7074 Commits

Author SHA1 Message Date
David Michael
983633e00d sys-kernel/coreos-sources: Bump 4.14.20 to 4.14.21 2018-02-22 16:14:56 -05:00
David Michael
3bb5d20e73 Merge pull request #3084 from glevand/for-merge-containerd
app-emulation/containerd: Fix GOARCH build problem
2018-02-21 19:44:45 -05:00
Geoff Levand
663b883809 app-emulation/containerd: Fix GOARCH build problem
A call to coreos-go_src_prepare in src_prepare will get go_export
called, which exports the correct GOARCH value.

Fixes https://github.com/coreos/bugs/issues/2366 (torcx store has
archives of wrong arch).

Signed-off-by: Geoff Levand <geoff@infradead.org>
2018-02-21 15:58:44 -08:00
David Michael
37a9734690 Merge pull request #3082 from dm0-/mirror
app-emulation/containerd: Fix the 1.0.* manifest entries
2018-02-21 13:14:18 -05:00
David Michael
c9bf58d918 app-emulation/containerd: Fix the 1.0.* manifest entries
The entries were defined for the release archives as Gentoo's
mirrors were serving, but the SRC_URIs were pointing at commit
archives.  This syncs our SRC_URI definition to match Gentoo's so
that the same sources are used and the Manifest verification works.
There is no difference in what gets built.
2018-02-21 12:45:19 -05:00
David Michael
84fb491b97 Merge pull request #3083 from dm0-/linux
sys-kernel/coreos-modules: Build the smartpqi driver
2018-02-21 07:58:54 -05:00
David Michael
64707fc097 sys-kernel/coreos-modules: Build the smartpqi driver
Requested for the HPE Smart Array P408i-c SR Gen10 Controller.
2018-02-20 19:32:48 -05:00
David Michael
58a7ac5ccd Merge pull request #3078 from dm0-/linux
Backport an iSCSI fix to Linux 4.14 in master
2018-02-20 16:05:04 -05:00
David Michael
4ca0d3dc75 sys-kernel/coreos-sources: Backport an iSCSI fix to 4.14 2018-02-20 15:36:22 -05:00
David Michael
4a89f7180b Merge pull request #3074 from dm0-/linux
Bump Linux 4.14.19 to 4.14.20 and 4.15.3 to 4.15.4 in master
2018-02-20 15:00:13 -05:00
David Michael
59c6605e6c Merge pull request #3073 from dm0-/go
Bump etcd and Go
2018-02-20 14:59:51 -05:00
David Michael
d3dc2a91cf sys-kernel/coreos-sources: Bump 4.14.19 to 4.14.20 2018-02-17 22:23:45 -05:00
David Michael
15554fac6e sys-kernel/coreos-sources: Bump 4.15.3 to 4.15.4 2018-02-17 22:20:00 -05:00
David Michael
5b66cf1730 chore(metadata): Regenerate cache 2018-02-16 17:00:31 -05:00
David Michael
42fb67306e eclass/coreos-go-depend: Set the default Go version to 1.10 2018-02-16 17:00:02 -05:00
David Michael
cf078d2478 dev-lang/go: Add Go 1.10 2018-02-16 16:57:47 -05:00
David Michael
88ed998c59 app-admin/etcd-wrapper: Bump 3.2.15 to 3.3.1 2018-02-16 16:52:18 -05:00
David Michael
12aca7b5bb dev-db/etcdctl: Bump 3.2.15 to 3.3.1 2018-02-16 16:51:21 -05:00
Andrew Jeddeloh
0a794eda1d Merge pull request #3062 from ajeddeloh/bump-google-cloud-sdk
net-misc/google-cloud-sdk: bump to 188.0.1
2018-02-15 15:15:54 -08:00
David Michael
263802e8eb Merge pull request #3071 from dm0-/linux
Add Linux 4.15.3
2018-02-14 07:35:06 -05:00
David Michael
04f2fa213e sys-kernel/coreos-sources: Bump 4.15.2 to 4.15.3 2018-02-13 19:58:00 -05:00
David Michael
07ce3624b3 Revert "sys-kernel/coreos-sources: Drop Linux 4.15"
Re-add the 4.15 kernel to continue testing on the master branch.

