Macbook Air 2013

neptuneK · 2013-09-23 00:11:16

While they are fixing the the speaker thingy upstream, the impatient can try:
> sudo hda-verb /dev/snd/hwC1D0 0x1 set_gpio_data 1
(auto-mute doesn't work.)

Update:
Like the brightness control, this does not work after resuming from suspend.

Last edited by neptuneK (2013-09-23 00:41:39)

MrMen · 2013-09-23 12:45:17

@hobarrera : thanks for this answer.

albertalbert · 2013-09-24 06:25:35

oboenerd wrote:

Your local pacman databases are out-of-date. Always perform a full pacman -Syu before installing a package group.
You shouldn't need to install this package group. Try removing 'synaptics' from the identifier line in your synaptics.conf. Also make sure you have an EndSection at the end of the file.
EDIT: Now I see what the problem could be....
https://bbs.archlinux.org/viewtopic.php?id=165854
https://bugzilla.kernel.org/show_bug.cgi?id=60181
Try updating your system (kernel) and see if it works.

Humm I ran pacman -Syu and I believe it installed a new kernel... Now booting I see "3.11.1-1-ARCH"
But I'm not positive about that...

I do know that now my wireless doesn't work, and it used to. Also booting with an ethernet cable plugged in (via usb adapter) doesn't get me online like it used to. A bit confused, didn't think just updating packages could cause this much fuss.

hobarrera · 2013-09-24 07:15:51

Run "uname -a" to know which kernel you're running.

Also make sure you enabled dkms, since I belive that rebuilds the wireless driver after an upgrade (I'm not 100% familiar with dkms, but I do know that you need to enable it for the driver needed for this laptop).

No issues here with 3.11.1-2-ARCH.

albertalbert · 2013-09-24 08:09:44

hobarrera wrote:

Run "uname -a" to know which kernel you're running.
Also make sure you enabled dkms, since I belive that rebuilds the wireless driver after an upgrade (I'm not 100% familiar with dkms, but I do know that you need to enable it for the driver needed for this laptop).
No issues here with 3.11.1-2-ARCH.

Thanks for the info I'll try that first thing tomorrow. What does the 2 signify in 3.11.1-2-ARCH ? And how would I upgrade to that?

Oh yeah just remembered I was trying to get the kernel version by typing uname -r for some reason

Last edited by albertalbert (2013-09-24 08:13:14)

karihre · 2013-09-24 20:12:48

albertalbert wrote:

Thanks for the info I'll try that first thing tomorrow. What does the 2 signify in 3.11.1-2-ARCH ? And how would I upgrade to that?

2 signifies the second package of kernel 3.11.1, usually with only minor changes. I believe this package is still in testing (see https://www.archlinux.org/packages/core/x86_64/linux/) so it will arrive soon.

hobarrera wrote:

1 out of about, 10 times, I've been seeing refind freeze and need to reboot. Haven't been able to find out what the issue is. VERY ocassional though.

After looking at this again tonight I think I might be onto something. If I remove the "initrd=\initramfs-linux.img" parameter, the kernel starts (but obviously quickly fails), so I'm wondering if it has something to do with my partition setup. Currently I have /dev/sda2 as /boot (ext2) and /dev/sda1 as /boot/efi (vfat). When I enter the refind menu I see the option "Boot vmlinuz-linux from 200 MiB ext2 volume" which includes the kernel boot parameters specified in /boot/refind-linux.conf with the addition of the "initrd=\initramfs-linux.img" parameter. Since this exact boot line is nowhere specified in any configuration file under the boot or efi partitions, I assume it's coming from some EFI variables? Can I be sure that the initramfs-linux.img is loaded from the correct place?

I tried to copy the vmlinuz-linux and initramfs-linux.img files to the efi partition as well, which resulted in a new item appearing in the refind menu at start (showing "boot vmlinuz-linux from EFI"). This actually booted but I had to manually alter the kernel parameters (which only contained the initrd part) which is tedious, is there a way to alter the default parameters of this entry? The existing (ext2) option remained unbootable, so it doesn't seem to be looking for the initramfs file in the EFI partition.

