c2f773a9d9
When STX Patches were ported to the top of QEMU 7.2 version, the auto-merging placed our custom cli args parsing in the wrong place at softmmu/vl.c file. These extra cli arguments are related to DPDK configurations and supposed to be passed to rte_eal_init() function, called only when dpdk is enabled. This wrong placing was actually causing an error for any instance that we tried to launch, since in the new base version there is QEMU code parsing the cli args BEFORE we properly handle and remove the DPDK extra args. This change re-orders the code position inside our patch, such that DPDK arguments are properly parsed, sent to rte_eal_init() and removed from argc and argv before proceeding with QEMU upstream logic. TEST PLAN: PASS - build-pkgs -c -p qemu PASS - build-image PASS - bootstrap a virtual AIO-SX PASS - apply stx-openstack PASS - 'openstack hypervisor list' shows a QEMU hypervisor UP PASS - launch a simple VM (no volumes, single network), it is properly built and end up in Running state. Closes-Bug: 2078230 Change-Id: Id14a2328e2fb249a4de8fc8ad6302cee0b1d48ab Signed-off-by: Thales Elero Cervi <thaleselero.cervi@windriver.com> |
||
---|---|---|
.. | ||
centos | ||
debian | ||
qemu | ||
scripts | ||
.gitignore | ||
PKG-INFO | ||
README |
Get the RedHat source rpm from: $ wget wget http://cbs.centos.org/kojifiles/packages/qemu-kvm-ev/2.3.0/31.el7_2.7.1/src/qemu-kvm-ev-2.3.0-31.el7_2.7.1.src.rpm Uncompress the source rpm: $ rpm2cpio qemu-kvm-ev-2.3.0-31.el7_2.7.1.src.rpm | cpio -idmv All the patches from the redhat source rpm have been applied in git/qemu using scripts/autopatch.sh. Some of these patches can't be applied cleanly with 'git am'. So this script attempd to execute 'git am' on each of them, if it fails it tries to patch the files manually with 'git apply --index'. A git tag 'qemu-kvm-ev-2.3.0-31.el7_2.7.1' has been added to the last patch applied. The Titanium Cloud specific patches will be added after this tag.