Acer C720/C720P Chromebook

n8henrie · 2014-04-11 22:44:07

Touchpad also broke for me with the kernel update, but the new script fixed it (many thanks!). As the others have mentioned, coming back from suspend now causes an instant reboot. I have tried repeating all the Wiki solutions that solved the suspend issues in the first place (cros-acpi-wakeup.conf, /usr/lib/systemd/system-sleep/cros-sound-suspend.sh, and the new grub parameters / rebuilding grub) but the problem persists.

therealarchdaemon · 2014-04-12 03:02:04

I can confirm that kernel 3.14 breaks suspend (it reboots rather than resumes when you open the lid after suspend). I've reverted to kernel 3.13 as a temporary fix.

Also, here's an updated script for the touchpad drivers that should work with any 3.14.x kernel version (not just 3.14.0 :-)).

http://pastie.org/9074242

Last edited by therealarchdaemon (2014-04-12 03:05:22)

nariox · 2014-04-12 19:27:42

Me too, ok, so kernel 3.14 breaks suspend. I wonder if some of the "fixes" we used for 3.13 is actually causing this behavior. I haven't heard anything about kernel 3.14 causing these issues upstream.

n8henrie · 2014-04-12 21:01:19

I was wondering as well. Tried removing the

/etc/tmpfiles.d/cros-acpi-wakeup.conf

and

/usr/lib/systemd/system-sleep/cros-sound-suspend.sh

files as well as the extra grub boot parameters, didn't seem to change anything at all.

nariox · 2014-04-12 22:41:24

I've tried adding:

dmesg > /var/log/olddmesg

to the cros suspend script. It seems like the suspend is working correctly, nothing weird on dmesg. But resume seems to be broken. Let's wait for 3.14.1 for a solution. I wonder if other haswell processors are experiencing the same bug.

nariox · 2014-04-14 20:51:33

Quick follow-up, kernel 3.14.1 does not fix suspend.

Sn1PeR · 2014-04-16 02:42:55

I set lidswitch behavior to hibernate (using a swapfile) as a temporary workaround for suspend being broken. Resume from hibernate seems to work, but the mouse/touchscreen no longer function and a USB mouse only occasionally works after resuming.

kocsenc · 2014-04-16 15:30:11

Hey, just ran a:

pacman -Syu'

- Mouse pointer breaks again, but use the easy fix found here: http://pastie.org/9074242 (from post #127)
- Resume works again and is tremendously fast.

Last edited by kocsenc (2014-04-16 15:30:36)

n8henrie · 2014-04-16 15:37:51

Huh, hibernate was not working for me either. Is hibernate working for you with 3.14? Or 3.14.1?

masmullin · 2014-04-16 23:07:52

If I understand correctly, the haswell GPU shares memory with the CPU. Is there anyway to up the 256M allocated to the GPU?

00:02.0 VGA compatible controller: Intel Corporation Haswell-ULT Integrated Graphics Controller (rev 09) (prog-if 00 [VGA controller])
Subsystem: Acer Incorporated [ALI] Device 0a11
Flags: bus master, fast devsel, latency 0, IRQ 60
Memory at e0000000 (64-bit, non-prefetchable) [size=4M]
Memory at d0000000 (64-bit, prefetchable) [size=256M]
I/O ports at 1800 [size=64]
Expansion ROM at <unassigned> [disabled]
Capabilities: <access denied>
Kernel driver in use: i915
Kernel modules: i915

ciarlill · 2014-04-18 02:04:07

Glad I popped in here before upgrading to 3.14, I'll try and make a habit of that

On another note: for those with the 2gb version are you finding it works well on linux? For the most part it is fine, except opening anymore than 4-5 tabs in Chromium. I read a lot of docs so tend to have a habit of using a lot of tabs. I even tried setting up a swap partition figuring it might still be decent because of the SSD. Now from a little bit of research I have done it looks like Chromium/Chrome will not use swap space? ChromeOS will use "tab discarding" and refresh tabs when you return to them but this not available in Chromium. I know I could use firefox but my current devices are all pretty integrated with goog so I would like to continue using Chromium. I also considered experimenting with going back to ChromeOS with chroot. It may suite my needs better. Just curious if anyone else has run into this and knows any tweaks for Arch or Chromium that may help.

thanks

LeaveAThousand · 2014-04-18 02:33:03

Hibernate works for me. Touchpad breaks after resume, though. I tried to manually unload the module cyapa before hibernating and then reloading it back after resuming, but that didn't work.

