Subject: IPL_LOCK oddity
To: None <tech-kern@netbsd.org>
From: Bill Studenmund <wrstuden@netbsd.org>
List: tech-kern
Date: 08/23/2006 17:08:55
--doKZ0ri6bHmN2Q5y
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable
In chatting with dbj, I was reminded that on a number of ports, we have=20
IPL_SERIAL higher than IPL_LOCK. IPL_IPI also is higher than IPL_LOCK on=20
x86 (I didn't check extensively).
The problem is that our serial port code, in an effort to be MP-safe,=20
takes locks. Including the code that runs at IPL_SERIAL. Thus IPL_SERIAL >=
=20
IPL_LOCK is very very bad. :-)
So is there any reason we don't just change IPL_LOCK to IPL_SERIAL?
Also, is the IPI code structured so that it doesn't have to take locks?
Take care,
Bill
--doKZ0ri6bHmN2Q5y
Content-Type: application/pgp-signature
Content-Disposition: inline
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.3 (NetBSD)
iD8DBQFE7O4XWz+3JHUci9cRAqi1AJ9P9qP2O/ZTfg84OWQrysN0XLekrwCfUmWR
3Nw3x6FZB5SY43r4CTajcc0=
=pCw+
-----END PGP SIGNATURE-----
--doKZ0ri6bHmN2Q5y--