The issue here is if an previous update succeeded but reboot not happen, for what ever unknown reason, this host lost the ability to update again.
I think some of the current update failure in the lab should be caused by this.
Review URL: http://codereview.chromium.org/2031002
Make standalone package for factory install, so it can be overlayed on top of an existing image.
Modify mod_image_for_test to do the overlay.
Review URL: http://codereview.chromium.org/1945004
A separate CL adds a test stanza to the metrics ebuild. More tests
(specifically, for the D-Bus MessageFilter) will also come in a
separate CL.
Review URL: http://codereview.chromium.org/1919005
Add restart_in_chroot_if_needed to common.sh, and modify the build scripts which referred to assert_inside_chroot to use it instead. The effect is that you don't ever have to explicitly enter_chroot.sh to build (still can, it work's fine).
Update mod_image_for_test.sh to use restart_in_chroot_if_needed
Review URL: http://codereview.chromium.org/1736025
TEST=Tested with image_to_live, memento_updater without image_to_live
This mimics the update process of the real image for the stateful partition.
By moving this the update into a separate script, I can add it to the gmerge
package such that developers will not have to use image_to_live. This change
also removes the dependency on tar.
Review URL: http://codereview.chromium.org/1774021
on the device under test and to use a not-normally-generated interface name, so that use of add-on ethernet devices is not affected in normal operation.
BUG=2962, 2941
(dup of 1810013, git goofiness)
Review URL: http://codereview.chromium.org/1826003
TEST=Tested by running image_to_live with correct formated images as well
as images that are missing either /var or dev_image directories.
Review URL: http://codereview.chromium.org/1705023
This allows appending extra kernel command line parameters to
the ARM kernel command line on the Voguev210 boards.
BUG=None
TEST=Built image
Review URL: http://codereview.chromium.org/1694021
* Leave ui.conf there, not run by default.
* dump autotest log into /dev/tty1
* Don't let tty1 eat keayboard events.
Review URL: http://codereview.chromium.org/1742018
Dropping caches (writing back all dirty pages to disk and clearing all the caches) is not necessary for factory tests. It can reduce the latency significantly. The scale is about 27s -> 3s, for running 2 empty tests.
Review URL: http://codereview.chromium.org/1780011
This adds some '/dev/sdb' options to the grub2 menu, so that we can keep the
default boot partition /dev/sda3 while still using the USB key as a recovery
image. This is work in progress for EFI BIOS development.
Review URL: http://codereview.chromium.org/1706014
This is useful for various JTAG and fastboot based flashing of
targets. It is also useful in that we will now have access to
USB card readers and mass storage devices inside the chroot.
BUG=None
TEST=Enter and exit the chroot cleanly.
Review URL: http://codereview.chromium.org/1711016
Choosing a default device is no good; the user should
tell us where the image should be written.
BUG=none
TEST=wrote an image
Review URL: http://codereview.chromium.org/1730012
Undoes a change from 5c3b457f. Hardcoding "--yes" in the
call to mod_image_for_test.sh instead of inheriting
image_to_usb.sh's "yes" was intentional. The user may not
pass --yes to image_to_usb.sh since they want to confirm
that they're writing to the correct device, but they already
told us that they want the image to be modified for test
with --test_image, so there's no reason to ask about it
again.
BUG=none
TEST=ran it
Review URL: http://codereview.chromium.org/1736005
Migrate only issues that are open or that are already in the
spreadsheet. This way, the spreadsheet doesn't get cluttered w/ closed
tasks.
Review URL: http://codereview.chromium.org/1734001
After discussing with drewry, we can't come up with a better way to inject these root certs. We considered putting them on the stateful partition, but that opens up an avenue of attack (if you can get a root cert into the magic directory, then you can MITM login). Thus, we put it on the rootfs instead. The script that sets up the hashes for vboot will verify that this directory is not present in production images. That work is tracked here: http://code.google.com/p/chromium-os/issues/detail?id=2693
Review URL: http://codereview.chromium.org/1566055
Add feature to mod_image_for_test to patch rootfs.
Change initctl path to get network but no chrome.
./image_to_usb.sh --install_mfg --install_autotest --test_image
which calls ./mod_image_for_test.sh --manuf
Review URL: http://codereview.chromium.org/1542011
As with build_packages, only retries by default if you passed --jobs
even though build_image will indeed spuriously fail with only one job.
sync_build_test now uses jobs=#cores by default for both package and
image building. Eventually will also use this for gclient sync'ing
once that is better vetted.
Review URL: http://codereview.chromium.org/1564035
