04-19-2014, 06:13 PM (This post was last modified: 04-19-2014, 06:14 PM by Brabbelbla.)
@kasar
Yes, you were right. Cut and insert. You appear to still have the suppress statement for the RAID entries edited, when they shouldn't be anymore. The suppress condition originally was the SATA Operation variable not having value 0x2, then edited to suppress if the value is 0x2, being as we know now, the RAID option. So you should revert that change, the original condition was correct.
Posts: 523
Threads: 0
Joined: Aug 2011
Reputation:
23
04-19-2014, 06:23 PM (This post was last modified: 04-19-2014, 06:27 PM by kasar.)
haha, well, I want to get the raid feature working and relase the v6 version of my L502X bios modifications
I made some changes since the v5, specially in the advanced security module, bringing up new options and like modify the asset tag or get rid of the computrace option by just switching options.
raid feature will be of course something nice to have and keep people from the nbr forums happy
thank you for all the help Brabbelbla
Posts: 88
Threads: 0
Joined: Feb 2013
Reputation:
6
04-27-2014, 04:47 AM (This post was last modified: 04-27-2014, 02:43 PM by follow_me.)
Sorry for a delay, had a long backup/restore cycle
I have tested the latest firmware :
- RAID mode works, I was able to create stripe array with 2 disks
- Legacy boot wortks, was able to boot Linux from this disk
- UEFI boot works , was able to boot UEFI Shell from this disk and it's fs was mapped
- TRIM is NOT WORKING , tested by removing file, forcing Linux to trim filesystem and then reading physical sectors of underlying SSD , and there was data,not zeroes as it shold be with trim working
* UPDATE *
as it turned out TRIM should be tested with the latest WINDOWS driver , I will make some addition tests to clarify weather TRIM working or not
To be able to boot from Raid disk it is required to add boot record using bcfg
for the RAID mode device handle is
UnknownDevice ExtendedSCSIPassThru EfiAtaPassThruProtocolGuid UnknownDevice UnknownDevice PCIIO DevicePath(PciRoot(0x0)/Pci(0x1F,0x2))
and there is one small issue - I have 3 ssd drives , one of whitch is replaced stock DVD-Drive , and I can't find a handle of it in RAID mode, may be it was just renamed to something else
so - most features are working with a little bcfg magic , TRIM, unfortunately, not
I will be glad to answer any questiong you have, or make some additional tests if neeeded
BTW: @brabbelbla do you have any other thoughts about TRIM to be enabled ? Can I have a FW with a fancy RAID menus ?
04-29-2014, 09:48 AM (This post was last modified: 04-29-2014, 09:58 AM by Brabbelbla.)
@follow_me
Thanks very much for testing. It is not really clear whether you have tested TRIM with the Windows drivers now, but I guess not.
About the 3rd SSD: with me (with the actual optical drive still in place), when I enable RAID mode it does not get an SCSI device adress but remains reported as SATA. Is that the case with you, too? Was it visible in the RAID OROM? Anyway, I don't think there is much to be done. There are no hidden settings for the SATA ports and the values for the settings that are there are the same as those for ports 0 and 1.
No other thoughts on TRIM. What do you mean with 'fancy RAID menus'? A proper UEFI driver? Maybe it is possible, I have seen some Lenovo notebooks with a UEFI RAID driver. Maybe, with some magic, you could get that one integrated in our firmware. If you are interested, I think it was the Lenovo T530. Phoenix SCT as well, but I think it is UEFI 2.31.
It is actually quite easy to replace the module yourself, using PhoenixTool, but I was curious so I did it anyway. It works, the driver is loaded, firmware volumes mapped. BUT THE CRAZY BUGGY EDITED SHELL WOULDN'T LET ME ADD SOMETHING AS A BOOT OPTION. Raaaahhh. It would not find the file. Nor let me MANUALLY add the boot variable for the internal shell. I tried this, the final part is a long series of hex values:
Where I took the diagnostics entry from the nvram dump and edited the text and the GUID. But it said it was the incorrect format or whatever. Cutting the data to 00010000180049006E didn't give the error, so I guess there is some limit as to how long the hex data can be. Grrr. The unedited but not so useful unedited old-style shell would run applications fine from the FV.
Of course, the good news is that replacing the driver works and it should enable you to do what you wanted to do.
04-29-2014, 02:49 PM (This post was last modified: 04-29-2014, 03:12 PM by follow_me.)
@brabbelbla tested with linux driver, not Windows one , will test it this weekend
about 3rd ssd - it is visible in the RAID OROM and can be added to array , but I dont want to do it because it is conected to slow (sata2) port
It is doesn't mapped as sata drive and I cant find it's handle
did you tried to add with bcfg ? or only with setvar ?
What I want to propose
I can build modified version of ffs driver with built in bcfg command which will add custom menu entries (buil in console and any other we want) and modify nvram to enable UEFI boot at start
test builds will be available this friday, I think
I propose to replace built in shell with a custom one
Any suggestions are welcomed
I have a shell where -p is for paggination
Posts: 81
Threads: 0
Joined: Jun 2012
Reputation:
11
04-29-2014, 03:26 PM (This post was last modified: 04-29-2014, 03:44 PM by Brabbelbla.)
I tried both. bcfg would not find the file, setvar would not add such long hex data. The modified shell wouldn't even run applications from the FV, while the unmodified old-style shell would. Of course, you seem to have a different modified shell, so things may work for you. Could you share yours? I will at least give you the one I use, it was edited by... dzamar or Slice, I think over at projectosx.
I will look at the UEFI RAID driver later.
BTW, I would love the pre-OS environment to be displayed in native resolution. If cou can manage that with your custom driver, sooner or later, I will be forever grateful
Posts: 88
Threads: 0
Joined: Feb 2013
Reputation:
6