Re: WEAK_REFERENCE?

2013-11-18 Thread Konstantin Belousov
On Mon, Nov 18, 2013 at 11:00:55PM +0100, Andreas Tobler wrote: > I prepared two patches, see below. The amd64 one is reviewed by bde@ and > the i386 is compile tested by me (runtime is theoretically also done, > but I'm not sure since I do not have 32-bit apps on my amd64). Use cc -m32. > > The

Re: UMA cache back pressure

2013-11-18 Thread Jeff Roberson
On Mon, 18 Nov 2013, Adrian Chadd wrote: Remember that for Netflix, we have a mostly non-cachable workload (with some very specific exceptions!) and thus we churn through VM pages at a presitidigious rate. 20gbit sec, or ~ 2.4 gigabytes a second, or ~ 680,000 4 kilobyte pages a second. It's quit

Re: UMA cache back pressure

2013-11-18 Thread Jeff Roberson
On Mon, 18 Nov 2013, Alexander Motin wrote: On 18.11.2013 21:11, Jeff Roberson wrote: On Mon, 18 Nov 2013, Alexander Motin wrote: I've created patch, based on earlier work of avg@, to add back pressure to UMA allocation caches. The problem of physical memory or KVA exhaustion existed there fo

random_harvestq eats much on geode-lx alix2c3

2013-11-18 Thread Stanisław Halik
Hey, random_harvestq eats much, much CPU on alix2c3: CPU: Geode(TM) Integrated Processor by AMD PCS (498.06-MHz 586-class CPU) glxsb0: mem 0xefff4000-0xefff7fff irq 9 at device 1.2 on pci0 Could you please add a sysctl/loader knob for it, or a way to throttle collection? Here's top output:

Re: WEAK_REFERENCE?

2013-11-18 Thread Andreas Tobler
On 18.11.13 23:56, Adrian Chadd wrote: > [snip] > > wiki.freebsd.org/FreeBSD/mips has links to the MIPS emulator setups. > > There's no excuse to avoid testing on MIPS. :-) Np, wasn't aware of that.. :) And I do not shy the work. Thanks, Andreas

Is there already staging support within /ports/Mk/bsd.java.mk?

2013-11-18 Thread Matthias Petermann
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Hello, while updating my SweetHome3D port I'd like to enable staging support. The build process is using bsd.java.mk to build via ant. After removing NO_STAGE and adding ${STAGEDIR} to the respective directories, I get the following message at *make

Re: WEAK_REFERENCE?

2013-11-18 Thread Adrian Chadd
[snip] wiki.freebsd.org/FreeBSD/mips has links to the MIPS emulator setups. There's no excuse to avoid testing on MIPS. :-) -adrian ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe,

Re: WEAK_REFERENCE?

2013-11-18 Thread Andreas Tobler
On 14.11.13 07:00, Konstantin Belousov wrote: > On Wed, Nov 13, 2013 at 10:18:27PM +0100, Andreas Tobler wrote: >> On 11.11.13 08:47, Konstantin Belousov wrote: >>> On Sat, Nov 09, 2013 at 11:16:08PM +0100, Andreas Tobler wrote: Hi all, anyone interested in this patch to remove the W

Re: UMA cache back pressure

2013-11-18 Thread Adrian Chadd
Remember that for Netflix, we have a mostly non-cachable workload (with some very specific exceptions!) and thus we churn through VM pages at a presitidigious rate. 20gbit sec, or ~ 2.4 gigabytes a second, or ~ 680,000 4 kilobyte pages a second. It's quite frightening and it's only likely to increa

Re: UMA cache back pressure

2013-11-18 Thread Alexander Motin
On 18.11.2013 21:11, Jeff Roberson wrote: On Mon, 18 Nov 2013, Alexander Motin wrote: I've created patch, based on earlier work of avg@, to add back pressure to UMA allocation caches. The problem of physical memory or KVA exhaustion existed there for many years and it is quite critical now for i

