From owner-freebsd-drivers@FreeBSD.ORG Tue Mar 8 18:46:30 2011 Return-Path: Delivered-To: freebsd-drivers@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 0CB3E106566C for ; Tue, 8 Mar 2011 18:46:30 +0000 (UTC) (envelope-from weiler@soe.ucsc.edu) Received: from mail-01.cse.ucsc.edu (mail-01.cse.ucsc.edu [128.114.48.32]) by mx1.freebsd.org (Postfix) with ESMTP id EC2058FC13 for ; Tue, 8 Mar 2011 18:46:29 +0000 (UTC) Received: from wraith.cse.ucsc.edu (wraith.cse.ucsc.edu [128.114.56.35]) by mail-01.cse.ucsc.edu (Postfix) with ESMTPSA id A298E1009C2C; Tue, 8 Mar 2011 10:46:29 -0800 (PST) Message-ID: <4D767985.90006@soe.ucsc.edu> Date: Tue, 08 Mar 2011 10:46:29 -0800 From: Erich Weiler User-Agent: Thunderbird 2.0.0.24 (X11/20100318) MIME-Version: 1.0 To: Jonathan Stewart References: In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: freebsd-drivers@freebsd.org Subject: Re: MFI Driver Behavior X-BeenThere: freebsd-drivers@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Writing device drivers for FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 08 Mar 2011 18:46:30 -0000 This was not resolved, although I tried in vain to get it to be more stable for over a month. The Perc H700 and H800 controllers do not work reliably with any version of FreeBSD. It's an MFI driver issue, such that the driver locks the controller temporarily under heavy disk load, but there seems to be little interest in fixing it. I thought at first it was a C-states issue, but then it locked again a week later. I punted and got an Adaptec controller, which is better supported under FreeBSD. Maybe they will make the MFI driver work better with these controllers one day, but I don't know when that will happen. Which is a bummer because more and more folks are buying these Perc H700/H800 controllers from Dell. On 03/08/11 10:38, Jonathan Stewart wrote: > [Please CC me I'm not on this list] > > Erich Weiler wrote: >>> We are testing a new theory. We believe that the C-states and C1E >>> control on this chipset (in the Dell R610), which are >>> configurable in the BIOS, are affecting the controller in some way. >>> We are disabling them and trying again. I'll post results soon. >> After disabling C-states and CIE in the BIOS on our R610, the Perc >> H800 controller now appears stable. My ZFS filesystem appears >> stable. Life seems good. >> >> I wonder what those items in the BIOS are doing in relation to the >> kernel to cause this. > > Have you had any further issues with this? I just updated 2 of my systems to > FreeBSD ???.net 8.2-STABLE FreeBSD 8.2-STABLE #2 r219138: Tue Mar 1 > 13:42:29 EST 2011 root@???.net:/usr/obj/usr/src/sys/SERENITY > amd64 > > from 8-stable sometime around 8 release and just had to have the > datacenter hit the reset button to bring the machine back. > > I'm reluctant to take this machine to current but if that fixes the problem > I can do it. The other option is calling the data center and having them > change the BIOS settings for me but that is also something I would rather > not do. > > Thanks, > Jonathan From owner-freebsd-drivers@FreeBSD.ORG Tue Mar 8 18:51:35 2011 Return-Path: Delivered-To: freebsd-drivers@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 307B21065676 for ; Tue, 8 Mar 2011 18:51:35 +0000 (UTC) (envelope-from cforgeron@acsi.ca) Received: from mta02.eastlink.ca (mta02.eastlink.ca [24.224.136.13]) by mx1.freebsd.org (Postfix) with ESMTP id EB2878FC08 for ; Tue, 8 Mar 2011 18:51:34 +0000 (UTC) MIME-version: 1.0 Content-transfer-encoding: 7BIT Content-type: text/plain; CHARSET=US-ASCII Received: from ip02.eastlink.ca ([unknown] [24.222.39.20]) by mta02.eastlink.ca (Sun Java(tm) System Messaging Server 7.3-11.01 64bit (built Sep 1 2009)) with ESMTP id <0LHR00E8G5PXJMC1@mta02.eastlink.ca> for freebsd-drivers@freebsd.org; Tue, 08 Mar 2011 14:51:33 -0400 (AST) X-CMAE-Score: 0 X-CMAE-Analysis: v=1.1 cv=KTm0Zd9FEOPabIM+9UH3RcOan8ug7Hn5ok0IYj/PkF8= c=1 sm=1 a=kj9zAlcOel0A:10 a=6I5d2MoRAAAA:8 a=0k8TjyC4y9tONcY83LYA:9 a=uy3iOMVH7-oqCVua-Z8A:7 a=-nhaIlfeOuAPfuYS4JZxKIXGcU4A:4 a=CjuIK1q_8ugA:10 a=SV7veod9ZcQA:10 a=Y4g+zi6NJtbRuBVJrbSZ6Q==:117 Received: from blk-222-10-85.eastlink.ca (HELO server7.acsi.ca) ([24.222.10.85]) by ip02.eastlink.ca with ESMTP; Tue, 08 Mar 2011 14:51:33 -0400 Received: from server7.acsi.ca ([192.168.9.7]) by server7.acsi.ca ([192.168.9.7]) with mapi; Tue, 08 Mar 2011 14:51:33 -0400 From: Chris Forgeron To: Erich Weiler , Jonathan Stewart Date: Tue, 08 Mar 2011 14:51:32 -0400 Thread-topic: MFI Driver Behavior Thread-index: AcvdwS4kZMmx1F9fTdKrDf56N0ct5wAAA21g Message-id: References: <4D767985.90006@soe.ucsc.edu> In-reply-to: <4D767985.90006@soe.ucsc.edu> Accept-Language: en-US Content-language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: acceptlanguage: en-US Cc: "freebsd-drivers@freebsd.org" Subject: RE: MFI Driver Behavior X-BeenThere: freebsd-drivers@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Writing device drivers for FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 08 Mar 2011 18:51:35 -0000 This may be of moderate use, but I have the exact same issue with the Dell PERC H700 and FreeBSD. I've tried 8.1, 8.2, and 9.0-CURRENT with similar results as you have had. I eventually gave up, and will reuse these in a Windows box, or I'll sell them on ebay. However, I haven't had any issues with a LSI1068 based controller that uses the same mfi driver. I'm also having good luck with the msp driver for the newer LSI2008 The Areca cards work well under FreeBSD - It's one of the cards I started investigating after I realized my PERC H700 wasn't a good option, and they provide source code as well. I'm mildly interested in trying to fix it, can program C, and I do have access to the card, but I'd need someone to help get me started, as I haven't done much with FreeBSD other than hack some kernel compiles and play with some custom NFS changes. -----Original Message----- From: owner-freebsd-drivers@freebsd.org [mailto:owner-freebsd-drivers@freebsd.org] On Behalf Of Erich Weiler Sent: Tuesday, March 08, 2011 2:46 PM To: Jonathan Stewart Cc: freebsd-drivers@freebsd.org Subject: Re: MFI Driver Behavior This was not resolved, although I tried in vain to get it to be more stable for over a month. The Perc H700 and H800 controllers do not work reliably with any version of FreeBSD. It's an MFI driver issue, such that the driver locks the controller temporarily under heavy disk load, but there seems to be little interest in fixing it. I thought at first it was a C-states issue, but then it locked again a week later. I punted and got an Adaptec controller, which is better supported under FreeBSD. Maybe they will make the MFI driver work better with these controllers one day, but I don't know when that will happen. Which is a bummer because more and more folks are buying these Perc H700/H800 controllers from Dell. On 03/08/11 10:38, Jonathan Stewart wrote: > [Please CC me I'm not on this list] > > Erich Weiler wrote: >>> We are testing a new theory. We believe that the C-states and C1E >>> control on this chipset (in the Dell R610), which are configurable >>> in the BIOS, are affecting the controller in some way. >>> We are disabling them and trying again. I'll post results soon. >> After disabling C-states and CIE in the BIOS on our R610, the Perc >> H800 controller now appears stable. My ZFS filesystem appears >> stable. Life seems good. >> >> I wonder what those items in the BIOS are doing in relation to the >> kernel to cause this. > > Have you had any further issues with this? I just updated 2 of my > systems to FreeBSD ???.net 8.2-STABLE FreeBSD 8.2-STABLE #2 r219138: Tue Mar 1 > 13:42:29 EST 2011 root@???.net:/usr/obj/usr/src/sys/SERENITY > amd64 > > from 8-stable sometime around 8 