nariox · 2014-04-18 03:18:24

@masmullin: I don't think it's possible ): These settings are usually set on BIOS. Maybe if you compile Seabios yourself you might be able to set that up manually
@ciarlill: I also find 2gb quite limited. Have you tried using zramswap? I think chromium/chrome should use swap. I don't see a reason why it wouldn't.
@kocsenc: by resume you mean suspend or hibernate?
@LeaveAThousand: it seems like the touchpad requires the i2c drivers to function correctly. I've tried removing some of the i2c-designware modules, but it was not enough. I haven't been adventurous enough to remove more i2c drivers, but the solution might be something like the ehci after suspend fix. See if you can unbind if before hibernation by echoing something to /sys/bus/i2c/drivers/cyapa/unbind and recover it with /sys/bus/i2c/drivers/cyapa/bind or maybe /sys/bus/pci/drivers/i2c-*

parchedas · 2014-04-18 10:13:13

ciarlill wrote:

Glad I popped in here before upgrading to 3.14, I'll try and make a habit of that
On another note: for those with the 2gb version are you finding it works well on linux? For the most part it is fine, except opening anymore than 4-5 tabs in Chromium. I read a lot of docs so tend to have a habit of using a lot of tabs. I even tried setting up a swap partition figuring it might still be decent because of the SSD. Now from a little bit of research I have done it looks like Chromium/Chrome will not use swap space? ChromeOS will use "tab discarding" and refresh tabs when you return to them but this not available in Chromium. I know I could use firefox but my current devices are all pretty integrated with goog so I would like to continue using Chromium. I also considered experimenting with going back to ChromeOS with chroot. It may suite my needs better. Just curious if anyone else has run into this and knows any tweaks for Arch or Chromium that may help.
thanks

I have a 2GB and haven't found any problems with Chrome and I regularly have between 20 and 40 tabs open. I have a 3GB swapfile but I'm not using any of that at the moment with 16 tabs open.

nariox · 2014-04-18 11:53:42

parchedas wrote:

I have a 2GB and haven't found any problems with Chrome and I regularly have between 20 and 40 tabs open. I have a 3GB swapfile but I'm not using any of that at the moment with 16 tabs open.

Are you using a DE? I find that most DEs take up a lot of memory (well, it's not that much if you have >2GB RAM)
Did you tweak your swappiness?

parchedas · 2014-04-18 22:17:53

nariox wrote:

Are you using a DE? I find that most DEs take up a lot of memory (well, it's not that much if you have >2GB RAM)
Did you tweak your swappiness?

Just using awesome (w/o compositing). No, I didn't change the swappiness.

gsilvers · 2014-04-20 04:04:08

How is this iGPU compared to a HD4400 or 4000?

nariox · 2014-04-20 04:35:28

gsilvers wrote:

How is this iGPU compared to a HD4400 or 4000?

According to this link: http://en.wikipedia.org/wiki/Intel_HD_Graphics
It says that it achieve 184GFlops, while the 4000 achieves 333 and the 3000 achieves 130.
But this seems to be the value for a GT1 with 10 execution units, AFAIK the Celeron 2955U has only 6 EUs, which would give something close to 110 GFlops.

According to my experience, the graphics card is pretty capable. I've played Guacamelee on 1080p (w/o AA) and it was very playable. I've tested Left4Dead 2, the FPS was about 40-50s on low, but would oscilate sometimes. Starcraft 2 on wine was playable as well, not for multiplayer, but I believe you can do the campaign.

[EDIT] In comparison to my HD3000 with a i7-2620QM, the 2955U is about the same for games. However, it might just be that the video drivers for Haswell GPUs is not as mature. We'll just have to wait to see. I'd say you'll probably be able to play any Source games and most other games that require an HD3000

Last edited by nariox (2014-04-21 02:26:56)

gothmog123 · 2014-04-20 07:25:32

Soon I'm getting this device too. Just to double check, if I am planning to make it boot into arch without Ctrl-L from the start, I need to remove the write protect screw (number 7 here), then I will have to reinstall ChromeOS because it will break from removing the screw (this is the procedure?), then set /usr/bin/set_gbb_flags.sh 0x489, then install arch.

Is this the correct procedure?

LeaveAThousand · 2014-04-20 12:58:06

gothmog123 wrote:

Soon I'm getting this device too. Just to double check, if I am planning to make it boot into arch without Ctrl-L from the start, I need to remove the write protect screw (number 7 here), then I will have to reinstall ChromeOS because it will break from removing the screw (this is the procedure?), then set /usr/bin/set_gbb_flags.sh 0x489, then install arch.
Is this the correct procedure?

Yea, that's what I did. The write protect screw turned out to be very difficult to remove with a manual screwdriver. It was easy to remove using an electric one, though. Also, my Chromebook has been emitting those high pitched sounds about which many people are complaining. I don't seem to recall hearing the sounds before I removed the screw. I haven't put back the screw. I think you can put it back without anything happening.

gothmog123 · 2014-04-20 23:31:03

Thanks. I think you are advised to put the screw back to prevent software access to the bios. You can do this right after you set the 0x489 flag, as after that it essentially becomes a normal laptop if my understanding is correct.

