Date: Sat, 29 Jan 2011 12:35:30 -0800 From: "Duane H. Hesser" <dhesser@accima.com> To: rank1seeker@gmail.com Cc: freebsd-hackers@freebsd.org Subject: Re: NVIDIA (port) driver fails to create /dev/nvidactl; 8.2Prerelease Message-ID: <20110129123530.0f4586d8.dhesser@accima.com> In-Reply-To: <20110129.064924.078.1@DEV> References: <20110128124312.a837f288.dhesser@accima.com> <20110129.064924.078.1@DEV>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sat, 29 Jan 2011 07:49:24 +0100 rank1seeker@gmail.com wrote: > > > echo 'debug.acpi.disabled="sysres"' >> /boot/loader.conf > > # reboot I added the above sysctl to loader.conf and rebooted. No change, but that is likely because no such sysctl exists in the stable kernel. Are you running current? # sysctl debug.acpi debug.acpi.suspend_bounce: 0 debug.acpi.reset_clock: 1 debug.acpi.do_powerstate: 1 debug.acpi.interpreter_slack: 1 debug.acpi.enable_debug_objects: 0 debug.acpi.acpi_ca_version: 20101013 debug.acpi.ec.timeout: 750 debug.acpi.ec.polled: 0 debug.acpi.ec.burst: 0 debug.acpi.batt.batt_sleep_ms: 0 debug.acpi.resume_beep: 0 Please let me take this opportunity to explain something (too everyone on this list who is willing to read it). My original post was not a plea for help. I appreciate the posts which have offered possible "tuning" items to resolve the problem, but my position is this: There is a significant regression in current stable sources which prevent the nvidia-driver port from initializing due to the absence of the /dev/nvidactl device. This is *probably* due to another bug in 'devfs'. I was willing to consider that I might have missed something in installing the port, but that is very unlikely...I've been doing this for a long time. I offered the opportunity for someone running a recent version of stable to report that they were running the nvidia-driver just fine, but that hasn't happened yet. Apparently it is working on current. It is NOT working on stable. I have examined the nvidia-driver port sources, and consider it unlikely that the problem is there (but again, ..could be wrong). Thanks again to those who have offered suggestions, but I submit that if a simple sysctl or the selection of port options determines whether device nodes are available, then 'things' have become far too complex. -- Duane H. Hesser <dhesser@accima.com>
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20110129123530.0f4586d8.dhesser>