View Full Version : P35/7.0 Beta Issues
I am unable to get the most recent nVidia driver to work on a new computer. After a "startx" the screen clears, and is blank, with the cursor in the upper left. It is a hard freeze, and the computer must be rebooted.
This is 32-bit FreeBSD 7.0-Beta4 and the most recent nVidia, xorg and ancillaries (cvsupped yesterday). The BSD agp was disabled in the kernel, and the proper options set in xorg.conf (NvAGP 1, nvidia driver, glx). The last log message was that the nVidia gart was loaded.
The board is a P35 (Gigabyte P35-DS3R) with 4GB memory, a Q6600 and an nVidia 8500GT. There seems to be no BIOS option to disable memory remapping.
The nv driver worked before I installed the nVidia driver (and before I took agp out of the kernel).
Any advice on how I might get this to work would be most appreciated!
More information: the memory remapping (or lack thereof) seems to be the issue. I removed 2GB of memory, and the nVidia driver loads and executes properly.
I count this as a driver bug, since I am unable to configure memory remapping in BIOS. If the BIOS can't do it, then the driver has to, does it not?
The NVIDIA FreeBSD graphics driver does not currently support systems on which RAM has been mapped above 4GB if the kernel will return such memory via malloc(); this likely won't change until the d_mmap() interface is improved with the functionality originally requested to enable FreeBSD/amd64, or until a vmap()-like kernel interface becomes available.
Thanks for the reply. It seems that the BIOS has to have memory remapping switches, or else the driver will not work on systems that have 4+ GB of memory. Not all motherboards do, sadly.
I'm actually surprised this even is a problem, since PAE kernels aren't supported and building against one should fail. I guess there's a possibility that the problem isn't actually the RAM mapped above 4GB itself, but something related (e.g. I/O range assignments, etc.). It'd be interesting to see the e820 memory maps, as well as a detailed summary of PCI resource allocations and kernel messages in both the 6GB and a working configuration.
I was thinking about this as well. FreeBSD sees about 3.5GB out of the 4GB installed (*not* 6 GB as you mention), and the remaining 0.5GB would be expected to be reserved for drivers and the like. There is no PAE enabled, so that is not the issue. Let me see what I can dig up regarding resource allocations. For my information, what is the e820 memory map?
Here's a first pass:
For the 4GB configuration:
The initial boot message is that 524MB of memory above 4GB is discarded.
% dmesg | grep mem
real memory = 3756916736 (3582 MB)
avail memory = 3664588800 (3494 MB)
acpi_hpet0: <High Precision Event Timer> iomem 0xfed00000-0xfed003ff on acpi0
nvidia0: <GeForce 8500 GT> port 0xb000-0xb07f mem 0xf6000000-0xf6ffffff,0xe0000000-0xefffffff,0xf4000000-0xf5ffffff irq 16 at device 0.0 on pci1
ehci0: <EHCI (generic) USB 2.0 controller> mem 0xfa104000-0xfa1043ff irq 18 at device 26.7 on pci0
atapci0: <JMicron AHCI controller> mem 0xfa000000-0xfa001fff irq 19 at device 0.0 on pci3
re0: <RealTek 8168/8111B PCIe Gigabit Ethernet> port 0xd000-0xd0ff mem 0xf9000000-0xf9000fff irq 16 at device 0.0 on pci4
ehci1: <EHCI (generic) USB 2.0 controller> mem 0xfa105000-0xfa1053ff irq 23 at device 29.7 on pci0
atapci2: <Intel AHCI controller> port 0xe600-0xe607,0xe700-0xe703,0xe800-0xe807,0xe900-0xe903,0xea00-0xea1f mem 0xfa106000-0xfa1067ff irq 19 at device 31.2 on pci0
orm0: <ISA Option ROM> at iomem 0xd2000-0xd4fff pnpid ORM0000 on isa0
vga0: <Generic ISA VGA> at port 0x3c0-0x3df iomem 0xa0000-0xbffff on isa0
% pciconf -l
hostb0@pci0:0:0:0: class=0x060000 card=0x50001458 chip=0x29c08086 rev=0x02 hdr=0x00
