Re: [ros-dev] [ros-diffs] [sginsberg] 43958: - Fix the "fix" of un-protecting the 7 IDT entries on P5 in HAL's BIOS call code when setting the custom Invalid Opcode handler

[Dmitry Gorbachev]

Hi!

This change causes a crash when booting, on Pentium.

[Stefan Ginsberg]

Ok...so how does it crash? And where? Can I get a backtrace or something?
> Date: Sat, 7 Nov 2009 22:30:05 +0300
> From: [e-mail]
> To: [e-mail]
> Subject: Re: [ros-dev] [ros-diffs] [sginsberg] 43958: - Fix the "fix" of	un-protecting the 7 IDT entries on P5 in HAL's BIOS call code	when setting the custom Invalid Opcode handler
>
> Hi!
>
> This change causes a crash when booting, on Pentium.
>
> Ros-dev mailing list
> [e-mail]
> http://www.reactos.org/mailman/listinfo/ros-dev

_________________________________________________________________
Keep your friends updated—even when you’re not signed in.
http://www.microsoft.com/middleeast/windows/windowslive/see-it-in-action/soc...

[Dmitry Gorbachev]

> Ok...so how does it crash? And where? Can I get a backtrace or something?

First time, HalpBiosDisplayReset() is called when booting, before
KiI386PentiumLockErrataFixup(). Recent changes cause a page to be
protected from writing. Then kernel tries to write to this page and
crashes. Another time, HalpBiosDisplayReset() is called when shutting
down; before r39959, crash happened to be there. From r39959 and
before r43958, write protection was established only in
KiI386PentiumLockErrataFixup().

It appears that a fix for a fix for a fix is now needed:)

[Stefan Ginsberg]

Should be fixed as of 44032.
> Date: Mon, 9 Nov 2009 00:13:51 +0300
> From: [e-mail]
> To: [e-mail]
> Subject: Re: [ros-dev] [ros-diffs] [sginsberg] 43958: - Fix the "fix" of	un-protecting the 7 IDT entries on P5 in HAL's BIOS call code	when setting the custom Invalid Opcode handler
> > Ok...so how does it crash? And where? Can I get a backtrace or something?
>
> First time, HalpBiosDisplayReset() is called when booting, before
> KiI386PentiumLockErrataFixup(). Recent changes cause a page to be
> protected from writing. Then kernel tries to write to this page and
> crashes. Another time, HalpBiosDisplayReset() is called when shutting
> down; before r39959, crash happened to be there. From r39959 and
> before r43958, write protection was established only in
> KiI386PentiumLockErrataFixup().
>
> It appears that a fix for a fix for a fix is now needed:)
>
> Ros-dev mailing list
> [e-mail]
> http://www.reactos.org/mailman/listinfo/ros-dev

_________________________________________________________________
Keep your friends updated—even when you’re not signed in.
http://www.microsoft.com/middleeast/windows/windowslive/see-it-in-action/soc...

[Dmitry Gorbachev]

Ok, it works. Thanks. :)