System very slow after suspend-to-ram

[Damien Cassou]

Binary package hint: pm-utils

Hi,

after I suspend my laptop by closing the lid and reopening the lid, the system is very very slow. There is a delay of some seconds between I type a key and the login screen displays it. The rest of the system is also very slow which forces me to reboot completely. I attache the result of gnome-power-bugreport.sh when the system works perfectly. If you want to, I can try to execute this program after a slow resume but it will be difficult.

This bug might be related to: bug #99675 or bug #229447.

Description: Ubuntu 8.04
Release: 8.04

pm-utils:
  Installed: 0.99.2-3ubuntu9
  Candidate: 0.99.2-3ubuntu9
  Version table:
 *** 0.99.2-3ubuntu9 0
        500 http://ch.archive.ubuntu.com hardy/main Packages
        100 /var/lib/dpkg/status

[Shouri Chatterjee (shouri)]

I can quantify the slowdown of the system after a suspend-to-RAM. I am not sure if the same result applies for a suspend-to-disk.

Steps to reproduce:
1. Boot up computer
2. Run a sample program and time it.
3. Suspend to RAM
4. Wake up.
5. Run the same sample program and time it.

Time taken will be about 3x more.

Attached are a ps report, a lsmod report, and a report of a sample program (lame encoder in my case), both before and after the suspend-to-RAM.

[Shouri Chatterjee (shouri)]

I forgot to mention that I have observed the above in Ubuntu Intrepid Ibex 8.10.
The above occurs both under the 2.6.27-9-generic, as well as the 2.6.27-11-generic linux kernels.

[Shouri Chatterjee (shouri)]

I reinstalled Ubuntu, but this time I used the 64-bit version. After all my computer uses AMD 64x2 Turion processors.

The problem I mentioned above (3x slowdown after a resume from suspend-to-RAM) no longer occurs. The earlier operating system was 32-bit installed on the same computer.

[Ilja S. (ilja-spam)]

Hi, I have absolutely the same problem on my Dell laptop with Ubuntu 8.10, 64-bit (2.26.27-11)
I used the same test to check the speed - encoding media file with "lame". The slowdown is about 4 times.

The good news is that I think I found why! (at least on 2 Dell laptops).
The problem is in AC Adapter. When I use one with lower voltage then standard one.
Here is how I reproduce this:
1. Start the system with standard (for my laptop) adapter - 90W. The system is fast.
2. Close the lid to suspend.
3. Change the adapter to 65W (you can do it with Dell, that is ok, just charge time is much longer)
4. Open the lid to return from suspend.
5. Now the system ~4x slower.
6. UNPLUG the 65W adapter to run on the battery power - and it is FAST again!! Plug 65W adapter back - and it is slow. I did few times - and it is definitely about adapter.

Now the interesting thing about this that if I do full restart with 65W adapter - everything works fine!
This quite annoying cause at work place we have lower voltage adapters and we need to restart OS every morning.

Hope this will help to fix this bug. Goog Luck!

[Damien Cassou (cassou)]

This is not related to my problem because I only have one kind of AC Adapter which is the default and official one.

[Shouri Chatterjee (shouri)]

Ilja S. wrote:
> Hi, I have absolutely the same problem on my Dell laptop with Ubuntu 8.10, 64-bit (2.26.27-11)
> I used the same test to check the speed - encoding media file with "lame". The slowdown is about 4 times.
>
> The good news is that I think I found why! (at least on 2 Dell laptops).
> The problem is in AC Adapter. When I use one with lower voltage then standard one.
> Here is how I reproduce this:
> 1. Start the system with standard (for my laptop) adapter - 90W. The system is fast.
> 2. Close the lid to suspend.
> 3. Change the adapter to 65W (you can do it with Dell, that is ok, just charge time is much longer)
> 4. Open the lid to return from suspend.
> 5. Now the system ~4x slower.
> 6. UNPLUG the 65W adapter to run on the battery power - and it is FAST again!! Plug 65W adapter back - and it is slow. I did few times - and it is definitely about adapter.
>
> Now the interesting thing about this that if I do full restart with 65W adapter - everything works fine!
> This quite annoying cause at work place we have lower voltage adapters and we need to restart OS every morning.
>
> Hope this will help to fix this bug. Goog Luck!
>

Not related to my problem either. I have only one adapter.
My problem went away by itself after I migrated from ubuntu-32 to ubuntu-64.

-Shouri

[Wojtek Kazimierczak (w-kazimierczak)]

Is this problem related only to Dell notebooks? Mine is D630 (Intel Core2). I think the slowdown is due to disk operations.

