Date: Wed, 30 Dec 2009 09:40:09 -0800 From: Dana Myers <dana.myers@gmail.com> To: John Baldwin <jhb@freebsd.org> Cc: Joerg Wunsch <joerg_wunsch@uriah.heep.sax.de>, freebsd-acpi@freebsd.org Subject: Re: FreeBSD 8.0 hangs on boot with ACPI enabled Message-ID: <4B3B9079.1060208@gmail.com> In-Reply-To: <200912301122.28030.jhb@freebsd.org> References: <20091230082556.GD1637@uriah.heep.sax.de> <200912300839.47463.jhb@freebsd.org> <20091230143943.GA1616@uriah.heep.sax.de> <200912301122.28030.jhb@freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
John Baldwin wrote: > No, when ACPI is enabled the BAR is getting reset to 0 for some reason > after > the boot has started and FreeBSD tries to guess at an address to use. > Unfortunately it picks an address that doesn't work. This is fixable, it's > just part of the much larger PCI resource management problem. > > OTOH, I'm not sure why initializing ACPI is trashing the BAR. If you want, > you can try to narrow down at what point the BAR gets reset to 0 I'm sure it's a side-effect of switching into ACPI mode. The Solaris port of ACPI CA has an option to leave the system in legacy mode during ACPI initialization which I've used to diagnose several instances of PCI config space being trampled by the ACPI-mode switch (done by SMM BIOS code). Sometimes an update to the BIOS solves the problem, but usually not. My guess is that Windows reconfigures PCI devices after switching into ACPI mode, and thus avoids this class of BIOS bug. Dana dana.myers@gmail.com
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4B3B9079.1060208>