Post-map PE blue screen cause panel discussion (E820cycles parameters)

Post-map PE blue screen cause panel discussion (E820cycles parameters)
Http://bbs.znpc.net/thread-6182-1-5.html
No, Dot. Posted on 2011-12-8 11:42:31

As you know, the workaround solution to the blue screen is to use map--e820cycles=0. But that does not solve the problem at all.

This post hopes to explore the true technical causes of the blue screen and possible solutions through extensive user testing and usage experience.

Clarification is required for the following issues:

1. When did you start the blue screen? That is, when the factory machine, start the blue screen? The goal is to determine whether the old computer has a blue screen phenomenon.

2. is the blue screen related to hardware (motherboard, video card) or software (driver)? Guessing the blue screen is a bug caused by some graphics driver (this bug may be artificially made or not, this topic is not discussed here). The graphics driver may "detonate" the bug under some hardware condition, and the bug is not "activated" under other hardware. If this is the case, replacing the graphics driver should solve the blue screen problem.

I think the WinPE blue screen may be related to the Firadisk or Winvblock Drive--now the PE mostly use Firadisk drive, and my netbook with Firadisk do WIN2003 Ramos, also easy to appear blue screen, and Winvblock seems better , but it seems that few use Winvblock to do PE drive.

Seemingly starting from Intel's 5 series of chips, the collection or switchable video card notebook more often appear, is estimated to be related to Intel's HD graphics, but as for the driver or hardware, but also to understand the PE production of people to help test
What does the BIOS mean? That's not the only motherboard that has BIOS. The video card also has a BIOS. Those boards, which are plugged into the motherboard, may have their own bios, or they may affect the original BIOS on the motherboard. Together with the motherboard, they form a unified, final BIOS. This final BIOS is sometimes referred to as the motherboard bios for the sake of simplicity. The board is plugged into the motherboard, which affects the final BIOS. Therefore, in the above discussion, it is wrong to say that the BIOS is not a problem.

You can "rule out" the problem with the motherboard, but you cannot "exclude" the BIOS issue. "Motherboard" and "BIOS" are two different concepts, that is, they have "cross" each other, but not exactly the same.

According to your description, I also think the problem may be caused by the video card. But strictly speaking, any board card (including the motherboard) may have problems. That is, theoretically there is the possibility that the problem is fundamentally caused by the motherboard, but the video card is unfortunately a "victim" (it is wronged, it becomes the biggest "suspicion").

If someone can prove that this video card is plugged into any motherboard with the same problem, then this basically proves that the problem is really a graphics card.
There is growing evidence that Intel's graphics card is causing problems. We continue to give "proof".
If possible, you will be given the grub4dos command line (after directly entering the command line) to perform the display of the Displaymem command and to perform the display results of the ' map--status ' command.
I opened this post, the original purpose is only to explore the default--e820cycles=-1 in the case of blue screen problem, the result, everyone out of this range.

The original goal was simply to figure out the specific technical details that led to the blue screen. is to determine which driver is causing the crash and whether the problem can be resolved by replacing the problematic driver. This is the purpose of this post.

To clarify again:

When the--e820cycles is not-1, any situation, is normal, that is, the crash, blue screen, are normal phenomena, I think, this will not be repeated many times to explain it? And this is not the focus of this post.
Sony VAIO VPCSA dual graphics Notebook PC:
Map--e820cycles=-1 blue screen, change to map--e820cycles=3 Normal
Take a closer look at "framebuf", indicating that the cache is showing

The cause of the blue screen was previously clear and was caused by the Windows protected mode driver. ZHAOHJ situation, in the blue screen has been the wrong driver name listed, it is called Framebuf.

If you can replace this driver, you can avoid the blue screen.

Stop doubting Grub4dos's int15. Completely unrelated to this. zw2312914 take a closer look, grub4dos int15, basically can be said, it is impossible to make mistakes, this has been repeated many times the temper. You should be mistaken. You can see from the beginning to the end that you will not make such a low-level mistake.

Why does it have nothing to do with Grub4dos's int15? As already proved, after Grub4dos's int15 takes over the control, does nothing, immediately uses the JMP instruction to go to the original entrance in the ROM, still freezes. And only completely put the int15 interrupt vector directly into the ROM entry, do not crash. This is why you should add the E820cycles parameter.

That is, as long as the interrupt number is modified to point to the normal memory below the ROM (even if you immediately turn to the ROM entrance), on the problematic machine, will be dead. Otherwise, if not taken over by grub4dos, there will be no blue screen. The problem is very clear, and this is completely unrelated to the process of Grub4dos's int15.

So, don't get tangled up in this issue. There is no solution except to replace the driver.

There is evidence that the same machine, swap with Win7 everything is OK, this shows that the problem is not on grub4dos, and it is the fault of the XP driver.

The above are proved, there is no ambiguity in the place.

The DISPLAYMEM command does not show whether the space occupied by the map. In other words, the DISPLAMEM command always shows the original memory layout given by the system BIOS.

Memcheck shows the memory layout modified by map--hook.

Currently feeling larger than 4G of memory machine, map--mem--top/iso.iso (0xff) This form, will destroy the simulation disk;
Sony notebooks that had tested 8G of memory above, and before and after map--mem--top a virtual disk, were unsuccessful.
e820 above 4G is not protected?
----------------------
Looking for a desktop with 8G memory, map--mem--top successfully.
It seems to be a problem with dual graphics boards.
But the dual graphics motherboard of the machine, without adding--top can be successful. Perhaps 4G or more of the memory area is not a continuous space, there are holes in it, we did not detect it.
As you know, the workaround solution to the blue screen is to use map--e820cycles=0. But that does not solve the problem at all.

Post-map PE blue screen cause panel discussion (E820cycles parameters)