release and just had to have the > datacenter hit the reset button to bring the machine back. > > I'm reluctant to take this machine to current but if that fixes the > problem I can do it. The other option is calling the data center and > having them change the BIOS settings for me but that is also something > I would rather not do. > > Thanks, > Jonathan _______________________________________________ freebsd-drivers@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-drivers To unsubscribe, send any mail to "freebsd-drivers-unsubscribe@freebsd.org" From owner-freebsd-drivers@FreeBSD.ORG Tue Mar 8 19:01:58 2011 Return-Path: Delivered-To: freebsd-drivers@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 21252106566B for ; Tue, 8 Mar 2011 19:01:58 +0000 (UTC) (envelope-from jonstew1983@gmail.com) Received: from mail-yx0-f182.google.com (mail-yx0-f182.google.com [209.85.213.182]) by mx1.freebsd.org (Postfix) with ESMTP id D1DC38FC0C for ; Tue, 8 Mar 2011 19:01:57 +0000 (UTC) Received: by yxl31 with SMTP id 31so2413223yxl.13 for ; Tue, 08 Mar 2011 11:01:57 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:date:message-id:subject:from:to :content-type; bh=b/XksEE6o+9kmbnDLFsBg34DV8BXmn15q1CGIn/2VIs=; b=nirR5GK842JtaI/S0OopVOp6wJfnXtKPvPTsXwWSZEEyf7GSKvGYfk79xGWjBCcO7Y uW1FJJTwFx7uTvh6A8XvU41/is6Qt9l2MqstN6XMdQapgyc7X+n0o9Y1vTv6gB2GRpC+ gp7Fup6Kiw+QK2zj61jtsFQdAyYbdhj71+dG0= DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:date:message-id:subject:from:to:content-type; b=wWvqni7hQVbQGKDkS48gQVyzlknWkwRUwpnZFTuGEBO2LdqNqqeH5ebOSXj3dBLW8W B1wPKxKQDTPiahPpXjpnZ6mA04I8onZUZmYwQVa0x1AHmi1NGyIRtRbQRRouTHsqJSYd B8KtLRwdF3CvM8d1zf64yOpR0sDKUNSQWNn2I= MIME-Version: 1.0 Received: by 10.150.61.20 with SMTP id j20mr6426502yba.175.1299609495779; Tue, 08 Mar 2011 10:38:15 -0800 (PST) Received: by 10.151.41.21 with HTTP; Tue, 8 Mar 2011 10:38:15 -0800 (PST) Date: Tue, 8 Mar 2011 13:38:15 -0500 Message-ID: From: Jonathan Stewart To: Erich Weiler , freebsd-drivers@freebsd.org Content-Type: text/plain; charset=ISO-8859-1 Cc: Subject: Re: MFI Driver Behavior X-BeenThere: freebsd-drivers@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Writing device drivers for FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 08 Mar 2011 19:01:58 -0000 [Please CC me I'm not on this list] Erich Weiler wrote: >> We are testing a new theory. We believe that the C-states and C1E >> control on this chipset (in the Dell R610), which are >> configurable in the BIOS, are affecting the controller in some way. >> We are disabling them and trying again. I'll post results soon. > > After disabling C-states and CIE in the BIOS on our R610, the Perc > H800 controller now appears stable. My ZFS filesystem appears > stable. Life seems good. > > I wonder what those items in the BIOS are doing in relation to the > kernel to cause this. Have you had any further issues with this? I just updated 2 of my systems to FreeBSD ???.net 8.2-STABLE FreeBSD 8.2-STABLE #2 r219138: Tue Mar 1 13:42:29 EST 2011 root@???.net:/usr/obj/usr/src/sys/SERENITY amd64 from 8-stable sometime around 8 release and just had to have the datacenter hit the reset button to bring the machine back. I'm reluctant to take this machine to current but if that fixes the problem I can do it. The other option is calling the data center and having them change the BIOS settings for me but that is also something I would rather not do. Thanks, Jonathan From owner-freebsd-drivers@FreeBSD.ORG Sat Mar 12 19:07:44 2011 Return-Path: Delivered-To: freebsd-drivers@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 44F3210656D5 for ; Sat, 12 Mar 2011 19:07:44 +0000 (UTC) (envelope-from brix@freebsd.org) Received: from solow.pil.dk (relay.pil.dk [195.41.47.164]) by mx1.freebsd.org (Postfix) with ESMTP id A1AD28FC0C for ; Sat, 12 Mar 2011 19:07:43 +0000 (UTC) Received: from druadan.brixandersen.dk (0x55534f5f.adsl.cybercity.dk [85.83.79.95]) by solow.pil.dk (Postfix) with ESMTPA id 51D321CC006 for ; Sat, 12 Mar 2011 19:49:38 +0100 (CET) From: Henrik Brix Andersen Content-Type: multipart/signed; protocol="application/pgp-signature"; micalg=pgp-sha1; boundary="Apple-Mail-8--473975120" Content-Transfer-Encoding: 7bit Date: Sat, 12 Mar 2011 19:49:31 +0100 Message-Id: <3550EA55-ADDE-40AC-9C22-1FAC441A0BC8@freebsd.org> To: freebsd-drivers@freebsd.org Mime-Version: 1.0 (Apple Message framework v1082) X-Pgp-Agent: GPGMail 1.3.2 X-Mailer: Apple Mail (2.1082) Subject: Allocating resources to isab children X-BeenThere: freebsd-drivers@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Writing device drivers for FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 12 Mar 2011 19:07:44 -0000 This is an OpenPGP/MIME signed message (RFC 2440 and 3156) --Apple-Mail-8--473975120 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset=us-ascii Hi, I am writing a driver for the GPIO part of the AMD CS5536 south bridge = (isab0), but have encountered a problem with resource allocation. The GPIO address (0x6100-0x61FF) is available through PCI BAR 1 as seen = in the output from pciconf(8): isab0@pci0:0:15:0: class=3D0x060100 card=3D0x20901022 = chip=3D0x20901022 rev=3D0x03 hdr=3D0x00 vendor =3D 'Advanced Micro Devices (AMD)' device =3D 'CS5536 [Geode companion] ISA' class =3D bridge subclass =3D PCI-ISA bar [10] =3D type I/O Port, range 32, base 0x6000, size 8, = enabled bar [14] =3D type I/O Port, range 32, base 0x6100, size 256, = enabled bar [18] =3D type I/O Port, range 32, base 0x6200, size 64, = enabled bar [20] =3D type I/O Port, range 32, base 0x9d00, size 128, = enabled bar [24] =3D type I/O Port, range 32, base 0x9c00, size 64, = enabled However, when I try to allocate the I/O port resource in the = minimalistic example below, I always get 0x1100-0x1100, not = 0x6100-0x61FF? I have modelled the code after recommendation from John Baldwin in = http://lists.freebsd.org/pipermail/freebsd-acpi/2007-August/003964.html = but I must be doing something wrong... Brix #include __FBSDID("$FreeBSD$"); #include #include #include #include #include #include #include #include #include #include #define AMD_GEODE_CS5536_DEV_ID 0x20901022 struct test_softc { device_t dev; int rid; struct resource *res; }; static void test_identify(driver_t *driver, device_t parent) { if (device_find_child(parent, driver->name, -1) !=3D NULL) return; if (pci_get_devid(parent) =3D=3D AMD_GEODE_CS5536_DEV_ID) { if (device_add_child(parent, driver->name, -1) =3D=3D = NULL) device_printf(parent, "could not add child\n"); } } static int test_probe(device_t dev) { device_set_desc(dev, "Test driver"); return (BUS_PROBE_DEFAULT); } static int test_attach(device_t dev) { struct test_softc *sc; int err; err =3D 0; sc =3D device_get_softc(dev); sc->dev =3D dev; sc->rid =3D PCIR_BAR(1); sc->res =3D bus_alloc_resource_any(dev, SYS_RES_IOPORT, = &sc->rid, RF_ACTIVE); if (sc->res =3D=3D NULL) { device_printf(dev, "could not allocate I/O port\n"); err =3D ENXIO; goto out; } /* prints out 0x1100-0x1100 - not 0x6100-0x61FF as expected */ printf("allocated: 0x%lx-0x%lx\n", rman_get_start(sc->res), rman_get_end(sc->res)); out: return (err); } static int test_detach(device_t dev) { struct test_softc *sc =3D device_get_softc(dev); int err; err =3D bus_generic_detach(dev); if (sc->res !