[CuteBirdy]

Huh, now you are saying there is no problem? Why? Wow this is complicated to explain, I'll try again...

AFTER THE MARVELL 2.2.0.1125 AHCI firmware update, on my X58A-UD7 rev 1.0, bios F9d - while Marvell was set to IDE (not AHCI):
1) Intel controller is working fine (both IDE and AHCI), when I keep BIOS F9d. BUT when I changed BIOS to F7 or F2 (to get rid of the memory freeze), the Intel controller was NOT fine when set to AHCI, it would freeze when scanning any connected hard drive at BIOS stage (but fine on IDE). In other words, the Intel controller was fine only when I have the stupid slow memory scanner problem!
2) Marvell controller I didn't actually test (both IDE and AHCI), like I didn't plug a hard drive into it to test, and I didn't look for version numbers of the Marvell in Windows. Because I didn't really care about the Marvell anymore. I don't plan to be using the Marvell after all that trouble, I just want to get the problem 1 above (Intel controller problem) fixed, together with the slow memory scanning problem, because those things are weird problems that Marvell shouldn't have interfered with! However, I enabled (under BIOS) Marvell for IDE and later AHCI, and booting up was fine with both settings. The Marvell controller didn't freeze the BIOS, unlike the Intel controller.

You say there is no problem with the Marvell - but there is: the Intel controller problem, and the slow memory scanning problem, happened straight after flashing the Marvell firmware, and can't be fixed even after reflashing Gigabyte BIOS back many times!

So I'm not trying to get the Marvell to work. I'm trying to fix what the Marvell firmware update screwed up.

You blame it on the memory - but why all 3 of my memory sticks go bad at once? Why so coincidentally it happens right after the Marvell update? And if it's just my memory problem, why is there another weird thing happening, the Intel AHCI not working when running F7 bios? (And yes, I put the RAM in the white slots, that's obvious)

All these weird things happen straight after the Marvell update! That's how I know the problems are caused by the Marvell. Even though the Marvell itself might be working properly after the flash, it has stuffed up my board's memory scanner (when it's running bios F9d) and Intel AHCI (when it's running bios F7 or lower).

[Lsdmeasap]

Sorry, I don't have time for a big reply this time.    Please stay on the latest BIOS, if you want to use older BIOS that's fine, but not suggested at all due to older BIOSes often having various issues and they do not include the latest fixes and computability enhancements.  All these BIOS flashes, are you shutting down and clearing CMOS fully after you flash?  If not that could be the reason of some of the problems.  Can you please post a list of your hardware as well, since we are now dealing with other hardware issues, thanks!

I understood you issues, you are not understanding that the Marvell flash or Marvell ROM cannot affect the Intel one or memory check like you are thinking.  The Marvell ROM can only affect the Intel one, when they are both in RAID, and that is only for certain BIOS versions.

It's great you are not trying to get the Marvell to work, as it's not too great of a controller.   Arrrr why did you do this in the first place then?      I know why, I'm just kidding - we all need a good laugh sometimes to help keep it cool

Please don't be getting mad or frustrated with me, I do understand the issues you are having, and am willing to help by trying to do the same to my board, but you need to understand the issues as you just put them have nothing to do with Marvell firmware.   It's just coincidence that they appeared at the same time you did that, it has to be.   Yes, things do happen like that, that's why the word was created, and it happens to us all throughout our lives, often called irony instead of coincidence and we hate it because of that.  Of course the BIOS version differences may make the issues appear differently, but it sounds like they are there no matter what to me.  So can you please install the Marvell Tray for me, and show me your firmware version currently installed, so I know if I need to flash mine in order to try and help you sort this out.  

I did not blame anything, I'm only trying to help you sort it out by suggesting testing other things which can possibly lead to the issues you mentioned.

Please do an Extended Clear CMOS, with the latest BIOS you have been using up until this flashing incident.   To do that unplug your PSU from the wall or the motherboard, or switch it off in the rear.  Then press and hold the case power on button for one minute, then press and hold the clear CMOS for one minute, then remove the CMOS battery and let the board sit overnight or at least 8-12 hours.    I know that may sound odd to you, but this is how you often can fix very odd issues such as this.  Please do it that amount of time, anything less and it may not work.    But, you can try a short CMOS clear first and see if that helps, if not do the long clear please, so we can see if it corrects your Intel AHCI problem.

