CVS commit: [netbsd-6] src/sys/arch/amd64

["Julian Coleman" <jdc%netbsd.org@localhost>]

Module Name:    src
Committed By:   jdc
Date:           Sun Jun  3 21:45:10 UTC 2012

Modified Files:
        src/sys/arch/amd64/amd64 [netbsd-6]: trap.c vector.S
        src/sys/arch/amd64/include [netbsd-6]: frameasm.h

Log Message:
Pull up revisions:
  src/sys/arch/amd64/include/frameasm.h revision 1.17-1.19
  src/sys/arch/amd64/amd64/vector.S revision 1.40-1.41
  src/sys/arch/amd64/amd64/trap.c revision 1.71
(requested by dsl in ticket #280).

Move all the XEN differences to a single conditional.
Merge the XEN/non-XEN versions of INTRFASTEXIT and
  INTR_RECURSE_HWFRAME by using extra defines.
Split INTRENTRY so that code can insert extra instructions
  inside user/kernel conditional.

Add a ';' that got deleted in a slight tidyup.

Rejig the way TRAP() and ZTRAP() are defined and add Z/TRAP_NJ() that
  excludes the 'jmp alltraps'.
Use the _NJ versions for trap entries with non-standard code.
Move all the KDTRACE_HOOKS code into a single block inside the
  IDTVEC(trap03) code. This removes a mis-predicted from every
  trap when KDTRACE_HOOKS are enabled.
Add a few blank lines, need some comments as well :-)
No functional changes intended.

Let the user of INTRENTRY_L() place a label on the 'swapgs' used
  when faulting from user space.

If we get a fault setting the user %gs, or on a iret that is returning
to userspace, we must do a 'swapgs' to reload the kernel %gs_base.
Also save the %ds, %es, %fs, %gs selector values in the frame so
they can be restored if we finally return to user (probably after
an application SIGSEGV handler has fixed the error).
Without this any such fault leaves the kernel running with the wrong
%gs offset and it will most likely fault again early in trap().
Repeats until the stack tramples on something important.
iret change works, invalid %gs is a little harder to arrange.

Treat traps in kernel mode during the 'return to user' iret sequence
as user faults.
Based heavily in the i386 code with the correct opcode bytes inserted.
iret path tested, arranging for segment register errors is harder.
User %fs and %gs (32bit apps) are loaded much earlier and any errors
will generate kernel panics - there is probably code to try to stop
the invalid values being set.


To generate a diff of this commit:
cvs rdiff -u -r1.69 -r1.69.2.1 src/sys/arch/amd64/amd64/trap.c
cvs rdiff -u -r1.38.8.1 -r1.38.8.2 src/sys/arch/amd64/amd64/vector.S
cvs rdiff -u -r1.16 -r1.16.8.1 src/sys/arch/amd64/include/frameasm.h

Please note that diffs are not public domain; they are subject to the
copyright notices on the relevant files.