ZitatCPU in C0 power state 50-60% of time even without cpu load, unresolved
I think because cpu0cst acpi table declare only MWAIT way to halt processor, old HLT doesnt work If i disable mwait cpu's support in bios, even win7 cannot enter c1+ states
Zitat von XPLives im Beitrag #189This only could be used for installing XP onto one machine but not migrate. After XP is installed in SATA AHCI on SkyLake you then move the hard drive / SSD to the Ivy Bridge computer and try booting in SATA AHCI mode it should do BSOD 7B or a Reboot Loop.
why you need install first on skylake and then copy on ivy ? why not install on ivy natively ? there are too many diff chipsets (non-intel including), where os will get bsod
Zitat von XPLives im Beitrag #185 This method ( pci\cc_0106) I also tried this a few weeks ago before stopping the tests to emulate how Vista method works but also failed. The reason is CC_0106 is not visible to XP because it is a "Compatible IDs" and the driver only sees "Hardware IDs" in XP during boot. Since Hardware IDs are not common among all the Intel SATA Controllers you cannot do a global detection. That is why Vista can do it natively since CC_106 exists on all Intel SATA controllers and Vista+ is programmed to accept "Compatible IDs" not just "Hardware IDs". If there is a way to brute force XP code to accept "Compatible IDs" for Boot storage device than this will work for all Intel SATA and can be made for all AMD SATA as well.
one of compatible id for intel works in xp without any hacks, my setup in TXTSETUP.SIF (windows xp sp1 32 bit): [HardwareIdsDatabase] PCI\VEN_8086&CC_0104 = "IaStor" PCI\VEN_8086&CC_0106 = "IaStor" tested on ivybridge and skylake, all works well, setup phase see all disks
Zitat von XPLives im Beitrag #184@Mov AX, 0xDEAD Is it more than just the code that has been removed in SkyLake+?
Why does this still work on Z87 and Z97 and not in Z170 if you understand it well enough to explain in more detail than good luck response which isn't helpful?
Intel's PCH specification on 100+ series chipsets: The PCH SATA controller support two modes of operation, AHCI mode using memory space and RAID mode. The PCH SATA controller no longer supports IDE legacy mode using I/O space. Therefore, AHCI software is required.
IviBridge/SandyBridge/etc: Standard ATA Emulation The PCH contains a set of registers that shadow the contents of the legacy IDE registers. The behavior of the Command and Control Block registers, PIO, and DMA data transfers, resets, and interrupts are all emulated.
Zitat von XPLives im Beitrag #175One day Intel USB 4.0 will arrive and the same USB installers will fail or the authors of these USB installers no longer care about XP.Most of these USB installer tools don't even work inside XP anymore and require Windows 7+ to create the USB install image
maybe is it time to check how work these installers ? they will still work with usb 4 and usb 5 too and still installing xp.
Zitat von XPLives im Beitrag #175https://www.win-raid.com/t4092f45-Guide-How-to-install-Windows-XP-OS-on-Modern-Hardware.html
#1 -- Internal SATA optical disc drive - CD, DVD, or Blu-ray #2 -- CD-RW disc for nLite slipstreaming of the SATA AHCI driver into your XP burned install disc #4 -- Desktop motherboard with a rear PS/2 physical port #5 -- PS/2 keyboard you are crazy, for installing window xp need nothing of these, why people buy that ?
