LSI 9270-8i IT mode

[vanfawx]

Well, we'll find out. I've sent @Freebsd1976 the SBR and SBP files I have from a retail 9270-8i.

 

[Freebsd1976]

Well, we'll find out. I've sent @Freebsd1976 the SBR and SBP files I have from a retail 9270-8i.

Will try in this weekend and report back .

1 first flash 512MB card (1G card in my nas ), if flash fail then stop , and convert it to 9207-8i HBA.

2 if card flash ok then see if it can run ok , if ok then we success

3 if run failed , will shutdown nas server , replace 512MB cache with 1GB cache ,see what happens

 

[vanfawx]

Just don't forget to save the SBP and SBR from your current cards, just so you have the ability to revert if things go wrong @Freebsd1976

 

[Freebsd1976]

there is report back , flash 1G m5110 RAID CARD
1 backup sbr and spd ok , and use md5sum to check , 9270 and m5110 card spd is same,but sbr is different , so only need flash sbr file
2 write emptysbr 512 ok
3 cleanflash ok
4 reboot ok
5 writesbr ok
6 flash rom ok (but show a message "Flash signature error .trying to clear it")
7 write sasadd not ok (sas2flsh error, don't know howto use megaoem or megascu to setaddress,if anyone know howto do this ,plz tell me)
8 reboot
9 card is ok ,no error ,system boot ok ,disk ok
seems crossflash is ok ,firmware version indeed changed, but what strange are :
1 when I use CRTL-P enter webbios , it still say it is an ibm m5110 ServerRAID Card,
2 so I boot to dos ,use megarec -readsbr 0 new.sbr , and compare with lsi9270 and backuped m5110 sbr. found newsbr different with 9270 ,same as backed M5110

anyone can help with this? I will do crossflash again and see what happen .


edit:
redo crossflash success twice , both write empty sbr (use dd if=/dev/zero of=nullbin.sbr bs=512 count=1)
1 use 2208_16.rom (firmware version 23.16） and lsi9270.sbr
2 use lsi2208 (firmware version 23.23）and lsi9270.sbr
but both still show m5110 card, not lsi 9270-8i

edit2:
after multi cleanflash and writesbr , I used megacli -adpallinfo -a0 found my sas address still there ,why this happend？

 

[Freebsd1976]

@vanfawx @fohdeesha did you know why crossflash successes but card still said it is an ibm serverraid m5110

 

[vanfawx]

@Freebsd1976 sorry, I don't.

 

[necr]

Thanks to @w0mbl3, finished flashing 9286CV-8e to 9207-8e IT firmware.

Took:
Downloads for Broadcom (LSI/Avago) SAS II HBA 9207-8e firmware+bios+sas2flash
lsi2208fixer.zip from the other thread, created a new dir in that .iso with contents above,
used the guide Flashing LSI 2208 with IT firmware to use as an HBA - My Wired House
had to carefully pull out a speaker from the PCB, looked at BIOS options and deleted BIOS after.
The only issue I see is with VPD read error by kernel, need to back it up next time with sasflash -o showvpd and write with sasflash -o -vpd <file.vpd>

Attaching BIOS options for you to decide if you need them.

 

[ServeTheHarry]

@BLinux I think I can help with that. I still have my SBR backup of the 9270-8i I converted to a 9207-8i IT mode.

I've sent you a PM.

Hope you don't mind me asking for PM too.
Little story/warning about the Intel Server Board M10JNP2SB:
Besides the rediculous cut-down to only 3 expansion possibilities (no proprietary PCEe and no 4th slot anymore like my previous RainbowPass), my favourite RMS25PB080 doens't work in that board. It's the 2208 chipstet, distributed with both chip revisions (C1 and D1, G27501-601 is C1 with PCIe2.0 and 256Byte SBR, G27504-610 is D1 with PCIe3.0 and 512 Byte SBR).
All other SAS controllers I tried do work. But the RMS25PB080 doesn't boot (only in this board, works in all other boards).
My intention is to transfor it into a 9271-8i to see if that changes the game.
But as others reported, there's no 2208 SBR (what ever this stands for) auto-healing when m0-flashing.
Since I have both (C+D) versions, I'm looking for 9266 and 9271 SBR binaries. Anybody able to share 768 bytes? ;-)
Thanks,
-harry

 

[vanfawx]

Hey @ServeTheHarry The SBR I have is from a retail 9270-8i, not a 9271-8i. I think the main difference is the 9271-8i includes a SKU to come with CacheCade Pro.

Send me a PM, and I can work out giving you the SBR I have.

