Hmm
Interesting find - if you keep the right SHIFT key pressed during power-on, you'll get the message "Option ROM will not be loaded".
This prevents loading of AHCI/RAID ROM.
But then, system can't boot off the RAID drive.
I cannot experiment with this more in the next few days, but maybe it would be possible to boot from the 3rd party SSD which is connected to the DVD SATA port then.
But it is questionable what would happen with the hardware detection then.
I guess the best method would be to install GRUB and patch it so it forces the SATA controller to AHCI (or RAID) mode just before OS boots.
-
I found what I need in Intel HM57 specs (see attached PIC)
So, the idea is:
1. Set the controller to IDE mode
2. Force boot from extra SSD (it is SATA #1, so it >should< work)
3. Load GRUB that is patched to poke the SATA controller and switch it back to RAID
Hopefull this will do the trick...
But it might just delay the trashing of the CPU tables....Attached Files:
-
-
http://www.intel.com/support/chipsets/iaa_raid/sb/CS-009335.htm - option ROM is required to be loaded and integrated into main ROM/BIOS if you want to use RAID. Ouch! -
Ok, finally a weekend
Here is the dump of SSDT ACPI tables - both OK and buggy
Looks like tables are not corrupted, they are perfectly OK, but... just wrong. For example, in case of the bug, T-states are defined even if disabled in BIOS, and PSS (performance state data) is in the different table (ssdt_2 instead of ssdt_4)
So the bug actually triggers some more complex chain of effects that render the SSDT ACPI tables different...Attached Files:
-
-
DAMN SONY #)@(&*
Apparantly, I managed to write code that switches SATA controller from IDE to RAID (or AHCI)... BUT...
It is impossible to boot from the RAID drive then (as expected, as it requires Option ROM) - I thought this can be simply avoided by booting from the 2.5" SSD which would then nicely boot Windows and load the RAID driver with it, making RAID drive accessible again...
Well, it appears that direct booting from the second SSD is just impossible - I am getting Operating System Not Found messages constantly even after fresh Windows install when built-in SSD is disabled.
Amazingly, system DOES boot if the USB DVD drive is plugged in with Windows setup and proper CD/DVD boot sector - which tells me that Sony really botched the BIOS and hardcoded HDD/SSD booting only from the SATA ports 0/1/3/4 - also, even in case of their HDD models, HDD is connected to SATA port 0 and not 2 which is a DVD drive connector.
So, as it seems the MBR/boot sector way of hacking the controller back to RAID/AHCI is not going to cut it. -
Annoying, indeed.
FYI my Z's SATA ports configuration is:
Port 0: SSD
Port 1: DVD
Port 2: SSD
Port 3: SSD
Port 4: SSD
So it's hardcoded to boot from port 0? Looks like we need the HDD cable. We have the part number but don't know where to get it. -
Yes, but with HDD cable you cannot use built-in SSD
Maybe, just maybe, there is some weird hardware limitation which also explains why is Sony using SATA0 even for the HDD model, instead of the natural and physically closest SATA1 (DVD) port.
I have one more thing I can try - I will write a small Windows driver that overwrites the ACPI tables completely... I'll see if I get some time to do this in the next week.
I personally don't like this solution as it will force us to re-dump / re-compile driver every time BIOS gets updated... assuming Sony doesn't fix this problem which is highly likely going to be the case -
OMG...
Well, the ACPI tables do get corrupted even if you run system software like e.g. Everest Ultimate or SiSoft Sandra!!!
But, that does not affect the P-states because they are pre-cached on boot, and later table corruption does not affect them.
However, this is the reason why brightness control stops working!!!
This is some seriously broken sh*t. I dunno what Sony did, but this is obviously much worse than the SSD problem. -
Did you have some idea to fix it ?
So this problem will occur only QuadSSD Model ? -
I fixed the bug!!!
I wrote a driver that patches the broken ACPI tables with the correct ones, and CPU scales down nicely now even with extra SSD
I will post the driver later... but I am afraid that it might require some user intervention - when the ACPI tables are somewhere else,, or different CPU
But I'll try to document it as much as I can -
PS. Does this drivers also fix the brightness problem ? -
Unfortunately, no - because ACPI table gets trashed "dynamically"
The only reason why running Everest/SiSoft Sandra/etc... does not break CPU power management is because that is loaded at the boot and cached in the driver.
And this is the reason why the patch works... But, if you run system software, it will trash the tables again, and brightness will stop working -
Bugfix is here:
http://forum.notebookreview.com/showthread.php?t=478621 -
So how can windows sleep-wakeup fix this problem. Does it load new acpi data when wakeup ? -
Probably EFI/BIOS does something as it is the first to get the control when waking up from sleep.
-
My z11 is the hdd-optioned one;
I was wondering, if I were to replace the HDD with a third-party SSD.
- Would it function normally, or might I have a possible bug simiar to the "Quad SSD + Extra SSD" bug? -
normally, zero problems
-
after reading this i think i will buy the simple 128GB (2x64GB) model and add my 160GB intel SSD..
would you put your system and boot from the sony SSD or the intel?
and is the lenovo slim bay still the way to go? or is the original sony HDD adapter from the no DVD model available? or better asked, whats the chapest way? -
-
Would it be possible to reconstruct the quad RAID setup as dual RAID?
New Z - serious problem with additional SSD/HDD
Discussion in 'VAIO / Sony' started by psyq321, Apr 16, 2010.