I wondering if I should use ext3/4 instead of ext2, or simply throw out the ext2 boot partition and use the efi vfat partition for /boot. Any ideas? How are you refined people doing it?

Thanks.

hobarrera · 2013-09-24 22:15:34

karihre wrote:

After looking at this again tonight I think I might be onto something. If I remove the "initrd=\initramfs-linux.img" parameter, the kernel starts (but obviously quickly fails), so I'm wondering if it has something to do with my partition setup. Currently I have /dev/sda2 as /boot (ext2) and /dev/sda1 as /boot/efi (vfat). When I enter the refind menu I see the option "Boot vmlinuz-linux from 200 MiB ext2 volume" which includes the kernel boot parameters specified in /boot/refind-linux.conf with the addition of the "initrd=\initramfs-linux.img" parameter. Since this exact boot line is nowhere specified in any configuration file under the boot or efi partitions, I assume it's coming from some EFI variables? Can I be sure that the initramfs-linux.img is loaded from the correct place?
I tried to copy the vmlinuz-linux and initramfs-linux.img files to the efi partition as well, which resulted in a new item appearing in the refind menu at start (showing "boot vmlinuz-linux from EFI"). This actually booted but I had to manually alter the kernel parameters (which only contained the initrd part) which is tedious, is there a way to alter the default parameters of this entry? The existing (ext2) option remained unbootable, so it doesn't seem to be looking for the initramfs file in the EFI partition.
I wondering if I should use ext3/4 instead of ext2, or simply throw out the ext2 boot partition and use the efi vfat partition for /boot. Any ideas? How are you refined people doing it?
Thanks.

Looks like you're missing the boot parameters (I'm guessing based on your comment above), and that's why it fails.
Run "refind-mkrlconf". This will create "/boot/refind_linux.conf". Add the missing parameters here. As a general reference, this file needs to be placed in the same directory as the image you're booting.

patrikjakobsson · 2013-09-25 19:41:00

neptuneK wrote:

I also get this error after resuming from suspend:
[ 4816.454350] ata1: SATA link up 1.5 Gbps (SStatus 113 SControl 310)
[ 4816.454995] ata1.00: unexpected _GTF length (8)
[ 4816.455587] ata1.00: unexpected _GTF length (8)
[ 4816.455598] ata1.00: configured for UDMA/33
[ 4816.455697] ata1: EH complete
[ 4816.555824] ata1: exception Emask 0x10 SAct 0x0 SErr 0x10000 action 0xe frozen
[ 4816.555834] ata1: irq_stat 0x00400000, PHY RDY changed
[ 4816.555839] ata1: SError: { PHYRdyChg }
[ 4816.555850] ata1: hard resetting link
I tried reformatting my the partition (ext4), use the mount option "discard", and changed the default
IO scheduler to deadline, and use noncq in kernel parameters. Does anyone else have this issue ?
The message does not seem to reduce usability; but I don't know if it is killing the SSD.

I'm also getting this error after suspend/resume (and also after hibernate) on kernel 3.12-rc1 with Ubuntu 13.04 and 13.10 (I know this is the ARCH forum). I've tried both BIOS, UEFI install and all combinations of kernel parameters. I've even hacked around in the kernel to see if there are any marvell specific quirks that works on this system. The error message indicates ata bus error and it's not noticable at first but after a while the sata performance degrades to unusable levels. I'm very curious why I'm the only one hitting this issue on never kernels. I've actually had a successful suspend/resume so it seems that it's not 100% reproducable.

Note that the libata=1:noncq parameter doesn't help at all and that it is correctly picked up by the kernel.

albertalbert · 2013-09-27 00:28:31

hobarrera wrote:

Run "uname -a" to know which kernel you're running.
Also make sure you enabled dkms, since I belive that rebuilds the wireless driver after an upgrade (I'm not 100% familiar with dkms, but I do know that you need to enable it for the driver needed for this laptop).
No issues here with 3.11.1-2-ARCH.