When you try the board again, please connect your hard drive to SATAII_0 and CD/DVD drive to SATAII_5.  Hopefully this will clear that up!  Before or after all that I'd also like to see your Marvell info like I posted above, so please install Marvell Tray, just so I can verify if the update did indeed do through properly, thanks!

[F5BJR]

*
For me i use BIOS modified with CBROM32

But only for the fun because i connect rarely disks to this port !!

I use always in RAID mode

Pierre

full firmware content for modify BIOS with CBROM32 ( rename zip to rar )

[Lsdmeasap]

Thanks Pierre, but I already know how to do it with CBROM too.   I just wanted confirmation from you, that it's possible to do with the flasher that comes with the Statio-Driver Firmwares.

However, I believe he flashed OK now, he just needs to verify for me, then I will test the exact procedure he used, and see if I also get the same errors and issues with the memory scan at startup, and the Intel AHCI functionality.

@ CuteBirdy

I forgot to mention this earlier this morning, please also go into your BIOS in the Integrated Peripherals section and disable Legacy USB Storage Detect, this may help with the slow memory scan at startup, even if you do not have USB devices connected.   And about those, also be sure you remove and front ones too, including card readers or hubs, just in case that's what is causing the memory scan to slow.

[CuteBirdy]

Yes after each BIOS flash, I am shutting down computer, pressing clear CMOS, turn on, load optimized settings, then enter my own changes.

My hardware is: i7 920; 6gb (2 x 3gb) GSkill NQ 1600mhz DDR3; WD 1tb blue; ATI Radeon 4850; Vantec 600W ION2+

I am not using RAID at all.

If Marvell flash cannot affect the Intel or memory check, then what else can? I didn't do anything else to my computer at the time when the memory check problem happened.

I didn't really need the Marvell, but tried to get it working, because I had a new SATA3 SSD to test, was curious if it could work better on Marvell SATA3 than Intel SATA2.

I'm not mad or frustrated at you, I know you are trying to help, it's just so hard to find what the problem is.

I installed the Marvell tray, right now it doesn't say anything (no device showing) because remember, I had just done "go -e" which erased the Marvell firmware. I still need to do what you asked, that is, install the Marvell again (sigh I hope no new problems reinstalling it), plug a HDD in the Marvell, then see what the tray says. Will tell you when I get the chance to do it (within an hour).

I've done a short CMOS clear, no difference. Will do an extended CMOS clear tonight, and report tomorrow.

HDD is currently connected to SATAII_1, will switch to SATAII_0 soon.

BIOS already has SATA0-3 mode set to "Native" (not sure if that's what you recommend). "USB storage", I have tried both "enabled" and "disabled", but no difference. I currently leave USB storage as "enabled" because I need it to recognise my USB stick, which contains the marvell firmware I need to flash soon. In case you ask, no i currently have no USB stick connected to the computer, I already said, even with no USB devices connected at all, I get the memory slow problem. (but as mentioned, no memory slow problem with older BIOS, but then instead I'll get Intel AHCI problem)

[Lsdmeasap]

Thanks for the further info!   Did you try an extended clear CMOS yet as I mentioned above?  If not please try it tonight while you sleep.

I told you what can affect that, USB items, memory errors, faulty drives, CD/DVD drives, KB/Mouse, ect.   Anything you connect to the board can affect it really.

No, SSD 6Gb/s will not work better on the Marvell anyway, just so you know    Thanks, I'm glad to see you are not mad or frustrated with me, I wasn't sure from your last post so I was hoping not!

Ahh ya, that would be why the Tray software isn't working then!   Go ahead and flash again with the Station-Driver files, if those are what you want me to test, if you don't care about that and don't want me to test, then you can just use force from the BIOS and flash in the version that's included in the BIOS.

No rush on any of that about the controller, as I may not be online much longer tonight anyway.

Native enabled or disabled shouldn't matter, but it could I suppose.  Try both ways on that and see if it affects your memory scan speeds or not.