Date: Thu, 5 Mar 2015 16:50:31 +0100 From: "J.R. Oldroyd" <jr@opal.com> To: =?UTF-8?B?SmVhbi1Tw6liYXN0aWVuIFDDqWRyb24=?= <jean-sebastien.pedron@dumbbell.fr> Cc: freebsd-x11@freebsd.org Subject: Re: [Call for testers] DRM device-independent code update to Linux 3.8 (take #2) Message-ID: <20150305165031.3e835cfb@shibato> In-Reply-To: <54F76B88.9090902@dumbbell.fr> References: <54F636B3.90701@FreeBSD.org> <20150304172802.0b285585@shibato> <54F76B88.9090902@dumbbell.fr>
next in thread | previous in thread | raw e-mail | index | archive | help
--Sig_/HMf_DfDD8levLxVJQhcJz4G Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable On Wed, 04 Mar 2015 21:31:04 +0100 Jean-S=C3=A9bastien P=C3=A9dron <jean-se= bastien.pedron@dumbbell.fr> wrote: > > The panic you have can only happen when the CP initialization failed. In > this case, the driver does some cleanup, including an mtx_destroy(9). >=20 > The panic happens because the code tries to use the destroyed mutex. The > previous core.txt.6 made me think the code we added for vt(4) didn't > check the state correctly. But core.txt.7 shows this happens with a > regular ioctl as well. >=20 > The crash was fixed in Linux 3.10. I will import the patch. > Ok. =20 > > The pattern that's emerging here is that older (mid-2014) kernels > > boot and init fine, warm or cold. The CP init problem started > > somewhere mid-2014. >=20 > I can't reproduce it here. I still need to work with the informations > you emailed me already a month ago. >=20 > > BTW, I now also have the ring CP init failure on a second system > > that was just updated to 10.1-RELEASE-p6. It has ATI Radeon HD 4200 > > which is RS780 and it also shows the ring init failure at CAFEDEAD > > but in r600_ting_test(). Unfortunately, this system is a production > > one so I can't easily play with it. Note, no mtx_lock panic here. > > Fortunately, the graphics not working isn't a problem on this system. >=20 > What version of FreeBSD were you running before the update? >=20 It was actually at 10.1-RELEASE-p5 before, but I am not sure that's helpful. There were CAFEDEADs there, too. The admin tells me that that system is generally always updated as binary updates using freebsd-update. I.e., no local compilation. That's been the case for all updates in the last six months, in any case. The init failures there are only occasional, not every boot. In fact, looking at its logs, it boots OK about 11 out of 12 times or so. About one of 12 times, it gets a ring init failure. That seems to have been the case on -p5, too. I cannot see boot logs from before -p5, and not sure they still exist. I can tell that most of the ring init failures seem, just like on my laptop, to be after warm reboots. But I can also see at least one ring init failure that appears to be after a cold power-off boot. -jr --Sig_/HMf_DfDD8levLxVJQhcJz4G Content-Type: application/pgp-signature Content-Description: OpenPGP digital signature -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iEYEARECAAYFAlT4e0sACgkQls33urr0k4nX8QCbBirpqLLYqNy8LECdmw//fSQ+ Q9UAnjwiVDLufhjmdiqBSUql3/NspM/K =j/QH -----END PGP SIGNATURE----- --Sig_/HMf_DfDD8levLxVJQhcJz4G--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20150305165031.3e835cfb>