=3D NULL) bus_release_resource(dev, SYS_RES_IOPORT, sc->rid, = sc->res); return (err); } static device_method_t test_methods[] =3D { DEVMETHOD(device_identify, test_identify), DEVMETHOD(device_probe, test_probe), DEVMETHOD(device_attach, test_attach), DEVMETHOD(device_detach, test_detach), { 0, 0 } }; static driver_t test_driver =3D { "test", test_methods, sizeof(struct test_softc), }; static devclass_t test_devclass; DRIVER_MODULE(test, isab, test_driver, test_devclass, 0, 0); --=20 Henrik Brix Andersen --Apple-Mail-8--473975120 content-type: application/pgp-signature; x-mac-type=70674453; name=PGP.sig content-description: This is a digitally signed message part content-disposition: inline; filename=PGP.sig content-transfer-encoding: 7bit -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (Darwin) iEYEARECAAYFAk17wDwACgkQv+Q4flTiePicZgCgof8J5jeXvHGnCRjYCEYWCE7V 92EAniR3QgxC/6Q5nHdtIr4JxISZo/MP =V5jy -----END PGP SIGNATURE----- --Apple-Mail-8--473975120-- From owner-freebsd-drivers@FreeBSD.ORG Sat Mar 12 19:48:04 2011 Return-Path: Delivered-To: freebsd-drivers@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id D99851065672 for ; Sat, 12 Mar 2011 19:48:04 +0000 (UTC) (envelope-from brix@FreeBSD.org) Received: from solow.pil.dk (relay.pil.dk [195.41.47.164]) by mx1.freebsd.org (Postfix) with ESMTP id 581D68FC08 for ; Sat, 12 Mar 2011 19:48:04 +0000 (UTC) Received: from druadan.brixandersen.dk (0x55534f5f.adsl.cybercity.dk [85.83.79.95]) by solow.pil.dk (Postfix) with ESMTPA id 341491CC012 for ; Sat, 12 Mar 2011 20:48:03 +0100 (CET) Content-Type: multipart/signed; protocol="application/pgp-signature"; micalg=pgp-sha1; boundary="Apple-Mail-9--470468019" Mime-Version: 1.0 (Apple Message framework v1082) From: Henrik Brix Andersen In-Reply-To: <3550EA55-ADDE-40AC-9C22-1FAC441A0BC8@freebsd.org> Date: Sat, 12 Mar 2011 20:47:58 +0100 Content-Transfer-Encoding: 7bit Message-Id: <0A707516-C7D1-4441-B17B-1273B6C256B0@FreeBSD.org> References: <3550EA55-ADDE-40AC-9C22-1FAC441A0BC8@freebsd.org> To: freebsd-drivers@freebsd.org X-Pgp-Agent: GPGMail 1.3.2 X-Mailer: Apple Mail (2.1082) Subject: Re: Allocating resources to isab children X-BeenThere: freebsd-drivers@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Writing device drivers for FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 12 Mar 2011 19:48:04 -0000 This is an OpenPGP/MIME signed message (RFC 2440 and 3156) --Apple-Mail-9--470468019 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset=us-ascii On Mar 12, 2011, at 19:49, Henrik Brix Andersen wrote: > I am writing a driver for the GPIO part of the AMD CS5536 south bridge = (isab0), but have encountered a problem with resource allocation. > The GPIO address (0x6100-0x61FF) is available through PCI BAR 1 as = seen in the output from pciconf(8): >=20 > isab0@pci0:0:15:0: class=3D0x060100 card=3D0x20901022 = chip=3D0x20901022 rev=3D0x03 hdr=3D0x00 > vendor =3D 'Advanced Micro Devices (AMD)' > device =3D 'CS5536 [Geode companion] ISA' > class =3D bridge > subclass =3D PCI-ISA > bar [10] =3D type I/O Port, range 32, base 0x6000, size 8, = enabled > bar [14] =3D type I/O Port, range 32, base 0x6100, size 256, = enabled > bar [18] =3D type I/O Port, range 32, base 0x6200, size 64, = enabled > bar [20] =3D type I/O Port, range 32, base 0x9d00, size 128, = enabled > bar [24] =3D type I/O Port, range 32, base 0x9c00, size 64, = enabled >=20 > However, when I try to allocate the I/O port resource in the = minimalistic example below, I always get 0x1100-0x1100, not = 0x6100-0x61FF? >=20 > I have modelled the code after recommendation from John Baldwin in = http://lists.freebsd.org/pipermail/freebsd-acpi/2007-August/003964.html = but I must be doing something wrong... Ahh, I just re-read Johns recommendation and realized that isab does not = support proxying resource allocations (yet). I will try adding that = functionality and report back. Brix --=20 Henrik Brix Andersen --Apple-Mail-9--470468019 content-type: application/pgp-signature; x-mac-type=70674453; name=PGP.sig content-description: This is a digitally signed message part content-disposition: inline; filename=PGP.sig content-transfer-encoding: 7bit -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (Darwin) iEYEARECAAYFAk17ze8ACgkQv+Q4flTiePizagCglAqvwEjTsO4NEdi8YNz6CQQc 7m4An3uvISR6w/lYA5AVg+E4aa5varoe =9/Uz -----END PGP SIGNATURE----- --Apple-Mail-9--470468019--