[Bug 285993] nvme device breakage in 14.2 STABLE n270867-25df691800f0

From: <bugzilla-noreply_at_freebsd.org>
Date: Wed, 16 Apr 2025 13:30:30 UTC
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=285993

John Baldwin <jhb@FreeBSD.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|New                         |Open

--- Comment #16 from John Baldwin <jhb@FreeBSD.org> ---
So it's kind of bizarre that pcib10 reports PFD (power fault detected) at the
very beginning of its probe.  This is why we think a card isn't present.  At
the start of attach for a pci bridge, we clear any pending events in the slot
status register (these bits are W1C):

static void
pcib_setup_hotplug(struct pcib_softc *sc)
{
...
        sc->pcie_link_sta = pcie_read_config(dev, PCIER_LINK_STA, 2);
        sc->pcie_slot_sta = pcie_read_config(dev, PCIER_SLOT_STA, 2);

        /* Clear any events previously pending. */
        pcie_write_config(dev, PCIER_SLOT_STA, sc->pcie_slot_sta, 2);


It would be handy I think to see the output of `pciconf -lc pcib10`.  The raw
register values of the slot capability register would be useful, it is at
offset 0x14 in the PCI-e capability registers.  If you aren't familiar with
using pciconf to read individual registers I can give you a pciconf command
after seeing the output of `pciconf -lc pcib10`.

-- 
You are receiving this mail because:
You are the assignee for the bug.