This reverts commit 2115be5c222890a9a9927d3cb0d593a9fd610bea.
2018-02-13 19:56:37 -05:00
David Michael
015fc75c67 Merge pull request #3070 from dm0-/linux-4.14
Bump 4.14.18 to 4.14.19 in master
2018-02-13 18:19:09 -05:00
Benjamin Gilbert
f334fa190e Merge pull request #3072 from bgilbert/mantle
coreos-devel/mantle: bump to 0.8.2
2018-02-13 15:14:10 -08:00
Benjamin Gilbert
df1bc157fd coreos-devel/mantle: bump to 0.8.2 2018-02-13 15:05:07 -08:00
David Michael
0410d6a401 sys-kernel/coreos-sources: Bump 4.14.18 to 4.14.19 2018-02-13 08:42:21 -05:00
Benjamin Gilbert
f469635349 Merge pull request #3067 from bgilbert/mantle
coreos-devel/mantle: bump to 0.8.0
2018-02-12 17:41:25 -08:00
Benjamin Gilbert
f18eed517f coreos-devel/mantle: bump to 0.8.0 2018-02-12 17:34:37 -08:00
David Michael
63797e6513 Merge pull request #3064 from dm0-/linux
sys-kernel/coreos-sources: Drop Linux 4.15
2018-02-11 17:07:00 -05:00
David Michael
c4aff6f546 sys-kernel/coreos-sources: Drop Linux 4.15
It fails to boot on AWS PV, so fall back to 4.14 for the next
release.  This can be reverted after tagging to resume testing.
2018-02-11 10:48:04 -05:00
David Michael
ef39090cd9 Merge pull request #3059 from dm0-/gcc
Upgrade GCC to 7.3.0
2018-02-11 10:43:56 -05:00
David Michael
4cbbb60492 chore(metadata): Regenerate cache 2018-02-10 18:33:55 -05:00
David Michael
aaf560f909 dev-lang/rust: Apply CoreOS changes
Prevent build failures with GCC 7 due to -Wimplicit-fallthrough.
2018-02-10 18:32:59 -05:00
David Michael
227218a3a5 dev-lang/rust: Import from portage-stable 2018-02-10 18:32:37 -05:00
David Michael
6ddd960c07 sys-kernel/coreos-sources: Disable -Werror as needed for GCC 7.3 2018-02-10 18:15:52 -05:00
David Michael
aceb94d63a profiles: Save build time by masking gcc-6.4.0-r1 2018-02-10 18:14:48 -05:00
David Michael
59d1a667f0 profiles: Match the arm64 board GCC build with the SDK 2018-02-10 18:14:48 -05:00
David Michael
131ceb6ec8 profiles: Disable building unused C++ libraries for ncurses 2018-02-10 18:14:48 -05:00
David Michael
65cf2cd8ff profiles: Upgrade to GCC 7.3 for retpoline support 2018-02-10 18:14:48 -05:00
Benjamin Gilbert
3c37b1db90 Merge pull request #3063 from bgilbert/key
coreos-base/coreos-init: bump for new signing subkey
2018-02-10 12:00:16 -08:00
Benjamin Gilbert
621ff9fc8d Merge pull request #2971 from ezequielgarcia/bfq
kernel: Enable BFQ I/O scheduler (for blk-mq)
2018-02-09 21:49:51 -08:00
Benjamin Gilbert
feb3c835ea coreos-base/coreos-init: bump for new signing subkey 2018-02-09 21:31:43 -08:00
Ezequiel Garcia
cbe1d6973d sys-kernel: enable BFQ I/O scheduler
BFQ is a proportional-share storage-I/O scheduler for the
blk-mq layer. Its main features are:

 * Low latency for interactive applications
 * Low latency for soft real-time applications
 * Higher speed for code-development tasks
 * High throughput
 * Strong fairness, bandwidth and delay guarantees

Note that the default blk-mq scheduler is "None", so
th BFQ scheduler must be explicitly chosen for a given
storage device.

Note also that BFQ only supports blk-mq layer, so drivers
using legacy blk layer are not affected by this change.

In particular, both HDD and SSD storage drives use
the legacy blk layer by default.

Signed-off-by: Ezequiel Garcia <ezequiel@vanguardiasur.com.ar>
2018-02-10 01:33:03 -03:00
Andrew Jeddeloh
29935edc95 net-misc/google-cloud-sdk: bump to 188.0.1
Bump google-cloud-sdk to latest. Remove python3 files from yet another
place in src_prepare. Do not "doman" since there are no longer any man
pages.
2018-02-09 17:07:56 -08:00
David Michael
56f138af4a Merge pull request #3061 from dgonyeo/ign-22
sys-apps/ignition: v0.21.0 -> v0.22.0
2018-02-09 18:43:29 -05:00
Derek Gonyeo
5cedac709c sys-apps/ignition: v0.21.0 -> v0.22.0 2018-02-09 15:38:03 -08:00
Andrew Jeddeloh
32e76a5bcf Merge pull request #3060 from ajeddeloh/gcs-pyopenssl
net-misc/google-cloud-sdk: depend on pyopenssl
2018-02-09 13:17:14 -08:00
David Michael
e363ee4cd2 Merge pull request #2970 from ezequielgarcia/master
kernel: Enable throttled background buffered writeback (CONFIG_BLK_WBT)
2018-02-09 15:03:28 -05:00
Andrew Jeddeloh
693d33b4a0 net-misc/google-cloud-sdk: depend on pyopenssl
pyopenssl is needed for gsutil signurl.
2018-02-09 10:52:50 -08:00
Ezequiel Garcia
b36ecb3403 sys-kernel: enable throttled background buffered writeback
This feature prevents the background writeback thread from stalling I/O,
e.g. reads.

For instance, without this option enabled, it's possible to stall reads
setting a writeback dirty threshold (/proc/sys/vm/dirty_writeback_centisecs)
sufficiently long.

Quoting the author of the feature:

"""
[..] If you are sick of Linux bogging down when buffered
writes are happening, then this is for you, laptop or server. The
patchset is fully stable, I have not observed problems. It passes full
xfstest runs, and a variety of benchmarks as well. It works equally well
on blk-mq/scsi-mq, and "classic" setups.
"""

"""
This feature adds a simple blk-wb code that keeps limits how much buffered
writeback we keep in flight on the device end. [..] it should be pretty
much auto-tuning.
"""

Signed-off-by: Ezequiel Garcia <ezequiel@vanguardiasur.com.ar>
2018-02-09 13:36:22 -03:00