Add two upstream commits for the Google gVNIC (gve) driver:
07993df56091 gve: Update QPL page registration logic
a2f19184014f gve: Enable reading max ring size from the device in
DQO-QPL mode
Patches are taken as-is from rpms/kernel commit b4decc5e6c1b, where they
were already verified on the GCP branch. 07993df56091 needed a backport:
this 6.12 based kernel predates the split of gve_buffer_mgmt_dqo.c out
of gve_rx_dqo.c and the rename of the RX alloc cfg queue config field.
Those adjustments are documented in the patch header. a2f19184014f is
unchanged from upstream.
Both patches apply with --fuzz=0 on top of the AlmaLinux 10 kernel
6.12.0-211.38.1.el10_2 source, and the resulting driver builds clean,
including with W=1, against kernel-devel-6.12.0-211.38.1.el10_2.
