I have a Lenovo B570 that works with Arch extremely well with the exception of one this. Ever since my last reinstall (swapped out the mSATA SSD for a full size SSD) I've had a very slow wake from suspend. Booting up the laptop takes 7 seconds or so (not counting the BIOS screen) but waking up the laptop from suspend takes up to 20 seconds sometimes.
If I let systemd take care of the suspend, I get a black screen (with the backlight on) for 20 seconds. If I let pm-utils take care of it, I see whatever I last had on the screen, frozen for 20 seconds. (Also, if I left pm-utils take care of it, I hear the sound being powered up at the end of the 20 seconds.)
Unfortunately, all the logs seem to start AFTER the 20 seconds which makes it impossible to figure out what's causing the delay. dmesg is the only exception but that stops logging during that pause, it seems.
Here are the outputs of:
systemd-analyze blame | http://pastebin.com/Wgd9fT4b
cat /var/log/pm-suspend.log | http://pastebin.com/CQVfZqQR
The relevant dmesg | http://pastebin.com/kbzSCBhP
The relevant journalctl -r | http://pastebin.com/Xycdrnvg
Any help in figuring this out would be great! (Or maybe it's all in my head... lol)
Last edited by ivanmmj (2014-05-31 16:59:23)
About a month ago I started seeing the same issue. I used to see resume from suspend in about one to two seconds, now it takes about 20 seconds (oddly consistently, like it's waiting for a timeout of some sort). I have watched this thread closely to see if anyone got back to you and was sad to see that there still have been no responses.
FIrstly, my logs look similar.
Secondly, hardware specs...
Laptop: Lenovo IdeaPad N586 7540
RAM: 8G ddr3 (plenty for suspend)
CPU: AMD A6-4400M APU
Graphics: Radeon HD 7520G
I figured that since this started for me about a month ago, it was something introduced into the kernel (especially with all the commits going in for the faster resume from suspend coming in 3.15), or a driver that needs to load (eg: ati), so I tested different kernels to confirm as well as uninstalled graphics drivers (no go on that unfortunatley). Thus far I have tested the most recent 3.14.1 and 3.14.2. I then tested the live usb isos from older dates, going back to 2013.04.01, which has kernel 3.8.4-1. Unfortunately, I still see the issue. Further, since that test was done using a live iso on a usb stick, I think it points to a potential hardware issue, not my locally installed instance, since it wasn't even mounted when I tested suspend on the live iso.
I'm going to do more testing here to see if it's hardware related (perhaps my old laptop will fix it - ugh) and hopefully narrow it down from there. I'll post back here as soon as I find out more.
Last edited by nullspoon (2014-05-10 17:26:23)
Thank you for replying. I'm sorry I didn't reply sooner.
I did not try a different kernel from within Arch but I did try a USB stick with Zorin OS 8 on it. It's an Ubuntu based distro. It also has the lag on it. I'm also starting to wonder if it's an issue with the hardware (or more specifically, the UEFI.) I've had issues with Lenovo's implementations of UEFI and I'm wondering if this is just yet another issue. If it isn't, it's odd that two people with two Lenovo's started having the same issue around the same time.
Thanks for responding. I'm sad to hear that your experience has been the same as mine.
A little to add. I tried Crunchbang with kernel 3.2, still slow resume. That also removed systemd from the equation amongst other things. Finally, I tried the same arch usb stick on my old intel centrino and resume from suspend was less than one second (super frownie face).
As of late I've been dabbling with catalyst settings in the hope of finding out that the graphics card has its own persistent memory (super long shot I know. I've never even heard of anything like that). I'll start dabbling with the uefi settings to see if I can get any change in the way this works. Seems a more likely scenario. One question for you, did you do the arch uefi install process or no?
My lenovo only gives me the option to boot in UEFI mode so I have that running on syslinux. On my last install, which didn't have the issue, I was using rEFInd.
As far as catalyst goes, I have an Intel IGP so not an issue here.
Ok. I fixed my issue and I feel like an idiot... I had an empty microSD card adapter in the SD tray. Removing it lets the laptop wake up in a split second. Putting it back in causes the slow wake.
Let me know if you also have the same issue. lol
I can confirm that.
This problem has been bugging me for a while, I'd never thought that an empty MicroSD adapter is the problem
Thanks for posting the solution.
(Lenovo IdeaPad G570, Linux 3.14.3-2)
Sorry for the long delay in response. I gave up on this issue for a while and just now returned to it.
That is simply amazing. I also had one in my microsd card slot. Popped it out and now resume from suspend occurs in less than one second. Hibernate is much faster now as well (hibernating and resuming). Who would have thought that was the issue. Good find!