pcib1@pci0:0:1:0: class=0x060400 card=0x50001458 chip=0x29c18086 rev=0x02 hdr=0x01
uhci0@pci0:0:26:0: class=0x0c0300 card=0x50041458 chip=0x29378086 rev=0x02 hdr=0x00
uhci1@pci0:0:26:1: class=0x0c0300 card=0x50041458 chip=0x29388086 rev=0x02 hdr=0x00
uhci2@pci0:0:26:2: class=0x0c0300 card=0x50041458 chip=0x29398086 rev=0x02 hdr=0x00
ehci0@pci0:0:26:7: class=0x0c0320 card=0x50061458 chip=0x293c8086 rev=0x02 hdr=0x00
none0@pci0:0:27:0: class=0x040300 card=0xa0021458 chip=0x293e8086 rev=0x02 hdr=0x00
pcib2@pci0:0:28:0: class=0x060400 card=0x50011458 chip=0x29408086 rev=0x02 hdr=0x01
pcib3@pci0:0:28:3: class=0x060400 card=0x50011458 chip=0x29468086 rev=0x02 hdr=0x01
pcib4@pci0:0:28:4: class=0x060400 card=0x50011458 chip=0x29488086 rev=0x02 hdr=0x01
uhci3@pci0:0:29:0: class=0x0c0300 card=0x50041458 chip=0x29348086 rev=0x02 hdr=0x00
uhci4@pci0:0:29:1: class=0x0c0300 card=0x50041458 chip=0x29358086 rev=0x02 hdr=0x00
uhci5@pci0:0:29:2: class=0x0c0300 card=0x50041458 chip=0x29368086 rev=0x02 hdr=0x00
ehci1@pci0:0:29:7: class=0x0c0320 card=0x50061458 chip=0x293a8086 rev=0x02 hdr=0x00
pcib5@pci0:0:30:0: class=0x060401 card=0x50001458 chip=0x244e8086 rev=0x92 hdr=0x01
isab0@pci0:0:31:0: class=0x060100 card=0x50011458 chip=0x29168086 rev=0x02 hdr=0x00
atapci2@pci0:0:31:2: class=0x010601 card=0xb0051458 chip=0x29228086 rev=0x02 hdr=0x00
none1@pci0:0:31:3: class=0x0c0500 card=0x50011458 chip=0x29308086 rev=0x02 hdr=0x00
nvidia0@pci0:1:0:0: class=0x030000 card=0x230a1682 chip=0x042110de rev=0xa1 hdr=0x00
atapci0@pci0:3:0:0: class=0x010601 card=0xb0001458 chip=0x2363197b rev=0x02 hdr=0x00
atapci1@pci0:3:0:1: class=0x010185 card=0xb0001458 chip=0x2363197b rev=0x02 hdr=0x00
re0@pci0:4:0:0: class=0x020000 card=0xe0001458 chip=0x816810ec rev=0x01 hdr=0x00
The 2GB case:
%dmesg | grep mem
real memory = 2146304000 (2046 MB)
avail memory = 2082365440 (1985 MB)
acpi_hpet0: <High Precision Event Timer> iomem 0xfed00000-0xfed003ff on acpi0
nvidia0: <GeForce 8500 GT> port 0xb000-0xb07f mem 0xf6000000-0xf6ffffff,0xe0000000-0xefffffff,0xf4000000-0xf5ffffff irq 16 at device 0.0 on pci1
ehci0: <EHCI (generic) USB 2.0 controller> mem 0xfa104000-0xfa1043ff irq 18 at device 26.7 on pci0
atapci0: <JMicron AHCI controller> mem 0xfa000000-0xfa001fff irq 19 at device 0.0 on pci3
re0: <RealTek 8168/8111B PCIe Gigabit Ethernet> port 0xd000-0xd0ff mem 0xf9000000-0xf9000fff irq 16 at device 0.0 on pci4
ehci1: <EHCI (generic) USB 2.0 controller> mem 0xfa105000-0xfa1053ff irq 23 at device 29.7 on pci0
atapci2: <Intel AHCI controller> port 0xe600-0xe607,0xe700-0xe703,0xe800-0xe807,0xe900-0xe903,0xea00-0xea1f mem 0xfa106000-0xfa1067ff irq 19 at device 31.2 on pci0
orm0: <ISA Option ROM> at iomem 0xd2000-0xd4fff pnpid ORM0000 on isa0
vga0: <Generic ISA VGA> at port 0x3c0-0x3df iomem 0xa0000-0xbffff on isa0
pciconf is the same.
There are some IRQ conflicts, but removing the devices that conflict with the video card makes no difference in the 4GB case.
Please let me know what else you would like.
Let me add one more thing: the computer does stay alive (namely, I can ssh into it) even though the local screen is frozen solid. top shows that all of one CPU is being used by X11. Killing the job shows the X11 error message "NVIDIA: failed to set MTRR 0xe0000000, 0M" on the local screen. The final line of the log file was "(II) NVIDIA(0): Initialized GART".
Please generate nvidia-bug-report.log files on the working and failing configurations; also, do you have the means to set up a serial console and run with a KDB enabled kernel (see the FreeBSD Handbook for setup details)?
Please generate nvidia-bug-report.log files on the working and failing configurations
Will do.