$ uname -a
Linux archer 3.11.1-1-ARCH #1 SMP Preempt SAT SEP 14 19:30:21 CEST 2013 x86_64 GNU/Linux

Before I had 3.10.10-1

Still can't get the ethernet to work, which is strange. After booting and logging in I ping and get nothing so I run `sudo dhcpcd` and it seems to set it up fine, get an ip but still ping doesn't work, browser either. Is there something I'm missing, or a way to maybe roll-back?

hobarrera · 2013-09-27 03:22:18

albertalbert wrote:

hobarrera wrote:
Run "uname -a" to know which kernel you're running.
Also make sure you enabled dkms, since I belive that rebuilds the wireless driver after an upgrade (I'm not 100% familiar with dkms, but I do know that you need to enable it for the driver needed for this laptop).
No issues here with 3.11.1-2-ARCH.
$ uname -a
Linux archer 3.11.1-1-ARCH #1 SMP Preempt SAT SEP 14 19:30:21 CEST 2013 x86_64 GNU/Linux
Before I had 3.10.10-1
Still can't get the ethernet to work, which is strange. After booting and logging in I ping and get nothing so I run `sudo dhcpcd` and it seems to set it up fine, get an ip but still ping doesn't work, browser either. Is there something I'm missing, or a way to maybe roll-back?

Sounds like the ethernet adapter itself is working, and your issue is elsewhere. Is the IP you're getting in the right range? Does DNS resolve? Are you pinging a local IP, or something like, say, google? What's the output of ifconfig after connecting?

albertalbert · 2013-09-27 18:49:38

hobarrera wrote:

albertalbert wrote:
hobarrera wrote:
Run "uname -a" to know which kernel you're running.
Also make sure you enabled dkms, since I belive that rebuilds the wireless driver after an upgrade (I'm not 100% familiar with dkms, but I do know that you need to enable it for the driver needed for this laptop).
No issues here with 3.11.1-2-ARCH.
$ uname -a
Linux archer 3.11.1-1-ARCH #1 SMP Preempt SAT SEP 14 19:30:21 CEST 2013 x86_64 GNU/Linux
Before I had 3.10.10-1
Still can't get the ethernet to work, which is strange. After booting and logging in I ping and get nothing so I run `sudo dhcpcd` and it seems to set it up fine, get an ip but still ping doesn't work, browser either. Is there something I'm missing, or a way to maybe roll-back?
Sounds like the ethernet adapter itself is working, and your issue is elsewhere. Is the IP you're getting in the right range? Does DNS resolve? Are you pinging a local IP, or something like, say, google? What's the output of ifconfig after connecting?

Yeah it looks like it's giving me an IP that's within range, though I don't have complete control over the network. Either way it worked fine before the upgrade.
Once I boot and login I run dhcpcd and get this:

$ sudo dhcpcd
dhcpcd[253]: version 6.0.5 starting
dhcpcd[253]: enp0s20u2: waiting for carrier
dhcpcd[253]: enp0s20u2: carrier acquired
dhcpcd[253]: enp0s20u2: soliciting an IPv6 router
dhcpcd[253]: enp0s20u2: rebinding lease of 192.168.1.21
dhcpcd[253]: enp0s20u2: NAK: from <removed IP>
dhcpcd[253]: enp0s20u2: soliciting an IPv6 router
dhcpcd[253]: enp0s20u2: soliciting a DHCP lease
dhcpcd[253]: enp0s20u2: offered 192.168.1.21 from 192.168.1.5
dhcpcd[253]: enp0s20u2: leased 192.168.1.21 for 86400 seconds
dhcpcd[253]: enp0s20u2: adding host route to 192.168.1.21 via 127.0.0.1
dhcpcd[253]: enp0s20u2: adding route to 192.168.1.0/24
dhcpcd[253]: enp0s20u2: adding default route via 192.168.1.5
dhcpcd[253]: enp0s20u2: forked to background, child pid 376

all of which looks fine to me, but hey, I'm no expert. <removed IP> was what looked like my external IP so decided not to share. Also when I ping I use www.google.com

Running ifconfig gives me this:

$ sudo ifconfig
enp0s20u2: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
        inet 192.168.1.21  netmask 255.255.255.0  broadcast 192.168.1.255
        inet6 fe80::5a6d:8fff:fed3:3ec4  prefixlen 64  scopeid 0x20<link>
        ether 58:6d:8f:d3:3e:c4  txqueuelen 1000  (Ethernet)
        RX packets 48115  bytes 4067374 (3.8 MiB)
        RX errors 0  dropped 48  overruns 0  frame 0
        TX packets 0  bytes 2653 (2.5 KiB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

lo: flags=73<UP,LOOPBACK,RUNNING>  mtu 65536
        inet 127.0.0.1  netmask 255.0.0.0
        inet6 ::1  prefixlen 128  scopeid 0x10<host>
        loop  txqueuelen 0  (Local Loopback)
        RX packets 0  bytes 0 (0.0 B)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 0  bytes 0 (0.0 B)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

Not sure what to look for here in relation to my problem, but I can see packets are being dropped.

