Comment 1 by Philippe Perrot, Dec 7, 2010
I get the same error on my system.
Comment 2 by Philippe Perrot, Dec 7, 2010
Using aquamac bug fixes, the error disappear, but the card doesn't work properly
Comment 3 by Engelbert Gal, Dec 10, 2010
The same error appears on a PNY GTX 460 (VCGBTX4601XPB) http://www.tonymacx86.com/viewtopic.php?f=15&t=8625
Comment 4 by Cosmosis Jones, Jun 2, 2011
if this is still a problem reopen referencing TRUNK
Comment 5 by Philippe Perrot, Jun 3, 2011
Hi, this issue is still existing under 10.6.7
Comment 6 by Cosmosis Jones, Jun 3, 2011
Comment 7 by Cosmosis Jones, Jun 3, 2011
yeah not 10.6.7 as much as are u using a boot compiled from the trunk? rev(944)
Comment 8 by Philippe Perrot, Jun 3, 2011
Yes, I got this issue in using this nvidia card with ATI 5680.
Comment 9 by Cosmosis Jones, Jun 3, 2011
Ah ok then.. you have 2 cards, so the graphics patcher said bad config for the ATI.. How exactly is everything setup (ie which slot is primary/which secondary)
Comment 10 by Philippe Perrot, Jun 3, 2011
The ati is primary and I get this message with the nvidia, in second. The graphic patcher said bad config for nvidia, and is fine for Ati
Comment 11 by Philippe Perrot, Jun 3, 2011
The ati is primary and I get this message with the nvidia, in second. The graphic patcher said bad config for nvidia, and is fine for Ati
Comment 12 by Pedro A. Aranda, Jul 3, 2011
"Coming from issue 96" and just arrived. Hi :-) Well my laptop is one those NVidia Optimus things with dual GPU. The NVidia on PCI 1 and I have an Arrandale IntelHD on PCI0. In my case, I see that the NVidia Patcher says it can't patch my graphics card and I have '-4095 MBytes of video RAM' It also says other things, but it goes a bit too fast for me. Can I stop Chameleon somehow to get the rest of the messages? Thanks a lot /PA
Comment 13 by Pedro A. Aranda, Jul 3, 2011
BTW, I forgot: I have a 1366x768 13 inch laptop (ASUS U30Jc) and I don't get beyond 1024x768. I'm using Chameleon-2.0-RC5-r1083 from olarila.com http://olarila.com/forum/viewtopic.php?f=9&t=70 Thanks /PA
Comment 14 by Azimutz, Jul 16, 2011
Hi guys... this is not very important for this issue but, i noticed no one cared about posting concrete info about the cards, e.g. model, device id, subsys id an vendor. Not good! About the issue, two things: - i have added pauses on the messages related to the issue (r1164) to help you guys in collecting feedback; please try the booter again and feedback - i followed the link on comment #3 and read some interesting stuff about this problem, when dual booting with Windows; anyone can check/confirm it? I have suspicions but i can't test nVidia atm and without proper feedback we can't fix any. So, remember the cards info... If there's no feedback soon i'll end up closing the issue! Thanks
Comment 15 by Philippe Perrot, Jul 16, 2011
Hi, got this trouble with Sapphire ati 5870 and evga gtx 580, under osx 10.6.6, and the nvidia board was the second board to get this message ; got this message with gtx285 as second board. Mother board is evga SR-2. The booter was r470. Won't able to test anything until august or october, as I'll be back home. Regards
Comment 16 by Azimutz, Jul 17, 2011
Hi Philippe... that doesn't help much. Test again when you can. If the issue is closed create a new one. Thanks
Comment 17 by Pedro A. Aranda, Jul 17, 2011
Hi Before going to the new bootloader, just another thing I have seen: Chameleon is not able to find any 1366x768 VESA mode on my machine...
Comment 18 by Azimutz, Jul 17, 2011
Pedro, that's not a Chameleon problem; your card's rom simply doesn't have that vesa mode; to confirm, when you are at the boot prompt, hit the "down arrow" key and select "Video info"; it will show you the available modes. The booter it's not new; i just added some pauses so you guys are alerted to the problem and can note down the error messages in case you can't boot. If you still can boot, the error messages can be checked on the boot log (bdmesg).
Comment 19 by Philippe Perrot, Jul 17, 2011
What will help much ? What do you need ? Regards
Comment 20 by Azimutz, Jul 18, 2011
Philippe, read comment 14 carefully; it's all there!! Just ignore the "pauses" added stuff; i removed them because you guys can check the messages on the boot log (bdmesg). Look for these: - no dcb table found - Bad display config block signature... - Bad Display Configuration Block signature... - ERROR: dcbtable_version is... Any of these messages will also print: - ERROR: nVidia ROM Patching Failed!