... also, do you have the means to set up a serial console and run with a KDB enabled kernel (see the FreeBSD Handbook for setup details)?
I can certainly include KDB in the kernel, but if it required use of a serial console and I/O redirection that is hard. The board does not have serial ports in the I/O plate, but only a header on the board. There is also no option to redirect console I/O to that serial port. If I can get that information over ssh, then I can do that. Please let me know if I misunderstand.
Here's the two log files. -2G works properly, -4G does not.
Strange, googling shows this as the only post with the exact error "failed to set MTRR 0xe0000000, 0M"
I have a P35-DS3R as well, with 4GB of RAM. I wonder why no one else seems to be reporting having that trouble with the nvidia driver...
I have a P35-DS3R as well, with 4GB of RAM. I wonder why no one else seems to be reporting having that trouble with the nvidia driver...
No idea, Saul -- I would have thought it was a common enough configuration. It is disappointing that there has been no response after the initial request for the bug reports (which were supplied).
There may be many people using this driver with 4gb of RAM & FreeBSD. (I'm using 6.2-stable)...I think it only happens on this particular motherboard though. Gigabyte support was totally unhelpful.
What did Gigabyte support say? The company is known for their many BIOS revisions. No help on that front?
Saul,
You have PM on this board.
Dear Sir,
In regards to your issue, If you installed total 4GB memory, the system will detect less than 4GB of total memory because of address space allocation for other critical functions, such as:
- System BIOS (including motherboard, add-on cards, etc..)
- Motherboards resources
- Memory mapped I/O
- configuration for AGP/PCI-Ex/PCI
- Other memory allocations for PCI devices
Different onboard devices and different add-on cards (devices) will result of different total memory size. In fact, when you do not install full 4GB memory, memory controller can use free 1GB memory address lines will be used for "PCI memory" to allocate PCI devices.
But, if you install full 4GB memory modules on this MB, memory controller will not have more address line to allocate PCI devices. Thus, BIOS will uses system memory space for PCI device allocation, then you will find shortage of total system memory capacity.
For example, a PCI-E VGA card needs 256MB system memory to allocate, and also needs extra system memory space for onboard audio chip, LAN controller and others. Usually, you will lose around 1GB memory space for that.
This limitation applies to most chipsets & Windows XP 32-bit version operating system.
If you install Windows XP 32-bit version operating system, we recommend that you install less than 3GB of total memory. Windows XP Pro X64 Ed. (64-bit) or other OS which can address more than 4GB memory
We also provide the FAQ on our website: http://www.gigabyte.com.tw/Support/Motherboard/FAQ_List.aspx?FAQID=1493
At last, if you still have any further question or suggestion about our products/service, please do not hesitate to contact with us directly. We will try our best to help you resolve the problem ASAP.
******** Thank you for supporting Gigabyte products ***********
I have P35 (Gigabyte GA-P35-DS4 Rev.2.0 with BIOS #F11) motherboard, a Intel C2D E8200, 1GB DDR2 x 4 and nVidia 8500GT. I encountered a system freeze when startx was executed on FreeBSD 7.0-RC1.
I use the modified kernel which is enabled some debugging option.
(reference: http://lists.freebsd.org/pipermail/freebsd-current/2007-September/077008.html )
But I can't get a memory dump because keyboard was not active and couldn't enter kdb mode while system hang-up.
This problem appear when "kern.smp.active" is set to "0" by sysctl.
I try the new version of nvidia-driver (169.07) with modified ports. But the problem is not cleared.
Attached bug report contains /var/log/Xorg.0.log. That Xorg.0.log was generated when startx is executed without nvidia.ko module. No Xorg.0.log was logged when system hang-up.
If you can read Japanese, please refer to my article:
http://taraijpn.blogspot.com/2008/02/nvidia-driverxhang-up.html
edhunterbg
04-08-08, 03:07 PM
I have same issue with Gigabyte P35-DS4 F12 bios, and there is no memory remapping options in bios. Hardware removing the 2 of 4 gigs makes X with nvidia driver start ok.
p.s. 7-stable 2008-4-3 sources and world and kernel
Bront_in_Hood
07-02-08, 08:20 AM
Hello everybody!
I have "P5B" motherboard, nvidia 9600 GT videocard, 4 Gbs RAM, FreeBSD 7.0 and the same problem with Xorg!!! I read this topic and tried to decrease my RAM to 2 GBs. And successfully started Xorg! (nana2)
So what is the problem? Is it motherboard, nvidia driver or FreeBSD bug? Or all together? And what can we do with this?
vBulletin® v3.7.1, Copyright ©2000-2010, Jelsoft Enterprises Ltd.