hi,
first of all, this isn't meant to be a distro trolling discussion, there
are better places for that ;-)
several days ago google HO started to fail after updating my debian/sid
setup. as discussed on G+ [1] and on googlegroups [2], that seemed to be
impacting anyone on debian/sid. when i realized that it was working on
debian/testing (Jessie) i decided just use 'testing'... however since last
evening update on 'testing', it started to fail too.. leaving me without
any working HO at all, which is quite inconvenient at Linaro..
after bisecting my system using snapshot.debian.org, i figured out which
package was the culprit:
# apt-cache policy libcairo2
libcairo2:
Installed: 1.12.14-4
Candidate: 1.12.14-4
Package pin: 1.12.14-4
Version table:
1.12.16-2 1001
500 http://ftp.fr.debian.org/debian/ testing/main amd64 Packages
*** 1.12.14-4 1001
500 http://snapshot.debian.org/archive/debian/20130927T214600Z/testing/main
amd64 Packages
100 /var/lib/dpkg/status
Simply downgrading libcairo2 [3] 'fixes' the problem. I know it's not
'fixing' it really, but i hope that this will be enough information so that
the real fix is uploaded.
In the mean time, if any of use use Debian , you can use that as a
workaround..
cheers
nico
[1] https://plus.google.com/u/0/108790110407014289037/posts/5q22d1ReavG
[2] https://productforums.google.com/d/msg/hangouts/vYsaeEnXJXs/hOy2HaW5LKoJ
[3]
http://snapshot.debian.org/archive/debian/20130927T214600Z/pool/main/c/cair…
Hi all,
I'd like to announce the availability of a preview/prototype of Go for
arm64 platforms.
This is based around gccgo but includes a version of the 'go' tool for
your familiar "go get" experience.
This release is based on the latest Ubuntu development version, Trusty
Tahr and contains gccgo 4.8 (implementing Go 1.1, and containing all
bugfixes I am aware of) and a version of the Go tool built from:
https://github.com/mwhudson/faux/tree/v1.1.2
(This is a fork of cmd/go from Go 1.1.2 with some bugfixes and
enhancements to work better when used with gccgo).
The easiest way to get started is to download the image I built for
use with the foundation model which is available here:
http://people.linaro.org/~mwhudson/go-preview/go-tool-preview-rootfs.img.xz
To use the image, follow the instructions from
https://wiki.ubuntu.com/ARM64/FoundationModel
but use my .img rather than the ubuntu-core one, and log in with
ubuntu/ubuntu rather than root/root.
If you just want to see the go tool itself, you can download binaries:
http://people.linaro.org/~mwhudson/go-preview/gohttp://people.linaro.org/~mwhudson/go-preview/go-static
("go" links against libgo.so.4, i.e. libgo from gccgo 4.8, and
"go-static" is, I hope unsurprisingly, statically linked).
I've also uploaded the rootfs I built the image out of:
http://people.linaro.org/~mwhudson/go-preview/go-tool-preview-rootfs.tar.xz
It ought to be possible to use this with qemu as described at:
https://wiki.debian.org/Arm64Qemu
but when I tried it segfaulted. YMMV!
I'd love to hear your experiences using this, I'm sure there are bugs
aplenty, but we have to start somewhere!
Cheers,
mwh
Hi Kevin,
I have updated the script you shared here:
https://wiki.linaro.org/WorkingGroups/PowerManagement/Doc/AdaptiveTickless?…
to get the time for which cpu wasn't getting interrupted.
Probably this might be useful for few other folks as well and so sharing it..
I am bad at scripting, so please ignore silly code in there :)
So, the script will take a parameter, i.e. number of samples you
need and will output the amount of time CPU1 (of a dual-core
machine) was isolated.
There is a variable "isdebug", which can be used to enable/disable
debug messages.
Let me know if there is something really stupid in it :)
--
viresh
Hi,
I am doing some work to optimize sqlite to utilize a GPU for acceleration.
One thing missing, I really should have some "reasonable" populated
sqlite databases of some amount of size. I can create contrived
examples but I think it might be advantageous to utilize some real
world sqlite databases. This presumes of course that it doesn't have
sensitive data and can be shared.
Appreciate any pointers. Thanks!
--
Regards,
Tom
"Where's the kaboom!? There was supposed to be an earth-shattering
kaboom!" Marvin Martian
Tech Lead, Graphics Working Group | Linaro.org │ Open source software
for ARM SoCs
w) tom.gall att linaro.org
h) tom_gall att mac.com
Is there any rational why flush_cache_louis is necessary in cpu_suspend()
funciton ?
After saving the context it does DCCIMVA ie writes to POC the affected
lines.
So why we should clean the entire L1 here ?
It rather should be left to the platform dependent function called after
__cpu_suspend ?
flush_cache_louis does not disable the d cache. This is not the
recommended sequence
for core power down.
Thanks
Sandeep Tripathy
Broadcom