 

[ServeTheHarry]

Hello @vanfawx, thanks, I did so.
I'd like to share that in the meantime I successfully converted a RMS25PB080 into a M5016. The WebBIOS reports so and subvendorXXXids are changed permanetly with holding valid NVDATA and booting.
Unfortunately this didn't chnage the problem with the Intel Server Board M10JNP2SB. No D3-D0 transition at powerup. I filed a Intel support case but they are going to refuse fixing anything. 1st level support insists that only "validated" controllers need to work, despite there PCI-SIG standard and both products are sold with explicit stated compliance to these standards. Trying to escalate, but after mentioning on the phone that I had the feeling that this card is intentionally black-listed , the shocking answer was »maybe«. Will have a look what other vendors provide – after more than 10 years of Intel Technology Provider in Gold :-(

-harry

 

[nthu9280]

Worth a shot to try with electric tape on SMBus pins on this card for this motherboard.

 

[ServeTheHarry]

Worth a shot to try with electric tape on SMBus pins on this card for this motherboard.

Thanks for the hint, already tried that without success unfortunately.
If I had more time I'd change the PCI device id just to see if it's black listed – my strong guess. Never had any compatibility issues with that controller…
I'd have to find out what device IDs megarec feels responsible for, otherwise a new controller was bricked :-(
And probably there are other similar severe consequences to think about. No idea if MR firmware refuses to boot with arbitrary vendor/chip IDs, subvendor will be evaluated at flash time at least. Maybe there's already more than one device ID for the 2208, otherwise lending one from 2308 could work. But no time to play unfortunately. And even if one manages to get the controller working in the intel board that way, one would have to patch the drivers to attach to foreign IDs – piece of cake for FreeBSD, but on machines utilizing 2208 one might run ESXi more likely and that implies at least switching to an acceptance level wich nobody agrees on in enterprise environment; and I don't have infrastructure to package patched ESXi drivers...

Thanks,
-harry

 

[BeTeP]

RMS25PB080 is an "integrated RAID module" which is Intel speak for "this controller will only work in specific Intel motherboards it was designed for". The M10JNP2SB is not on the list. There is nothing to fix. It is by design.

 

[ServeTheHarry]

RMS25PB080 is an "integrated RAID module" which is Intel speak for "this controller will only work in specific Intel motherboards it was designed for". The M10JNP2SB is not on the list. There is nothing to fix. It is by design.

It claims to be compatible with PCI Express Spezifikation Revision 3.0 (G42520-002). So even if UEFI wouldn't support management, the card has to boot itself.
I'm using this controller for quiet some time in various servers and never had any compatibility issues – most of these servers the controller was not designed for.
And even after turning the card into a M5016 it doesn't boot in that specific board. Smells like black listing – especially because of the comment from the 1st level ITP support, who answered my suspicion with "maybe".
OTOH, some kind of blacklisting could also be implemented controller side and for some reason I was lucky that all server systems I used so far are able to trigger the blacklist-killswitch on the controller.
I decided to omit other vendors because of their practice of interoperability blacklisting/branding and raising artificial problems. Now the reference designer does the same :-( probably I'm not thinking enterprized enough…

 

[gdclub]

@Devinirv The SBR is from a retail boxed MegaRaid 9270-8i and is 512 bytes, as is the SBP.

If you want to create an empty 512 byte SBR, from Linux: dd if=/dev/zero of=512byte.sbr bs=512 count=1

Past that, I'm not really sure what you're asking for. If you'd like a copy of the SBR I sent @BLinux, I'm more than happy to do that.

vanfawx my friend, I'm sorry for my disturbing that I'd like for a copy for the SBR files,
and the DeviceID/VendorID for stock 9270-8i.
I'm working for modding an IBM M5110 to an stock 9270-8i and made some progress.

 

[Freebsd1976]

vanfawx my friend, I'm sorry for my disturbing that I'd like for a copy for the SBR files,
and the DeviceID/VendorID for stock 9270-8i.
I'm working for modding an IBM M5110 to an stock 9270-8i and made some progress.

could you share your method or more information?

I tried but it fail, firmware indeed upgrade ,but card still show M5110 and flash 9270 SPR , but after reboot and dump SPR show it change back to orginal M5110 SPR.

 

[vanfawx]

@gdclub send me a PM with the email you want me to send it to.

 

[gdclub]

could you share your method or more information?

I tried but it fail, firmware indeed upgrade ,but card still show M5110 and flash 9270 SPR , but after reboot and dump SPR show it change back to orginal M5110 SPR.

It seems like lsiulti's needed
I should findout what's the difference between stock 9270 and M5110 by getting the SPRs and DeviceID(the VENxxxx_DEVxxxx for the hardware)

 

[Freebsd1976]

It seems like lsiulti's needed
I should findout what's the difference between stock 9270 and M5110 by getting the SPRs and DeviceID(the VENxxxx_DEVxxxx for the hardware)

what your card show when enter webbios after crossflash? m5110 or 9270-8i

 

[gdclub]

what your card show when enter webbios after crossflash? m5110 or 9270-8i

M5110
So I need the sbr to check the principle for the recovery