diff options
Diffstat (limited to 'results/classifier/108/other/1722857')
| -rw-r--r-- | results/classifier/108/other/1722857 | 27 |
1 files changed, 27 insertions, 0 deletions
diff --git a/results/classifier/108/other/1722857 b/results/classifier/108/other/1722857 new file mode 100644 index 000000000..1c9590bae --- /dev/null +++ b/results/classifier/108/other/1722857 @@ -0,0 +1,27 @@ +vnc: 0.644 +device: 0.613 +semantic: 0.524 +graphic: 0.519 +boot: 0.394 +socket: 0.354 +network: 0.298 +other: 0.259 +files: 0.202 +debug: 0.185 +performance: 0.156 +PID: 0.123 +permissions: 0.076 +KVM: 0.013 + +Missing PCI "programming interface" ID emulation for USB host controllers + +Qemu doesn't currently emulate the correct value of the "register level programming interface" field in the PCI config space of USB host controllers. As a result, some guest operating systems (most notably Windows) fail to load e.g. generic xHCI drivers, and instead ask for a vendor-specific driver, which may not be always available. + +Example: "-device nec-usb-xhci" emulates a Renesas (formerly NEC) uPD720202 xHCI controller. However, in the PCI config space, it shows us as class 0C, subclass 03, prog-if 00 (UHCI) where as the real uPD720202 (and all real xHCI controllers) have prog-if 30 (xHCI). A Windows guest booted with this option will not recognize devices attached to the XHCI controller unless drivers from Renesas are manually installed first, even though recent Windows versions include a generic xHCI driver that works perfectly well with real uPD720202 hardware. + +That sounds surprising ... according to the source code: +https://git.qemu.org/?p=qemu.git;a=blob;f=hw/usb/hcd-xhci.c;hb=v5.1.0#l3386 +... QEMU already sets 0x30 as programming interface byte there. Could you please double-check whether your problem still occurs with the latest version of QEMU? + +[Expired for QEMU because there has been no activity for 60 days.] + |