p.s. want ide compatibilty bios mod on ahci-only hardware (skylake+)? good luck )
There many unresolved problems running Windows XP x32 on sky/kaby/coffee lake hardware my own list: - ACPI BSOD: .......... 0xA5 (0x03, X, 0xС0140001, Z) ACPI 2.0 syntax, 99% solved (custom acpi.sys), 1% - "QWordAcc" still 32bit only .......... 0xA5 (0x02, X ,Y ,Z) error in ACPIRangeValidatePciResources, PCI vs E820 list conflict, solved (custom acpi.sys) .......... 0xA5 (0x11, 0x08 ,Y ,Z) error in ACPIInitializeDDB, solved (custom acpi.sys) .......... other possible BSODs, unresolved - all integers in original ACPI 1.0 XP are 32 bits, ACPI 2.0 requires all ints to be 64 bits, unresolved. Acpi.sys from Longhorn v.5048 has 64 bit integers and compatible win Winxp - Broken acpi pm_timer, solved (custom hal.dll, forced to use TSC counter) or enable hidden bios option "TCO Timer" - Acpi tables patcher at boot time like Chameleon/Clover, solved,( http://reboot.pro/topic/22051-acpi-patcher-at-boot-time) - CPU in C0 power state 50-60% of time even without cpu load, solved (custom intelppm.sys, 1) only C1 c-state or 2) C1/С7s/C10 c-states) - USB 3.x drivers for integrated intel chipsets, solved (patched AMD USB3 driver) - Resource manager show conflict VGA card with ACPI IOTR/IOTraps device, solved 1) manual patching dsdt or 2) custom acpi.sys - UEFI x64 Boot, partial solved for WinXP x64 only (bootmgr from vista beta+vesa vga driver) - NVMExpress drivers, partial solved (modified OFA 1.3/1.5 driver+storport.sys from windows 2003) - Access to additional 3TB+ Disks/GPT partitions, solved (Paragon GPT Loader) - Boot from GPT partitions, unresolved (require uefi boot support first) - TRIM support in NTFS/Storport/??? driver unresolved
Zitat von XPLives im Beitrag #169This will make XP SATA installation without the need of AHCI slipstream first method. Laptops could use this to install XP lacking an optical disc drive requirement. This will solve the headaches many are experiencing since USB installation on SkyLake+ can be problematic and fail
maybe is it time to forget ancient methods and requirements ? today no need prepare media for installing windows in "MS Way", we have modern&fast equalents to install xp (easy2boot, firadisk, winntsetup, yumi, ...). installing from usb3 disk, installing on ahci controller - no problem
Zitat von Fernando im Beitrag #1854Don't forget to give us a feedback regarding the result you got with these drivers.
Original intel's usb3 driver(at least 5.0.4.43) hardcoded to "know" DEV_ID, so DEV=A63D and any "unknow" driver will eject feedback is simple - "device can't start, code xx"
Zitatz390 USB host driver is open source and can be ported easily. the problem is to get signature to not disable windows protection
Fake. No one manufacturer has published a source codes for usb controllers at win platfrom. MS created driver for win8+ based on usb 3.0 specification, but with vendor "quirks" (see linux usb3 sources), 3.1 is compatible with 3.0, so generic driver works with 3.1 too, but maybe not optimal
Zitat von diderius6 im Beitrag #148 And maximal unclear is for me, why I have no success with ACPI on my Asrock z370 board. The bios DSDT table is nearly one by one like the Bios DSDT table from gigabyte H110M-S2H. May be, another device is called, where until now I do not prevent it from reserved words from Acpi 2.0,
New syntax will generate A5(03,xxxx,С014001,yyyy) error, if you have other errcodes this not directed linked to new syntax I dont know much more about res conflicts at acpi level, but acpi 2 has important difference comparing to 1.1b - Integer implemenation. All integers in acpi 2.0 are 64 bit, so we can define QWORD fields and read/write to it from integer variables. In WinXP integers are 32 bit and it support only one "64 bit" QWordMemory instruction and i dont understand how this OS set/read 64 bits fields inside this structure.
Zitat von diderius6 im Beitrag #148@Mov AX, 0xDEAD I wonder, if another table for example SSDT in Bios has to do with the pm_timer.
acpi specs dont say anything about controling timer in acpi tables, timer is hardware level OS can only: 1) route interrupt to itself(SCI mode) or to bios (SMI) 2) disable or enable interrupts
p.s. because win7+ use TSC and we all have "right" cpus, no need anymore to fix this buggy timer, just patch os/bios to avoid using it
Some info about bug with Acpi Timer on Skylake+ platforms: - XP preSP2 uses TSC, not affected - XP SP2,SP3 uses acpi timer, affected - Vista search for HPET Timer and uses it, frequency is 14.31818 MHz (4 x freq of Acpi timer), not affected - Vista uses acpi timer if didn't find HPET, affected - Win7+ check quality of processor counter (TSC), skylake+ cpus have constant rate counter, so these OSes use TSC and not affected. Unlike XP, processor raw frequency divided to 1024, for example 3600Mhz/1024=3.579545Mhz (avoid arithmetic 32bit overflow in old-gen apps ?) - If Win7+ forced to not use TSC (bcdedit.exe /set useplatformclock true), it search for HPET and uses it, not affected - If Win7+ forced to not use TSC and didn't find HPET, it uses acpi timer, affected
Regarding acpi specs and intel specs on pch chipset, acpi timer generates hardware SCI interrupt (IRQ #9) if TMR_EN bit is SET. This bit is SETted on live system, but windows doesnt receive any interrupts. Near to TMR_EN exist PWRBTN_EN bit (generates SCI at power button press), and if we press button, windows receive interrupt (surprise!)
I came to the conclusion that this is a bug at the platform level. Maybe bios doesnt set some hardware regs or intel forget to say something about additional requirements, who knows....