[Damien Cassou (cassou)]

I have a macbook v4.1 and confirm the problem on Ubuntu 9.04. Please click on the "I'm affected link".

[mmxbass (mmxbass)]

There's something larger at play here. This goes beyond Ubuntu or laptops specifically. I can confirm this exact same problem in a custom-made DESKTOP machine running Debian Squeeze.

[mmxbass (mmxbass)]

As an additional comment, I should note that only X is slow. You can switch into one of the other terminal views (CTRL+ALT+F2 for example) and the responsiveness is fairly normal.

[Nicholas Kourtzis (kourtzis)]

In my system (a Dell XPS M1330 laptop with a C2D processor and nVidia M2400GS graphics) the slowdown continues even if I reboot the system without completely shutting it down. And it is even visible in POST and grub screens (which are drawn slowly), so it should not be related to X. A complete shutdown and restart brings the system back to normal.

[Nicholas Kourtzis (kourtzis)]

I forgot to mention that I am currently running 10.04. It is the first time I encounter this problem; never had it with earlier versions of Ubuntu.

[Damien Cassou (cassou)]

@Nicholas, the fact that you need to shutdown your computer and the fact that the bug only appears with recent versions of Ubuntu seem to show that the bug is different. You might wish to report a new bug.

[Rabi Alam (statokinetics)]

I had been getting the problem in previous versions of ubuntu on this Macbook 4.1. I was still getting them as of recently on Lucid. I followed a link from a launchpad bug report page (https://bugs.edge.launchpad.net/ubuntu/+source/linux/+bug/400413) to a kernel bug tracker report page (https://bugzilla.kernel.org/show_bug.cgi?id=14718) and saw the suggestion to try booting with the kernel option 'irqpolled' enabled. So, I tried adding an entry for 2.6.32-24-generic with irqpoll enabled in grub.cfg, and since booting with that option, I haven't been having problems with sluggishness after suspend. But I wonder, why does changing the hardware interrupt type to polling fix the issue?

[Alecz20 (alexguzu)]

I have the same problem.

What I can add is that the slowness comes from the HDD working on "something"

I managed to open gnome-system-monitor, and I saw that the system was using huge amounts of memory:

1.5 GB of RAM and 1.1 GB of swap.

The computer was slow as when you are copying files.

I have no idea why this happens only when resuming from suspend.

[Laryllan (laryllan)]

Using the kernel parameter "irqpoll" does not resolve the issue for me.

[Adam Łyskawa (a-lyskawa)]

I have the same problem with Ubuntu 10.10 x64 and desktop PC.
Waking up takes ca 10-15 minutes with huge disk activity in process.
The problem occurs randomly, once every few suspends.

[gdi2k (gdi2k)]

I have this issue too on a Thinkpad X200. I had Maverick 64-bit installed from Oct-2010 without issue, but I did a fresh install the other day and have been having the issue ever since. Both were 64-bit Maverick. Using current kernel (Linux x200 2.6.35-28-generic #49-Ubuntu SMP).

Contrary to others, no swap is being used and only an average 800 MB of RAM is being used (I have 4 GB total) when this happens.

Initially I thought it had to do with CPU frequency scaling, but forcing the cores to run at max speed doesn't change anything. Also, when the issue is not occurring and I force the CPU cores to run at min speed (800 MHz), it's still much faster than when in this post-suspend slow state.

top shows nothing out of the ordinary - no run-away processes or the like. Nothing noteworthy in syslog. But the cores are completely maxed out at the smallest effort - running gmail in Firefox for example.

I'm wondering if this is the same issue:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/371212

[tedaste (toivo-arach)]

I have the same problem with a Dell Studio XPS.

When it resumes the disk light is continuously on and you can hear it seeking. The system is paging in processes, I don't know what it's done to the processes memory when it suspends but when it's resuming you can see the process size increasing again.

I do see the following error when it suspends, I don't know if it's related though.