Comment 21 by Cosmosis Jones, Jul 18, 2011
Owner: azimutz
Comment 22 by Cosmosis Jones, Jul 18, 2011
Comment 23 by Philippe Perrot, Jul 19, 2011
Is this pict helpful ? Best regards
- IMG_5055.jpg - 760.22 kB
Comment 24 by Evan Lojewski, Jul 20, 2011
Azimutz: I was taking a look at the nvclock source code and found this for reading the vbios from the NV50 family of cards: /* On NV5x cards we need to let pramin point to the bios */ if (nv_card->arch & NV5X) { uint32_t vbios_vram = (nv_card->PDISPLAY[0x9f04/4] & ~0xff) << 8; if (!vbios_vram) vbios_vram = (nv_card->PMC[0x1700/4] << 16) + 0xf0000; old_bar0_pramin = nv_card->PMC[0x1700/4]; nv_card->PMC[0x1700/4] = (vbios_vram >> 16); } We currently are not doing this with the NV50 cards, so this change (or equiv) may be needed. I don't have a NV50 card, so I cannot test this.
Comment 25 by Pedro A. Aranda, Jul 22, 2011
I collected a couple of things, both Linux and MacOS from my Asus U30Jc. I have also tried to get the NVidia ROM with FreeDOS + NVFlash but it says EEPROM not recognised... I have edited the Linux DMESG output because of its length. I just hope I didn't delete anything important. Anyhow, I have a backup :-)
- AsusU30Jc_bdmesg.txt - 13.98 kB - view
Comment 26 by Azimutz, Jul 23, 2011
Hi guys; i've been without network for last couple of days so, i need some time to get back on top of stuff... Philippe, the pic is a start, yes :) Pedro, will take a look. Evan, yep... i suspect of at least two things for now: - windows may be setting something on the card that causes the bad config diplay, when booting OS X by rebooting directly from windows - we need to update the code for new cards, like you suggest For now, i can't test nVidia; i have the GT 220 to install on the desktop, but i need the only available pci express slot for ATi testing; the laptop has a GT 450M, but it will take a while for me to get my hands on it. Besides, i never used nVidia so, the knowledge is very limited.. keep that in mind ;)
Comment 27 by Philippe Perrot, Jul 23, 2011
yes, it's an old pict of start, took in february. As I'll be able to access to my computer, I'll wil test with last booter and lion... But I think the trouble remains. I done some test with lion in may, it was stopped by new "pci config" stuff. I've read the solution to pass through this trouble in adding something in boot.plis file.
Comment 28 by Cosmosis Jones, Jul 25, 2011
Comment 29 by A Sto, Sep 7, 2011
I'm experiencing the same issue. The computer boots to Lion fine unless there is a power outage or booted from Windows prior. The only way I have been able to fix this is to take the GTX 580 out of the computer and leave it overnight. The next reboot after the GTX580 boots chameleon as expected. I have a Gainward GTX 580 3GB Phantom.
Comment 30 by Locmer -, Oct 22, 2011
Trying to add a: Gainward GeForce (512 MB) (4718462009726) Graphics Card to a 10.7.2 on a Gigabyte GA-Z68X-UD3H-B3 Results in: ERROR: dcbtable_version is 0x3 ERROR: nVidia ROM Patching Failed! for details of how Mac OS X sees the GEForce 8400 see attached jpeg
- geforce8400gs.jpg - 83.45 kB
Comment 31 by Moik The, Apr 25, 2012
Howdi! Not sure if this is still an issue, but here's my solution to it ;) Actually I stumbled upon the very same issue and tried to find out what's wrong. When comparing the current nvidia injector code with the nouveau codebase (especially the dcb reading part) reading multibyte integers need to be swapped on some cases. In this patch I change the following: - some of the uint16 and uint32 reading issues - swapped the order of rom fetching from pramin-prom to prom-pramin (guess that's just a matter of taste) I just fixed the necessities as currently I'm in the process of rewriting the entire nvidia injector from the scratch based on information from the current nouveau codebase. Will share my progress in case anyone is still interested in having most of the nvidia issues resolved ;)
- 0001-tidy-up-of-nvidia-scanning.patch - 5.58 kB - view
Comment 32 by Philippe Perrot, Apr 25, 2012
Hi Michael, I'm still interested by solving this issue. Thanx
Comment 33 by Cosmosis Jones, Apr 25, 2012
submit your patch as a patch on forge. also join irc and #Chameleon message me to get invited into the dev channel if you are going to rewrite it thanks cosmo1t
Comment 34 by Cosmosis Jones, May 8, 2012
merged patch (some of it) restructured some things on my own as well
Owner: cosmo1t
Comment 35 by bu11etpr00f, Aug 21, 2012
It appears that this issue might be related to using dual video cards (one ATI & one NVIDIA). All my details are included in my post here: http://www.tonymacx86.com/graphics/68171-dual-graphics-video-cards-at i-radeon-5770-nvidia-geforce-8800-gt-trinary-3-three-displays.html Thanks!
Comment 36 by Anmol Mishra, Nov 10, 2012
This issue is still a problem, and there are many posts regarding this. bu11etpr00f has given an excellent summary in his post above.
Comment 37 by Cosmosis Jones, Nov 11, 2012
thats nice. dual video card issue IS not the same as bad block sig. Chameleon has never supported injecting multiple video cards from different manufactures on the same GE=yes flag. Maybe it was intended to do so , but never worked. Instead of continuing to pile crap on this post, open a new one with an appropriate title and information. Nothing is more annoying than reading new posts on closed tickets. This is an issue in our ticketing system that i am going to get fixed. So open a new issue, with the real information, the real bdmesg's outputs, multiple dev ids, the ioregs, etc, not a link to another forum site. mmmmmk?
Reported by rabbit 74, Dec 3, 2010