From: Lars Erdmann <lars DOT erdmann AT arcor DOT de>
Newsgroups: comp.os.msdos.djgpp
Subject: Re: Hooking protected mode interrupts
Date: Thu, 30 May 2002 08:05:25 +0200
Organization: Arcor
Lines: 42
Message-ID: <3CF5C125.2E882E2E@arcor.de>
References: <3CF4EF43 DOT BC5325C9 AT arcor DOT de> <1022611296 DOT 485030 AT queeg DOT ludd DOT luth DOT se>
NNTP-Posting-Host: dialin-212-144-128-001.arcor-ip.net
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Trace: newsread1.arcor-online.net 25355 NG bGFycy5lcmRtYW5u
X-Complaints-To: abuse AT arcor-online DOT net
NNTP-Posting-Date: Wed, 29 May 2002 06:08:51 +0000 (UTC)
X-Priority: 1 (Highest)
X-Mailer: Mozilla 4.61 [de] (OS/2; U)
X-Accept-Language: de
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp
Reply-To: djgpp AT delorie DOT com

Hi,

I understand, that I have to lock my OWN code and data.
However, the wrapper function that I allocate via the below mentioned
functions are NOT MY code.
So my question was if those wrapper functions are locked by the API
calls themselves.
It's a problem for me to do that since (at least for
_go32_chain_protected_mode_interrupt_vector) I do not get the address of
the wrapper function returned and besides, I don't know its length.

But the other post from Eli Zaretskji clarified that the wrapper
function is locked by the APIs themselves, so there is nothing to worry.

Thanks for your help,

Lars

Martin Str|mberg schrieb:
> 
> Lars Erdmann <lars DOT erdmann AT arcor DOT de> wrote:
> : I have successfully hooked a protected mode interrupt.
> : Now I have read in some note, that the protected mode wrapper that is
> : obtained via _go32_dpmi_alloc_iret_wrapper (or so) or via
> : _go32_chain_protected_mode_interrupt_vector (or so) is not locked by
> : djgpp.
> : Since it is essential to lock all code and data accessed during
> : interrupt time, I wonder if I risk a trap in my app.
> 
> Yes.
> 
> : Has this been fixed/changed in the newest release, maybe ?
> 
> There's nothing to fix.
> 
> : If not, what is the workaround ?
> 
> To lock it (ummm, I must misunderstand you because this is obvious)!
> 
> Right,
> 
>                                                 MartinS