DMARC-Filter: OpenDMARC Filter v1.4.2 sourceware.org DD66C3858C53
MIME-Version: 1.0
References: <b7a37aeabe4ce0df6b33f55088a28150@tiscali.it>
In-Reply-To: <b7a37aeabe4ce0df6b33f55088a28150@tiscali.it>
Date: Wed, 14 Jun 2023 16:48:41 +0200
Message-ID: <CAB8Xom8c6V5cc-54a5tjxxGaR-CO4nTCs23yo3KkUPWTJrO8sg@mail.gmail.com>
Subject: Re: fork bug in cygheap
To: paolo.zambotti@tiscali.it
Cc: cygwin@cygwin.com
From: marco atzeri via Cygwin <cygwin@cygwin.com>
Reply-To: marco atzeri <marco.atzeri@gmail.com>
Content-Type: text/plain; charset="utf-8"
Sender: "Cygwin" <cygwin-bounces+archive-cygwin=delorie.com@cygwin.com>
Content-Transfer-Encoding: 8bit

On Wed, Jun 14, 2023 at 2:20 PM Paolo via Cygwin <cygwin@cygwin.com> wrote:
>
>   Hi all,
>
> I just found a bug (I think) in cygheap_fixup_in_child
> function (cygheap.cc).
> the effect is that, sometimes a command from a
> script ends with the following error message
>
> 0 [main] openssl (17784)
> child_copy: cygheap read copy failed, 0x0..0x8003D8538, done 0, windows
> pid 17784, Win32 error 299
>
> Digging a little bit in the code I found
> that commit_size in cygheap_fixup_in_child is wrongly calculated if the
> condition (child_proc_info->cygheap_max > (void *)
> CYGHEAP_STORAGE_INITIAL) is met.
>
> Currently, in that case, its value is
> evaluated as
>
> commit_size = allocsize
> (child_proc_info->cygheap_max);
>
> but I think this is wrong and it should
> be
>
> commit_size = allocsize (child_proc_info->cygheap_max) -
> CYGHEAP_STORAGE_LOW;
>
> The wrong value avoids the successful execution of
> the following second VirtualAlloc and the the error in child_copy
> function.
>
> I tried to recompile the cygwin dll with the above change and
> indeed the error disappears.
>
> I'm I correct?
> Paolo.
>

it looks like
https://cygwin.com/pipermail/cygwin-developers/2023-April/012620.html

-- 
Problem reports:      https://cygwin.com/problems.html
FAQ:                  https://cygwin.com/faq/
Documentation:        https://cygwin.com/docs.html
Unsubscribe info:     https://cygwin.com/ml/#unsubscribe-simple