fix(hostkernel): patch qemu_fw_config deadlock

[bleggett]

This fixes an issue where qemu_fw_config can deadlock on module load under QEMU with a PV dom0:

  watchdog: BUG: soft lockup - CPU#2 stuck for 26s! [(udev-worker):975]
  CPU: 2 UID: 0 PID: 975 Comm: (udev-worker) Tainted: G  W  6.18.44-edera-host #1 PREEMPT(none)
  RIP: e030:fw_cfg_sysfs_probe+0x7f4/0x930 [qemu_fw_cfg]
  RBX: 00000001017c9fa0   RDX: ffff8881017c9fa0
  Call Trace:
   platform_probe+0x43/0x80
   really_probe+0xda/0x2d0
   __driver_probe_device+0x87/0x170
   driver_probe_device+0x24/0xf0
   __driver_attach+0x92/0x190
   bus_for_each_dev+0x80/0xd0
   driver_attach+0x22/0x30
   bus_add_driver+0xee/0x220
   driver_register+0x62/0x120
   ? __pfx_fw_cfg_sysfs_init+0x10/0x10 [qemu_fw_cfg]
   __platform_driver_register+0x22/0x30

As far as I can tell, all of the following have to be true before this can be hit:

• The QEMU fw_cfg device exists
• The Xen dom0 is PV, and it's booted under QEMU
• CONFIG_VMCORE_INFO=y && FW_CFG_SYSFS=y in the dom0 kernel
• qemu_fw_cfg.ko is available to load