Thanks again for your time and help.

hobarrera · 2013-09-27 23:04:13

You should try pinging 192.168.1.5. Also, make sure you have DNS resolution. Looks like you issue is not hardware related but elsewhere - DHCP would not work if it were a hardware issue.

albertalbert · 2013-09-28 00:34:52

hobarrera wrote:

You should try pinging 192.168.1.5. Also, make sure you have DNS resolution. Looks like you issue is not hardware related but elsewhere - DHCP would not work if it were a hardware issue.

So strange, couldn't get it to work but after running dhcpcd I unplugged the ethernet, then back in and now I'm typing this from in Arch! Now to get Wi-Fi back up and running. I had it working pretty consistently before, but for the life of me can't remember how I got it installed and working. Should have written it down!

Looking here: https://bbs.archlinux.org/viewtopic.php?id=145884&p=4 it seems like I need to rebuild the module since the Kernel update. Trying that now, running

[albert@archer ~]$ sudo dkms install broadcom-wl/6.30.223.30

Kernel preparation unnecessary for this kernel.  Skipping...

Building module:
cleaning build area....
make KERNELRELEASE=3.11.1-1-ARCH -C /usr/lib/modules/3.11.1-1-ARCH/build M=/var/lib/dkms/broadcom-wl/6.30.223.30/build.....
cleaning build area....

DKMS: build completed.

wl.ko:
Running module version sanity check.
 - Original module
   - No original module exists within this kernel
 - Installation
   - Installing to /usr/lib/modules/3.11.1-1-ARCH/kernel/drivers/net/wireless/

depmod....

DKMS: install completed.

Cool, gonna reboot and see what's up, I'll edit and update when I'm done.

Edit: It's alive! Ran my little internet script that does all the busy work for me and I connected with Wi-Fi without a problem after running the code above and restarting! Yay!

Now I need to get my trackpad to let me right-click...

Last edited by albertalbert (2013-09-28 00:42:41)

hobarrera · 2013-09-28 11:47:57

albertalbert wrote:

hobarrera wrote:
You should try pinging 192.168.1.5. Also, make sure you have DNS resolution. Looks like you issue is not hardware related but elsewhere - DHCP would not work if it were a hardware issue.
So strange, couldn't get it to work but after running dhcpcd I unplugged the ethernet, then back in and now I'm typing this from in Arch! Now to get Wi-Fi back up and running. I had it working pretty consistently before, but for the life of me can't remember how I got it installed and working. Should have written it down!
Looking here: https://bbs.archlinux.org/viewtopic.php?id=145884&p=4 it seems like I need to rebuild the module since the Kernel update. Trying that now, running
[albert@archer ~]$ sudo dkms install broadcom-wl/6.30.223.30

Kernel preparation unnecessary for this kernel.  Skipping...

Building module:
cleaning build area....
make KERNELRELEASE=3.11.1-1-ARCH -C /usr/lib/modules/3.11.1-1-ARCH/build M=/var/lib/dkms/broadcom-wl/6.30.223.30/build.....
cleaning build area....

DKMS: build completed.

wl.ko:
Running module version sanity check.
 - Original module
   - No original module exists within this kernel
 - Installation
   - Installing to /usr/lib/modules/3.11.1-1-ARCH/kernel/drivers/net/wireless/

depmod....

DKMS: install completed.
Cool, gonna reboot and see what's up, I'll edit and update when I'm done.
Edit: It's alive! Ran my little internet script that does all the busy work for me and I connected with Wi-Fi without a problem after running the code above and restarting! Yay!
Now I need to get my trackpad to let me right-click...

