Yes, I have the same Plextor M8Pe 256GB & 512GB M.2 devices which worked perfectly at 10.12.x.However not working at 10.11.6 yet with either Piker's patches by Clover's config.plist, RehabMan's HackNVMeFamity.kext, or IONVMeFamilyBorg.kext.Hope someone can help for this issue.Maybe the kext from 12.whatever can load under 11?
I can't load the 12 kext on 11. Two possibilities: either get the newer IOBlockStorageDevice in place on the system, or find a way to hack the vtable import?
I believe there's only one additional virtual function that was added, something about provisions I don't quite understand. Or the third possibility: I go out and write a driver for these nvme drives using FreeBSD code like the NVMeGeneric one, and fix power management.
I think the easiest would be first: just add the IOBlockStorageDevice from 10.12 to SLE then test it.
System might bomb though.
Sorry folks, this is only for PCI/PCIe chips!
the good news for USB is that it's a generic mass storage chip, so it should work out of the box with the proper matching injection.
I can't help with that much, don't have the hardware
that's good to see. the speed I'm seeing seems pretty constant around 4/9. I don't have one of these super fancy fast sd cards, can't test that.
in theory since we're attached to a pci bus the driver could go blazing fast. it depends on what clocking the driver configures
@Micky1979 very cool! I'm surprised hah.
The program I'm using to test disk speed is blackmagic Disk Speed Test.
device id is 10de 5229. Why not add the "IOPCIClassMatch" instead of the "IOPCIMatch"? Well, this chip is not a generic SDHCI device so it's a bad idea to match on anything. i'll go with adding all the id's the openbsd driver tries to match on, that should do it. there's at most 5-6 of them i think.
I'll collaborate! ;)great i need testers. have fun. if a card gets stuck try ejecting and inserting back. Sinetek-rtsx.kext-0.1-btest.zip 35.25KB 55 downloads
just realized, you'll need to edit the plist to add your pci id, mine is 5227, that's the only one i bothered to add in.
report back if the driver loads on your hardware after that. thx
For more information just see the OpenBSD man page for rtsx, since that's what the driver is based from.
macOS kext_rtsx Hi, making a new thread. This driver is for Realtek SDHC card readers on a pci/pcie bus, most commonly found in laptops.
I would like to get testers. The driver is not yet finalized, but I'm making the tread now anyway since there was interest. expect this post to change in the future. [read below for current status] -- sinetek things to fix eventually - get more pci id's in plist- async task in its own thread instead of blocking the os- prealloc DMA buffer?- cards with sector_size != 512, haven't been tested at all.- clean up src. git push.- power management? what happens when you sleep the machine?- the eject button in finder is broken, do not use for now.- spams the syslog with debug messages. release config?- basic namespace pollution a bit out of hand
Well Okay, I made some progress using a regressive read method.
I'm not getting panics anymore after the card gets attached, but there's a slight problem with speed using this method. blockslow.png 525KB 2 downloads
@quinielascom sorry for derailing your thread, i'll start my own!
For Realtek chips, please hop on over to the other thread!
Hello, I'm running X99 Final1ty Gaming i7 motherboard + a 6900K processor. x99+i7.jpg 14.22KB 7 downloads
I managed to get my own IOPCIFamily compiled with a MAX_BAR_SIZE fix (should also fix PFM64_SIZE?? to remove -npci=0x2000 problem. need to probe further) and now the system boots up well.
I'm running a custom kernel to avoid binary patching.. I just added in the entry for BROADWELLE (family 0x4F) in cpuid.c.
For Booting I'm ignoring Clover completely (had many problems with it on this board, the UEFI memory map is totally borked), instead use Enoch r2839.
For the USB problems I just rolled back the kexts using spakk's method, see this. This way all (Intel) ports seem to work just fine.
Hello, I'm running X99 Final1ty Gaming i7 motherboard + a 6900K processor.
I managed to get my own IOPCIFamily compiled with a MAX_BAR_SIZE fix (should also fix PFM64_SIZE??) and now the system boots up well.
I'm using enoch, and don't get any instability due to the funky EFI memory map on these boards. I'm also running a custom kernel to avoid binary patching.. I just added in the entry for BROADWELLE (family 0x4F) in cpuid.c.
USB problems I just rolled back the kexts using spakk's method, see this.
Awesome, glad to see so much interest! The driver is basically done, but right now it is insanely buggy and hangs up my machine under weird circumstances.
So I'm still racking my brains out finding that bug, sadly.
Initial performance indicates around 8.5 MB/s read speads on my crappy testing card. Not bad at all
Also I notice that some users are getting opcodeemu errors on their system. In particular the instruction are "pextrd" and "pextrq", belonging to SSE41.
So it might be necessary nowadays to add SSE41 to the opemu, not sure why that was removed