Apple's 2016 MacBook Pro & Linux Don't Mix

[starshipeleven]

How would you parse it? x8 with each lane 2,5GT/s or as 2,5GT/s summed up?

Now that's weird.
It says PCI Express (blah blah) version 2 on both, so it should theoretically be able to run as 5 GT per lane.

I'm 100% sure that the card in the minipcie slot is a pcie 1.0 controller itself because there is no sane reason for it to have a more powerful pcie controller onboard, so the 2.5 GT per lane will mean that the card is running at PCIe 1.0, as the 1.0 runs at 2.5 per lane.

For the GPU I don't know, maybe that GPU has only PCIE 1.0 controller onboard so it is running at 2.5 per lane too?

Or maybe it's just nonsense as the kernel didn't read stuff right somehow.

Anyway, my laptop's output of lspci -vv (only relevant parts are shown, it dumps a TON of text):

Capabilities: [70] Express (v2) Endpoint, MSI 00
LnkCap: Port #0, Speed 2.5GT/s, Width x1, ASPM L0s L1, Exit Latency L0s <4us, L1 <64us
LnkSta: Speed 2.5GT/s, Width x1, TrErr- Train- SlotClk+ DLActive- BWMgmt- ABWMgmt-
LnkCtl2: Target Link Speed: 2.5GT/s, EnterCompliance- SpeedDis-

In my desktop PC where I know there is a 2.0 x1 port and I know the card in it is a 2.0 x1 card (a NEC usb 3.0 controller) it says

Capabilities: [40] Express (v2) Root Port (Slot+), MSI 00
LnkCap: Port #8, Speed 5GT/s, Width x1, ASPM L0s L1, Exit Latency L0s <512ns, L1 <4us
LnkSta: Speed 5GT/s, Width x1, TrErr- Train- SlotClk+ DLActive+ BWMgmt+ ABWMgmt-
LnkCtl2: Target Link Speed: 5GT/s, EnterCompliance- SpeedDis-


By reading this, LnkCap is the port's max speed, so my laptop has a 1.0 x1 port even if the APU in it has a pcie 2.0 controller, maybe it is set by board firmware (in the sense that the board firmware sets that on boot and I can't change it)?

https://community.mellanox.com/docs/DOC-2496

You might want to run lspci -vv on your system as that tool should be on BSDs too.

[aht0]

ppb0 at pci0 dev 3 function 0: vendor 0x1022 product 0x1708 (rev. 0x00)
ppb0: PCI Express capability version 2 <Root Port of PCI-E Root Complex> x8 @ 2.5GT/s
pci1 at ppb0 bus 1
pci1: i/o space, memory space enabled, rd/line, wr/inv ok
radeon1 at pci1 dev 0 function 0: vendor 0x1002 product 0x6741 (rev. 0x00)
ppb1 at pci0 dev 4 function 0: vendor 0x1022 product 0x1709 (rev. 0x00)
ppb1: PCI Express capability version 2 <Root Port of PCI-E Root Complex> x1 @ 2.5GT/s
pci2 at ppb1 bus 2
pci2: i/o space, memory space enabled, rd/line, wr/inv ok

How would you parse it? x8 with each lane 2,5GT/s or as 2,5GT/s summed up?
portion of boot log from NetBSD.

[aht0]

Ummm, yeah, keep dreaming. http://pinoutguide.com/Slots/mini_pcie_pinout.shtml https://en.wikipedia.org/wiki/PCI_Ex...ress_Mini_Card
It theoretically has contacts for a second Pcie lane but I've never ever seen them connected in laptops, so it is one lane physically, period.

I said 1.0 because at the times the controller in the laptop was 99% likely 1.0, nowadays it could be 2.0 or even 3.0 too.... hmmm.... Need to do some tests with my laptop, it's from 2013, it should have a 2.0 controller at least.
But considering that 99.999% of wifi cards don't saturate pcie 1.0, I doubt they allow the port to support more than that, due to power budget or for limiting the expandability or whatever.

Okay, you appear to be correct.

One of my laptops reported "4x Mini-PCIe" in dmesg some months a go. But I could neither remember which OS I was playing with at the time nor which exact machine and after couple of days, I gave up trying reproduce it.

[starshipeleven]

4x PCIe 2.0 as well.

Ummm, yeah, keep dreaming. http://pinoutguide.com/Slots/mini_pcie_pinout.shtml https://en.wikipedia.org/wiki/PCI_Ex...ress_Mini_Card
It theoretically has contacts for a second Pcie lane but I've never ever seen them connected in laptops, so it is one lane physically, period.

I said 1.0 because at the times the controller in the laptop was 99% likely 1.0, nowadays it could be 2.0 or even 3.0 too.... hmmm.... Need to do some tests with my laptop, it's from 2013, it should have a 2.0 controller at least.
But considering that 99.999% of wifi cards don't saturate pcie 1.0, I doubt they allow the port to support more than that, due to power budget or for limiting the expandability or whatever.

[aht0]

It's a fuckton of bandwith, back then you had what, miniPCIe which was a x1 PCIe 1.0? Also Expresscards. NEVER FORGET EXPRESSCARDS.

4x PCIe 2.0 as well.
ExpressCard 2.0 (2009+) looks actually to have usable performance but have yet to see any laptop with 2.0 slot. Waste of space compared to USB connectors

[Jesse2004]

Apple is real shit. Just bought a dell xps 13 with kaby lake processor.

[Passso]

probaly because 16:10 are stupid, we don't need more vertical space but more horizontal we scroll things from up to down not from left to right

God. Give me the strengh to believe this is irony. Please God.

[Passso]

because they are the LAST vendor that offers a 16:10 screen! Am I the only one who misses vertical space at a 13" 16:9 laptop?

Gotta like it 100x!

MS Surface is 16/10 too BTW but you cannot run Linux on it neither...

[juno]

Heh, i hoped to see some results of that low watt RX 460... revision ef

Yeah, as that's the interesting part in there. Not the pseudo touchscreen, all the soldered parts etc.

[spykes]

The door is open if some wants to write drivers for it.

You mean Apple actually provides technical specifications for their hardware ?