Linux 3.15 Can Resume From Suspend 7~12x Faster

[phoronix]

Phoronix: Linux 3.15 Can Resume From Suspend 7~12x Faster

Early on in the Linux 3.15 merge window there were improvements to significantly speed-up suspend and resume for systems, but now there's another late merge of a patch that has the capability of speeding up the resume time from suspend by 7~12x for at least some laptop/desktop systems...

http://www.phoronix.com/vr.php?view=MTY2MjE

[Nille]

Nice nice, i would be more happy if i don't have to unload and reload the half of all important kernel modules for my system.

[asdfblah]

This is really cool!
But, what happens if I run a command like this?

$ pm-suspend && dmesg > /tmp/dmesg

Or this:

$ pm-suspend && dd if=/dev/sda of=/dev/sdb

It shouldn't cause problems, right?

[ichk]

 $ pm-suspend && dd if=/dev/sda of=/dev/sdb

Why would someone want to do this?

[ichk]

$ pm-suspend && dd if=/dev/sda of=/dev/sdb

It shouldn't cause problems, right?

Why would someone want to do this?

[Caleb]

According to the Intel blogpost (https://01.org/suspendresume/blogs/t...mpler-approach) there are 2 patches, one for ATA and another for SCSI devices.

It seems that the one that was merged in is for SCSI devices (who has SCSI devices on his laptop?). What happened to the ATA patch?

[ryanplusplus]

This is really cool!
But, what happens if I run a command like this?

$ pm-suspend && dmesg > /tmp/dmesg

Or this:

$ pm-suspend && dd if=/dev/sda of=/dev/sdb

It shouldn't cause problems, right?

Based on the overview on 01.org it sounds like this will work -- all ATA commands are queued and executed once resume completes.

It's suspicious that all of the benchmarks are on systems with HDDs. I suspect that there is far less benefit for machines with SSDs.

[asdfblah]

Based on the overview on 01.org it sounds like this will work -- all ATA commands are queued and executed once resume completes.

Ah, I overlooked that part... thanks.

Why would someone want to do this?

The first command is used when debugging resume from suspend; the second one... just to give an example of something that could go wrong, although it's a bad example... a better one would be something like:

$ pm-suspend && dd if=/dev/sdb of=/home/user/sdb.backup

[Caleb]

The original Intel blog post had both ATA and SCSI patches, but the patch that was merged into 3.15 was only the SCSI one.
What happened to the ATA patch? And who has SCSI devices in their laptops anyway?

[Azrael5]

Sata is missed from the benefit?