Apr 18 20:30:33 lapie kernel: [104018.415883] [fglrx] Preparing suspend fglrx in kernel.
Apr 18 20:30:33 lapie kernel: [104018.416633] ehci_hcd 0000:00:1a.7: PCI INT C disabled
Apr 18 20:30:33 lapie kernel: [104018.416644] uhci_hcd 0000:00:1a.2: PCI INT C disabled
Apr 18 20:30:33 lapie kernel: [104018.416655] uhci_hcd 0000:00:1a.1: PCI INT B disabled
Apr 18 20:30:33 lapie kernel: [104018.416665] uhci_hcd 0000:00:1a.0: PCI INT A disabled
Apr 18 20:30:33 lapie kernel: [104018.439367] kworker/u:6: page allocation failure. order:10, mode:0x4020
Apr 18 20:30:33 lapie kernel: [104018.439370] Pid: 14836, comm: kworker/u:6 Tainted: P 2.6.38-8-generic #41-Ubuntu
Apr 18 20:30:33 lapie kernel: [104018.439372] Call Trace:
Apr 18 20:30:33 lapie kernel: [104018.439379] [<ffffffff811147c4>] ? __alloc_pages_nodemask+0x604/0x840
Apr 18 20:30:33 lapie kernel: [104018.439424] [<ffffffffa02de5d2>] ? gal_map_virtual_space+0x62/0xc0 [fglrx]
Apr 18 20:30:33 lapie kernel: [104018.439427] [<ffffffff81149f45>] ? alloc_pages_current+0xa5/0x110
Apr 18 20:30:33 lapie kernel: [104018.439429] [<ffffffff811107ee>] ? __get_free_pages+0xe/0x50
Apr 18 20:30:33 lapie kernel: [104018.439432] [<ffffffff81154e7f>] ? kmalloc_order_trace+0x3f/0xb0
Apr 18 20:30:33 lapie kernel: [104018.439434] [<ffffffff81155dfa>] ? __kmalloc+0x13a/0x160
Apr 18 20:30:33 lapie kernel: [104018.439461] [<ffffffffa02ed73f>] ? hal_wait_CPDMA_idle+0xf/0x30 [fglrx]
Apr 18 20:30:33 lapie kernel: [104018.439479] [<ffffffffa02c1b8e>] ? KCL_MEM_SmallBufferAllocAtomic+0x1e/0x20 [fglrx]
Apr 18 20:30:33 lapie kernel: [104018.439501] [<ffffffffa02d4d63>] ? firegl_save_fb+0x233/0x360 [fglrx]
Apr 18 20:30:33 lapie kernel: [104018.439525] [<ffffffffa02d3ef5>] ? firegl_pm_save_framebuffer+0x1f5/0x280 [fglrx]
Apr 18 20:30:33 lapie kernel: [104018.439548] [<ffffffffa02d687d>] ? firegl_cail_powerdown+0x8d/0x210 [fglrx]
Apr 18 20:30:33 lapie kernel: [104018.439563] [<ffffffffa02bef07>] ? fglrx_pci_suspend+0x87/0x150 [fglrx]
Apr 18 20:30:33 lapie kernel: [104018.439567] [<ffffffff81300586>] ? pci_legacy_suspend+0x46/0xe0
Apr 18 20:30:33 lapie kernel: [104018.439569] [<ffffffff813013f5>] ? pci_pm_suspend+0xd5/0x120
Apr 18 20:30:33 lapie kernel: [104018.439573] [<ffffffff813c0ab9>] ? pm_op+0x1b9/0x220
Apr 18 20:30:33 lapie kernel: [104018.439576] [<ffffffff8108fa60>] ? async_run_entry_fn+0x0/0x180
Apr 18 20:30:33 lapie kernel: [104018.439578] [<ffffffff813c0ce3>] ? __device_suspend+0x133/0x1b0
Apr 18 20:30:33 lapie kernel: [104018.439580] [<ffffffff813c0f71>] ? async_suspend+0x21/0x60
Apr 18 20:30:33 lapie kernel: [104018.439582] [<ffffffff8108fae4>] ? async_run_entry_fn+0x84/0x180
Apr 18 20:30:33 lapie kernel: [104018.439585] [<ffffffff8108269d>] ? process_one_work+0x11d/0x420
Apr 18 20:30:33 lapie kernel: [104018...

[xtsbdu3reyrbrmroezob (xtsbdu3reyrbrmroezob)]

This bug seems to be affecting me too even with all the latest Ubuntu updates. Never saw it before in previous releases.

[Jason Smith (sddfdds)]

Im getting this in natty too, or something similar. Suspend/resume work fine and are fast, but after resuming, hard drive access speed is limited to ~4MB/s. USB hard drives plugged into the system are not affected though. It's not quite unusable, but extremely frustrating when doing anything with the hard drive.

[Albert Wilkes (o-null07)]

My impression is that I was able to get rid of the issue after disabling "Spin down hard disks when possible" and the issue seemed to only affect my Dell D630, not my D820 despite running both running 2.6.32-38-generic #83-Ubuntu SMP

[Phillip Susi (psusi)]

Damien, if this is still happening for you, can you check with top after resuming to see if something is hogging the cpu?