Date: Sat, 27 Nov 1999 18:57:28 -0800 (PST) From: Matthew Dillon <dillon@apollo.backplane.com> To: David Greenman <dg@root.com> Cc: Julian Elischer <julian@whistle.com>, current@FreeBSD.ORG Subject: Re: Which is the truth? (sycalls and traps) Message-ID: <199911280257.SAA40292@apollo.backplane.com> References: <199911252215.OAA14800@implode.root.com>
next in thread | previous in thread | raw e-mail | index | archive | help
:>Am I also right in assuming that all the registers that the user was
:>running when they did the KERNCALL have been saved on the KERNEL stack by
:>the time that the above routines are called?
:
: No, that's what the pushal (push all) does.
:
:-DG
:
:David Greenman
pushing/popping integer registers doesn't really take a whole lot of
time. The real overhead is with all the junk we have on entering and
exiting from a kernel context from/to a user context.
In fact, for kernel threads it's even easier: If you are calling
the switch code synchronously then, of course, the scratch-on-call
registers need not be saved or restored either. It's as simple as
the code I've shown below, taken from one of my other projects:
pushfl
pushl %ebx
pushl %ebp
pushl %esi
pushl %edi
pushl $myresumefunction
movl %esp,ta_Ctx+mc_SP(%ecx)
Switch to new task:
movl ta_Ctx+mc_SP(%ecx),%esp
ret
And the restore function:
popl %edi
popl %esi
popl %ebp
popl %ebx
popfl
...
Enhance as required (i.e. other registers need to be added if an
interrupt, segment registers, and so forth.
note: pushal/popal does not mess with the segment registers or
the FP registers. It is not all-inclusive. Only the main registers
are pushed/popped.
-Matt
Matthew Dillon
<dillon@backplane.com>
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199911280257.SAA40292>