Jump to content

Need advice on USB3 PCI card


frankiee
 Share

41 posts in this topic

Recommended Posts

Yeah, the MacPro Definitons are actually in SSDT-2 ;)

 

I think those messages arrive when OS X is sort of trying to detect the available ports when they are not explicitly specified. But I do not see this as a problem (and my card surely has a different port setup than nMP anyways)

Link to comment
Share on other sites

Regarding the port injecting, your IOReg is the same as mine. So the nMP injections should work fine (we have the same number of ports with the same addresses). If you add that code to your SSDT you'll find the createports messages should go away and your ports will be labelled HS01 to SSP4 in IOReg. Whether there's any other benefit I can't say, but it should be safe to try if you want to.

 

Well, I think we have the same number of ports anyway. There's 4 in your XH01 entry, but 5 in the picture of the card? Anyway, if there's more ports you can just add extra entries for them making sure to use the right address. Or if the others are under XH02 then you could just try XH01 first.

Link to comment
Share on other sites

Regarding the port injecting, your IOReg is the same as mine. So the nMP injections should work fine (we have the same number of ports with the same addresses). If you add that code to your SSDT you'll find the createports messages should go away and your ports will be labelled HS01 to SSP4 in IOReg. Whether there's any other benefit I can't say, but it should be safe to try if you want to.

 

Well, I think we have the same number of ports anyway. There's 4 in your XH01 entry, but 5 in the picture of the card? Anyway, if there's more ports you can just add extra entries for them making sure to use the right address. Or if the others are under XH02 then you could just try XH01 first.

 

The card not only has 5 ports on the back of the card, but an internal header as well, so that should be 7(!) ports in total. As I mentioned, they seem to use some hubs, so there is one (USB3) hub appearing before the internal ports, but also an USB2(!) hub, which seems to connect to the outer two ports on the card. I dont know why this happens, maybe they are cheating ;) Looks like this in System Profiler:

 

2jNSfQe.png

Link to comment
Share on other sites

It's strange that only four of the physical ports are showing up in the IOReg (based on your screenshot there). Both hubs also seem to be attached to the same physical port (#3).

 

Maybe its because the FL1100 chip has only 4 ports? I think they then used two hubs to make them 7 total. But I still wonder why one of the hub only shows as USB2 hub.

Link to comment
Share on other sites

Yep, that's what it is. The FL1100 is a 4-port controller. And the VIA hub chip (there's only one in the card pic) is attached to port #3 and operates in 2.0 and 3.0 mode like the host controller.

 

So my IOReg isn't the exact same, I wasn't taking into account the hub on port #3.

 

I don't know if using my SSDT port injections would disable the hub on port 3 or if it would just add itself regardless. Only one way to know for sure...

Link to comment
Share on other sites

Yip, but other than showing up as USB2, the ports seem to work. When I connect a stick on the internal ports (the ones hanging on the USB3 hub), it shows as USB3 and seems to work at the correct speed. So since I wanted my case ports being functional, I think I have to live with that, or else I would actually need two cards .... I know there is one model with two external ports and one internal header, so this also should be hub-free.

 

What bugs me more is that charging thing, it would be sooo nice if the card would deliver more power (especially since it needs additional power anyways!) and charge while sleeping. Well, seems like you cannot have everything.

 

Any idea about that problem with booting from USB? That also bugs me, since I cannot completely test sleep on my 10.11 test system. Since my hack is actually a production machine, I am pretty vary of just switching to a new system without having completely tested everything.

Link to comment
Share on other sites

Yeah that's the model I have. Two ports on the back and the header for front-panel ports.

 

I'll have a look at the usb booting thing later. I have a sata hot-swap bay that I use for dual-booting in times like this but I can throw the Mavericks drive into a USB3 dock and try that.

 

As for the iPad charging, are you connecting it to any of the ports that are attached to that hub? Because that might be splitting the power from port #3 (you have the iPhone on port #2 in that shot).

Link to comment
Share on other sites

Nope, the iOS devices are connected to the "hub-less" ports. And yes, I have connected that additional power cable ;)

 

As for USB booting - try it, but be aware that the only way out might be pressing the reset button. Still I am curious if the same happens to you.

 

Think I will connect my iPad to the plain old Apple charger when I am not using it for debugging my Web Apps, and I think I should buy a SATA enclosure as well ...

 

PS: and you are right, the card only has one hub. Still no clue why it (partially) shows as USB2 though.

Link to comment
Share on other sites

Yeah I have the same unresponsive system here on wake with Mavericks booted from the card (in USB3 mode). Makes sense really as anything attached to the card is dead on wake there. So when the system wakes, the boot drive is lost.

 

I'd expect it to work OK if the disk in the dock had 10.11 as I don't have the same problem there.

Link to comment
Share on other sites

I'd expect it to work OK if the disk in the dock had 10.11 as I don't have the same problem there.

 

Yeah but as soon as you would boot from USB -> boom. Maybe its related to the "improperly disconnected" message you also get on real macs when ejecting USB drives. So it seems more like a OS X bug in general, and not a hack specific problem. But it's weird that you cannot sleep under OS X when being booted from an USB drive. I think that should work, but of course no big problem, except for testing sleep with an installation on an external drive.

 

Ah and I also cleared that USB Hub mystery! So its definetly one hub, kinda posing as an USB2 or USB3 hub. Now, it actually depends on what you connect to it. When I connect an USB2 device it shows under the USB2 hub entry. Connect an USB3 device on the same port, and then I see it under the USB3 hub in SysProfiler. Seems like the two outer USB ports, as well as the header show this behavior and therefore are used with this hub. Also fits to what you said. So that apparently works as intended.

Link to comment
Share on other sites

Ah and I also cleared that USB Hub mystery! So its definetly one hub, kinda posing as an USB2 or USB3 hub. Now, it actually depends on what you connect to it. When I connect an USB2 device it shows under the USB2 hub entry. Connect an USB3 device on the same port, and then I see it under the USB3 hub in SysProfiler. Seems like the two outer USB ports, as well as the header show this behavior and therefore are used with this hub. Also fits to what you said. So that apparently works as intended.

No mystery here...

 

Each USB3 port is actually two ports in one. Each has USB2 pins and USB3 pins.

 

As such, two hubs, one for USB2 and another for USB3.

Link to comment
Share on other sites

No mystery here...

 

Each USB3 port is actually two ports in one. Each has USB2 pins and USB3 pins.

 

As such, two hubs, one for USB2 and another for USB3.

 

Yeah, but it was a mistery for me ;) Never had this specific configuration before, and therefore was scratching my head. But you always learn, thanks for the insight!

Link to comment
Share on other sites

  • 2 months later...

I have a FL1100 card with two ports on the back and two internal. The card was always working fine but after upgrading to 10.11.3 it causes an instant wake from sleep. Did someone experince the same?

 

My first try was to inject clock-id, but unfortunately that doesn't work in 10.11 anymore. So my second try was to inject Name(_PRW, Package() { 0x0D, 0 }) according to RehabMan that should work for Intel but it doesn't for Fresco.

 

 

Now I found a workaround with GenericUSBXHCI.kext but I would prefer do use the native kext. Someone a hint?

Link to comment
Share on other sites

 Share

×
×
  • Create New...