Check the entire thread, all these things have been discussed and explained

nnkken · 2013-09-29 04:30:52

I'm using Macbook Air 2013 and I have a strange problem:
Kernel 3.11.2 (both the official linux and the linux-ck kernel) maps my "`" key into "<" key.

It only affects the Macbook Air keyboard, i.e. the "`" key is normal when I use external keyboard.
If I press "Shift+`", it results as ">".
It behaves normal on kernel 3.11.1, so I downgraded my kernel.
It is not related to WMs or X as the problem remains when I'm typing username in the command line login screen after disabling kdm.

Does anyone have the same problem?

Edit:
The problem is solved by options hid_apple iso_layout=0.
Sorry for not searching enough before asking.
I'm just a little bit confused about why it works normally before 3.11.2 but not after.

Last edited by nnkken (2013-09-29 06:52:45)

GSnake · 2013-09-29 17:15:17

Hello guys, I'm interested in buying the latest MBA BUT... is it linux-friendly? This is VERY important since I'll trash OSX as soon as possible to jump to the penguin side. Please let me know!

patrikjakobsson · 2013-09-30 02:14:29

GSnake wrote:

Hello guys, I'm interested in buying the latest MBA BUT... is it linux-friendly? This is VERY important since I'll trash OSX as soon as possible to jump to the penguin side. Please let me know!

I'm still having issues with the SSD Marvell controller. It seems some MBA come with a Samsung controller and some come with a Marvell controller. The Samsung controller seems to be working with the quirks people have mentioned but I'm unable to get the Marvell one to behave properly after supend/resume. I have the 128GB version which seems to come with the Marvell controller.

On the bright side there are some sound card fixes that just arrived in 3.12-rc3 which detects the headphone hotplug properly.
See http://www.kernelhub.org/?msg=336772&p=2

But I wouldn't buy one just yet...

GSnake · 2013-09-30 06:09:19

Will everything be fixed? How's battery life? What about the previous macbook models?

albertalbert · 2013-09-30 17:33:49

I just updated again to 3.11.2-1-ARCH and my trackpad seems to be working a lot better. Got two finger scrolling and right-click at least now just have to tweak it for better performance.

My battery life in Arch has been pretty outstanding, and once I get the back-light issues sorted I can only imagine it'll get better.

patrikjakobsson · 2013-09-30 21:49:26

I managed to find the problem for the SATA issue with the Marvell controller. You must set /sys/class/scsi_host/host[0/1]/link_power_management_policy to min_power instead of max_performance before going into suspend. pm-utils gets in the way so it needs to be fixed by making sure sata_alpm does the right thing before suspend and after resume.

See bug report: https://bugzilla.kernel.org/show_bug.cgi?id=62351

Edit: pm-utils in archlinux might do the right thing, I've only tinkered with this on Ubuntu 13.10 beta2

Last edited by patrikjakobsson (2013-09-30 21:50:40)

tommytomatoe · 2013-10-01 02:30:27

nnkken wrote:

I'm using Macbook Air 2013 and I have a strange problem:
Kernel 3.11.2 (both the official linux and the linux-ck kernel) maps my "`" key into "<" key.
It only affects the Macbook Air keyboard, i.e. the "`" key is normal when I use external keyboard.
If I press "Shift+`", it results as ">".
It behaves normal on kernel 3.11.1, so I downgraded my kernel.
It is not related to WMs or X as the problem remains when I'm typing username in the command line login screen after disabling kdm.
Does anyone have the same problem?

Edit:
The problem is solved by options hid_apple iso_layout=0.
Sorry for not searching enough before asking.
I'm just a little bit confused about why it works normally before 3.11.2 but not after.

I had the same issue! Bugged me for the last day. I tried that from the Wiki and all is well. :-)

albertalbert wrote:

I just updated again to 3.11.2-1-ARCH and my trackpad seems to be working a lot better. Got two finger scrolling and right-click at least now just have to tweak it for better performance.
My battery life in Arch has been pretty outstanding, and once I get the back-light issues sorted I can only imagine it'll get better.

The trackpad has worked out of the box for me. Two finger click and scroll are great. I switched the two-finger and three-finger actions since I like the two-finger "right" click. Now we just need to get two finger back action in the browser.