Last edited by gothmog123 (2014-04-20 23:31:29)

therealarchdaemon · 2014-04-21 00:02:19

Alright, here's an update on the resume from suspend issue for kernel 3.14 (no I have not found a solution yet, sorry).

My initial temporary workaround to get suspend working again was to revert to kernel 3.13, but I'm not really comfortable using an unsupported kernel for any extended period of time. So, I decided to switch to the linux-lts package (which is on kernel 3.10) until there's a proper fix or workaround for the latest kernel.

I've found that everything works fine with kernel 3.10, including suspend (as long as you make sure you've followed the instructions in the wiki, of course). To get the touchpad working, just use the script for kernel 3.12 or earlier from the wiki, and make sure you've installed linux-lts-headers before you run the script (or just modify the script to install that instead of linux-headers).

On a side note, has anyone tried filing a bug report for this yet? This does seem to be a kernel issue, since suspend still works with older kernels.

Last edited by therealarchdaemon (2014-04-21 00:04:36)

nariox · 2014-04-21 02:38:31

therealarchdaemon wrote:

Alright, here's an update on the resume from suspend issue for kernel 3.14 (no I have not found a solution yet, sorry).
My initial temporary workaround to get suspend working again was to revert to kernel 3.13, but I'm not really comfortable using an unsupported kernel for any extended period of time. So, I decided to switch to the linux-lts package (which is on kernel 3.10) until there's a proper fix or workaround for the latest kernel.
I've found that everything works fine with kernel 3.10, including suspend (as long as you make sure you've followed the instructions in the wiki, of course). To get the touchpad working, just use the script for kernel 3.12 or earlier from the wiki, and make sure you've installed linux-lts-headers before you run the script (or just modify the script to install that instead of linux-headers).
On a side note, has anyone tried filing a bug report for this yet? This does seem to be a kernel issue, since suspend still works with older kernels.

We should probably do it. I'm guessing the reason we hadn't seen any bug reports on it yet is because most distros are not using 3.14 yet.
One thing that I've done in the 3.14 kernel is disabling some patched that didn't apply. These might be the cause of our problems. Have you tried suspend on lts without applying the patches?

gsilvers · 2014-04-21 03:30:04

So this thing seems like it would work wonders for an on the go machine. Anyone mind answering a question or two?

1 - Is the 2gb of ram enough, I'm relatively spoiled with a 8gb main machine but is 2gb enough to run gnome and a web browser? I saw a 4gb touch floating around on amazon. Worth the 30 dollars extra?

2 - Is the battery life similar to what is advertised on chrome os?

n8henrie · 2014-04-21 04:49:46

I have tried suspend both with and without the (suspend) patches. No luck I haven't tried it without the trackpad patch, though. Going through the list of grub parameters from the wiki to see if any of them make a difference.

Arch Linux

#126 2014-04-11 22:44:07

Re: Acer C720/C720P Chromebook

#127 2014-04-12 03:02:04

Re: Acer C720/C720P Chromebook

#128 2014-04-12 19:27:42

Re: Acer C720/C720P Chromebook

#129 2014-04-12 21:01:19

Re: Acer C720/C720P Chromebook

#130 2014-04-12 22:41:24

Re: Acer C720/C720P Chromebook

#131 2014-04-14 20:51:33

Re: Acer C720/C720P Chromebook

#132 2014-04-16 02:42:55

Re: Acer C720/C720P Chromebook

#133 2014-04-16 15:30:11

Re: Acer C720/C720P Chromebook

#134 2014-04-16 15:37:51

Re: Acer C720/C720P Chromebook

#135 2014-04-16 23:07:52

Re: Acer C720/C720P Chromebook

#136 2014-04-18 02:04:07

Re: Acer C720/C720P Chromebook

#137 2014-04-18 02:33:03

Re: Acer C720/C720P Chromebook

#138 2014-04-18 03:18:24

Re: Acer C720/C720P Chromebook

#139 2014-04-18 10:13:13

Re: Acer C720/C720P Chromebook

#140 2014-04-18 11:53:42

Re: Acer C720/C720P Chromebook

#141 2014-04-18 22:17:53

Re: Acer C720/C720P Chromebook

#142 2014-04-20 04:04:08

Re: Acer C720/C720P Chromebook

#143 2014-04-20 04:35:28

Re: Acer C720/C720P Chromebook

#144 2014-04-20 07:25:32

Re: Acer C720/C720P Chromebook

#145 2014-04-20 12:58:06

Re: Acer C720/C720P Chromebook

#146 2014-04-20 23:31:03

Re: Acer C720/C720P Chromebook

#147 2014-04-21 00:02:19

Re: Acer C720/C720P Chromebook

#148 2014-04-21 02:38:31

Re: Acer C720/C720P Chromebook

#149 2014-04-21 03:30:04

Re: Acer C720/C720P Chromebook

#150 2014-04-21 04:49:46

Re: Acer C720/C720P Chromebook

Board footer