DomainKey-Signature: a=rsa-sha1; c=nofws; d=sourceware.org; h=list-id
	:list-unsubscribe:list-subscribe:list-archive:list-post
	:list-help:sender:subject:to:references:from:message-id:date
	:mime-version:in-reply-to:content-type; q=dns; s=default; b=XWhx
	wkJ/6oS1aTlTYsRt0ioOnawLCn4dwxDjHHOMrs2tBwWyeIWbdEFMk3KWzzbt+bSE
	Ggx/Q9BM28MnKbxl2KqW0rf8dWJiUPAQutYMWfYin5EvpGwVJiY7V3L2IrgeckcV
	c2e5evP4TBcEA+rjtNrSHSoGhYyYwQVNWsRA5wY=
Mailing-List: contact cygwin-help@cygwin.com; run by ezmlm
Sender: cygwin-owner@cygwin.com
Mail-Followup-To: cygwin@cygwin.com
Subject: Re: Is async-signal-safety applicable to Cygwin?
To: cygwin@cygwin.com
References: <f80444ec-69b9-0177-dd75-c994092380fd@gmail.com>
From: Eric Blake <eblake@redhat.com>
Openpgp: preference=signencrypt
Message-ID: <9e5ce156-19c2-8379-6e29-8997c1700d5b@redhat.com>
Date: Tue, 19 Mar 2019 09:35:19 -0500
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101 Thunderbird/60.5.1
MIME-Version: 1.0
In-Reply-To: <f80444ec-69b9-0177-dd75-c994092380fd@gmail.com>
Content-Type: multipart/signed; micalg=pgp-sha256; protocol="application/pgp-signature"; boundary="7V9SnJ5MdHGL2Mu7723BekRxN10c6iJI5"

--7V9SnJ5MdHGL2Mu7723BekRxN10c6iJI5
Content-Type: multipart/mixed; boundary="BIemsF0heglGvAWvjVab8Ih3xZ3pT2odR";
 protected-headers="v1"
From: Eric Blake <eblake@redhat.com>
To: cygwin@cygwin.com
Message-ID: <9e5ce156-19c2-8379-6e29-8997c1700d5b@redhat.com>
Subject: Re: Is async-signal-safety applicable to Cygwin?
References: <f80444ec-69b9-0177-dd75-c994092380fd@gmail.com>
In-Reply-To: <f80444ec-69b9-0177-dd75-c994092380fd@gmail.com>


--BIemsF0heglGvAWvjVab8Ih3xZ3pT2odR
Content-Type: text/plain; charset=utf-8
Content-Language: en-US
Content-Transfer-Encoding: quoted-printable

On 3/19/19 9:28 AM, LRN wrote:
> ASS[0] is documented for Linux (and for POSIX) but i'm failing to find any
> information on how this applies to Cygwin. I assume, since there are no "=
real"
> signals on Cygwin, that it's not possible for a signal handler to break t=
he
> state of some function that happened to be executed when the signal handl=
er was
> triggered. But i could be wrong.

You are wrong. It IS applicable. Cygwin has signals and signal handlers;
although Windows may not have POSIX signals as such, it DOES have the
ability to interrupt a process mid-function when a timer expires, and
cygwin's implementation of signals on top of the semantics that windows
provides can indeed result in interrupting the middle of a
non-async-safe function.  Hence, a signal handler that calls a
non-async-safe function at the same time that the signal has interrupted
another non-async-safe function is indeed observable on Cygwin and can
indeed result in deadlocks (a classic example being the case if you
malloc() from a signal handler that interrupted an ongoing malloc() call).

--=20
Eric Blake, Principal Software Engineer
Red Hat, Inc.           +1-919-301-3226
Virtualization:  qemu.org | libvirt.org


--BIemsF0heglGvAWvjVab8Ih3xZ3pT2odR--

--7V9SnJ5MdHGL2Mu7723BekRxN10c6iJI5
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: OpenPGP digital signature
Content-Disposition: attachment; filename="signature.asc"

-----BEGIN PGP SIGNATURE-----

iQEzBAEBCAAdFiEEccLMIrHEYCkn0vOqp6FrSiUnQ2oFAlyQ/icACgkQp6FrSiUn
Q2pSVQgAj2HYJjXqfmgL+BZnT/hpBr0RD1N0Vd+u4mofGq8k1n8aCSa5P8Qyf6Op
o8S/kQ5hAo6VCsQuxW+pW4zWvxO10wPzN9b3C410Ao3PPUsStqhAzSDP60T+b2Nx
nPDd9BhiQjNNXyYCHWkt2Q3x5gkO2i51jKVIDJ6DAp/L+m0Yn/7zVV5CpY1w+fZ9
nf8ZSlOsezbV9m4EcVR0p4VCLA9qTitfGUNK+zZNbQeZLXcyVCXZ9NgH/ZeTo41w
lqr4x/idcokR31EnYJHmwuC89SzXwkgtv8+L4Fy7VWkvl2JqeEE1vGR5oEK89ZO2
Dk64YTCwAXOvfugN7Tg72agogyDe5A==
=LeYy
-----END PGP SIGNATURE-----

--7V9SnJ5MdHGL2Mu7723BekRxN10c6iJI5--