Which back-light issues are you having? With KDE the kb backlight works with the keyboard bindings. So does screen brightness.

If you're in tty, you can use this little script. It accepts an argument from 0-255.

#!/bin/bash

BRIGHTNESS=$1

if [[ "$BRIGHTNESS" == "off" ]]; then
    BRIGHTNESS=0
fi

cd /sys/devices/platform/applesmc.768/leds/smc\:\:kbd_backlight
echo $BRIGHTNESS | tee brightness

patrikjakobsson wrote:

I managed to find the problem for the SATA issue with the Marvell controller. You must set /sys/class/scsi_host/host[0/1]/link_power_management_policy to min_power instead of max_performance before going into suspend. pm-utils gets in the way so it needs to be fixed by making sure sata_alpm does the right thing before suspend and after resume.
See bug report: https://bugzilla.kernel.org/show_bug.cgi?id=62351
Edit: pm-utils in archlinux might do the right thing, I've only tinkered with this on Ubuntu 13.10 beta2

Good find! Look forward to hearing more about your results. The periodic hangups can be annoying.

Kivutar · 2013-10-01 22:11:35

You can boot arch on a macbook without bootloader. You need an HFS+ partition of 128mo + the usual EFI partition of 200mo. There is a blog post about it but I don't find it anymore. Of course you have to compile your kernel for that, using builtin command line and initrd.

tommytomatoe · 2013-10-01 22:15:21

Kivutar wrote:

You can boot arch on a macbook without bootloader. You need an HFS+ partition of 128mo + the usual EFI partition of 200mo. There is a blog post about it but I don't find it anymore. Of course you have to compile your kernel for that, using builtin command line and initrd.

The Arch Linux kernel has EFISTUB enabled by default. No compiling necessary. :-)

hobarrera · 2013-10-02 13:26:38

Kivutar wrote:

You can boot arch on a macbook without bootloader. You need an HFS+ partition of 128mo + the usual EFI partition of 200mo. There is a blog post about it but I don't find it anymore. Of course you have to compile your kernel for that, using builtin command line and initrd.

No need for two EFI partitions, you can keep just one of the ones you mention above. (1).

GSnake · 2013-10-03 08:24:12

I bought the MBA at last! Still waiting to make Arch Linux stable though... (Linux 3.12 too..)

Arch Linux

#51 2013-09-23 00:11:16

Re: Macbook Air 2013

#52 2013-09-23 12:45:17

Re: Macbook Air 2013

#53 2013-09-24 06:25:35

Re: Macbook Air 2013

#54 2013-09-24 07:15:51

Re: Macbook Air 2013

#55 2013-09-24 08:09:44

Re: Macbook Air 2013

#56 2013-09-24 20:12:48

Re: Macbook Air 2013

#57 2013-09-24 22:15:34

Re: Macbook Air 2013

#58 2013-09-25 19:41:00

Re: Macbook Air 2013

#59 2013-09-27 00:28:31

Re: Macbook Air 2013

#60 2013-09-27 03:22:18

Re: Macbook Air 2013

#61 2013-09-27 18:49:38

Re: Macbook Air 2013

#62 2013-09-27 23:04:13

Re: Macbook Air 2013

#63 2013-09-28 00:34:52

Re: Macbook Air 2013

#64 2013-09-28 11:47:57

Re: Macbook Air 2013

#65 2013-09-29 04:30:52

Re: Macbook Air 2013

#66 2013-09-29 17:15:17

Re: Macbook Air 2013

#67 2013-09-30 02:14:29

Re: Macbook Air 2013

#68 2013-09-30 06:09:19

Re: Macbook Air 2013

#69 2013-09-30 17:33:49

Re: Macbook Air 2013

#70 2013-09-30 21:49:26

Re: Macbook Air 2013

#71 2013-10-01 02:30:27

Re: Macbook Air 2013

#72 2013-10-01 22:11:35

Re: Macbook Air 2013

#73 2013-10-01 22:15:21

Re: Macbook Air 2013

#74 2013-10-02 13:26:38

Re: Macbook Air 2013

#75 2013-10-03 08:24:12

Re: Macbook Air 2013

Board footer