Crossflashing of LSI 9341-8i to LSI 9300-8i. Success (but no SMART pass through)

[Cemal]

By the way @vorticity_flux

If you have looked to the SMART pass through with Hard Disk Sentinel, there was a programing issue, I have informed about it to them, and they have corrected this issue in beginning of 2018! Your post date for this problem is 30 September 2014.

So the problem you have faced with SMART pass through, possibly not related with your flashing! (If you have tested with Hard Disk Sentinel in year 2014.)

 

[nthu9280]

@Cemal

Expand drivers. There are Win10/server-2016 drivers. p16 under current and older ones under archive on the Broadcom support site

9311-8i Downloads

 

[Cemal]

@nthu9280

I am using Windows 7, P11 software is the latest possible drivers.... P12 and up, for Windows 8, and newer!

 

[BLinux]

I have an original 9311-8i IR controller. I am using Windows 7 Ultimate x64. I have upgraded it to the latest P16 bios, (not meaning the P16 software drivers). But, after upgrading it, ATTO disk write performance lowered too much. I have 8 units of Corsair GS 128GB disks with Raid 0. The write performance degraded from 3500MB/s to 2000MB/s.

So I need to revert back to P11, not to P12, P13, P14 and P15 because, there is no driver higher than P11 for Windows 7.

That was not easy. It is not accepting back flashing. So I have to erase the bios totally.

sas3flsh -o -e 7 --> deleted the current bios.
sas3flsh -f 9311_IR.bin -b mptsas3.rom -b mpt3x64.rom --> re-programmed to P11 bios.
sas3flsh -o -sasadd 500605B001xxxxxx --> added sas adress.

Everything went fine! Write performance increased to 3500MB/s again.

I did not tried to flash 9341-8i bios yet. But, @vorticity_flux I have no SMART pass through problem. I did not played with jumpers, I did not do any modifications on card.

Note: mpt3x64.rom is inside SAS3_UEFI_BSD_P11.zip file, downloadable under Bios tab of Broadcom Downloads page, upto P14. P15 and P16 9311-8i bioses have mpt3x64.rom file in "uefi_bsd_rel" directory of, for example: "9311_8i_Package_P16_IR_IT_FW_BIOS_for_MSDOS_Windows.zip" package.

you really only need to do "sas3flsh -o -e 6" , not -e 7. That will erase both the firmware and bios/uefi, but not the mfg data, which will preserve the SAS address and allow you to flash which ever version of firmware you need. people use -e 7 waaay too much.

 

[Cemal]

@BLinux

"sas2flsh -o -e 6" is for sas2 controllers, for sas3 you need to use "sas3flsh -o -e 7" when you type sas3flsh only, there is not command option being 6, but 7!

 

[BLinux]

@BLinux

"sas2flsh -o -e 6" is for sas2 controllers, for sas3 you need to use "sas3flsh -o -e 7" when you type sas3flsh only, there is not command option being 6, but 7!

No, please check your information first. Here's the manual directly from LSI:

https://docs.broadcom.com/docs-and-downloads/oracle/files/sas3Flash_quickRefGuide_rev1-0.pdf

The erase function (-e) has options 1~8, see page 13-14. "7" is for complete flash erase, including MPB, which is unnecessary for just switching from IR->IT. "6" excludes MPB.

Granted, not all versions of sas3flash support the advanced mode (-o) required for the erase function. but where it is available, the options for erase are on page 13-14 in above linked manual.

 

[Cemal]

@BLinux

Yes you are right, it writes 7 parameter is not specific for sas3 controllers, my mistake!

6 - Clean flash (erase everything except manufacturing area)
7 - Erase complete flash

But, nevertheless, it inserted the manufacturers data automatically, after re-flashing.

Read this post, it explanins how:

Changing the PCI IDs on LSI 2008/2108-based MegaRAID controllers

Soon after, I have realized that, the problem I have faced with low write issue, is not because of latest bios of 9311-8i controller, but possibly changing the mainboard that I used. They both are same brand and model motherboard, Asus Rampage IV Black Edition, but somehow, one of them performs lower write score. That mainboard also can not power save and do not lowers to x12 divider, all the way it runs at processor frequency. (Checked bios no issues.)

I will re-load the SSD to be sure. If not, maybe something wrong with the second Asus Rampage IV Black Edition itself...

 

[nthu9280]

Nice link on MegaSCU. Does it work on HBAs? Many MegaRAID utilities only work with cards in IR / iMR mode.

 

[Cemal]

@nthu9280 The list at the link all Sas Controllers and the id's are all for Sas Raid controllers, not HBA's.

To use the same way, HBA id's for those cards first have to be known. After it, can be tried, why not!

But, flashing procedure, flashing files are different. I am not so sure....



The link: Changing the PCI IDs on LSI 2008/2108-based MegaRAID controllers worked for my IBM ServeRaid M5014 card lately very well, fast boot with latest 0239.rom bios, become a real LSI 9260-8i controller in the end, (except the half cache). The subject opened in year 2012 and closed with %100 success in year 2018. I have tried to update the bottom link, but can not see the result of the update, possibly because of being a very old dated topic! (Topic closed....)

But, it was not possible to learn that procedure earlier than this year, because it says: "This document was first published 2018-01-25 21:58:57, and last modified 2018-01-25 21:58:57."

https://www.servethehome.com/flash-ibm-serveraid-m5014-lsi-92608i-firmware/#comment-463698

 

[kamiccolo]

Hey everyone!! So I managed to grab a 9341 8i but im having huge issues with 850 pros. So the write on a 850 pro is 500 mb/s but on the raid controller with the stock firmware and bios the writes are at like 33mb/s abysmal. Will flashing my card to a 9300 8i remove this weird issue with the 850 pro's?? I have 8 of them and wanted to use the 9341 8i for raid 0. I get 3800 mb/s read and 380 mb/s write in raid 0 8x 850 pro. Can anyone comment on this??

 

[kamiccolo]

Holy cow ok got it fixed that's lucky.
So this one is pretty interesting.....
All my ssds individually were down to like 50mb/s write using my raid card only. If I poped them in my mobo they work at the proper speeds.
I was like ok so at this point the raid was already broken down this was testing each drive individually. First thing I thought was the new windows update since im on x99 it already screwed my overclock but that drove me to install Linux right.

So I usb Linux and checked the drive speeds in Linux.
The exact same scores perfect read craped on write at like 30mb/s it was bad.
It was interesting tho because there were 3 partitions on all my hard drives that Linux picked up!! I was like what I thought the drives are clean in one partition nope apparently not somehow.
So then I went into the raid card bios and switched each drive from Jbod to unconfigured good through erased each drive separately. Then I actually turned them all back Jbod for the moment to check in Linux.
I checked each one in Linux and the write speeds have almost tripled back up to almost 400 for each drive separately in Linux.
So I was like going on out now so I secure erased each drive separately with Samsung secure erase on usb.
Recreated the raid 0 in the 9341 8i bios and tweaked all the settings for best performance.
set my max oc back up and installed windows 10 64 bit latest version.

Loath and behold the write issue with the 9341 8i was gone. No longer bottlenecked at 330mb/s sequential write with 8x Samsung 850 pros raid 0 lol.

Now the interesting thing is that I had this working before the windows 10 july update at full speeds but I don't know if its it for sure but it could have been.

The other screwed up thing is that all the drives worked fine on my onboard intel controller at max speed but not on the 9341 8i.
I think thorough erase in the 9341 8i bios is what really did the trick but even the secure erase from Samsung does wonders too.

But anyways here are the results!

[/url][/IMG]