Re: UMA cache back pressure

2013-11-18 Thread Jeff Roberson
On Mon, 18 Nov 2013, Alexander Motin wrote: Hi. I've created patch, based on earlier work of avg@, to add back pressure to UMA allocation caches. The problem of physical memory or KVA exhaustion existed there for many years and it is quite critical now for improving systems performance while

Re: UMA cache back pressure

2013-11-18 Thread Alexander Motin
On 18.11.2013 14:10, Adrian Chadd wrote: On 18 November 2013 01:20, Alexander Motin wrote: On 18.11.2013 10:41, Adrian Chadd wrote: So, do you get any benefits from just the first one, or first two? I don't see much reason to handle that in pieces. As I have described above, each part has ow

Re: UMA cache back pressure

2013-11-18 Thread Adrian Chadd
On 18 November 2013 01:20, Alexander Motin wrote: > On 18.11.2013 10:41, Adrian Chadd wrote: >> >> Your patch does three things: >> >> * adds a couple new buckets; > > > These new buckets make bucket size self-tuning more soft and precise. > Without them there are buckets for 1, 5, 13, 29, ... ite

Re: UMA cache back pressure

2013-11-18 Thread Alexander Motin
On 18.11.2013 11:45, Luigi Rizzo wrote: On Mon, Nov 18, 2013 at 10:20 AM, Alexander Motin mailto:m...@freebsd.org>> wrote: On 18.11.2013 10:41, Adrian Chadd wrote: Your patch does three things: * adds a couple new buckets; These new buckets make bucket size self-tu

Re: UMA cache back pressure

2013-11-18 Thread Luigi Rizzo
On Mon, Nov 18, 2013 at 10:20 AM, Alexander Motin wrote: > On 18.11.2013 10:41, Adrian Chadd wrote: > >> Your patch does three things: >> >> * adds a couple new buckets; >> > > These new buckets make bucket size self-tuning more soft and precise. > Without them there are buckets for 1, 5, 13, 29,

Re: UMA cache back pressure

2013-11-18 Thread Alexander Motin
On 18.11.2013 10:41, Adrian Chadd wrote: Your patch does three things: * adds a couple new buckets; These new buckets make bucket size self-tuning more soft and precise. Without them there are buckets for 1, 5, 13, 29, ... items. While at bigger sizes difference about 2x is fine, at smallest

Re: zpool requires re-import on reboot

2013-11-18 Thread Andriy Gapon
on 18/11/2013 10:51 Beeblebrox said the following: >> Is there anything "non-standard" about your configuration? >> All pools that are present in /boot/zfs/zpool.cache on a root filesystem of a >> root pool should be automatically imported. > > Yes I know that, hence the reason I posted. > > ll /

Re: zpool requires re-import on reboot

2013-11-18 Thread Allan Jude
On 2013-11-18 03:51, Beeblebrox wrote: >> Is there anything "non-standard" about your configuration? >> All pools that are present in /boot/zfs/zpool.cache on a root filesystem of a >> root pool should be automatically imported. > Yes I know that, hence the reason I posted. > > ll /boot/zfs shows

Re: zpool requires re-import on reboot

2013-11-18 Thread Beeblebrox
> Is there anything "non-standard" about your configuration? > All pools that are present in /boot/zfs/zpool.cache on a root filesystem of a > root pool should be automatically imported. Yes I know that, hence the reason I posted. ll /boot/zfs shows recently updated zpool.cache => -rw-r--r-- 1

Re: UMA cache back pressure

2013-11-18 Thread Adrian Chadd
Hi! Your patch does three things: * adds a couple new buckets; * reduces some lock contention * does the aggressive backpressure. So, do you get any benefits from just the first one, or first two? -adrian On 17 November 2013 15:09, Alexander Motin wrote: > Hi. > > I've created patch, base

Re: zpool requires re-import on reboot

2013-11-18 Thread Andriy Gapon
on 18/11/2013 09:47 Beeblebrox said the following: > I have root on zfs, which mounts fine on start. I have two other pools, which > do not get mounted and must be imported each time. Boot falls to single-user > mode because it cannot find the zfs-related mounts for the two pools in > question. The

Re: zpool requires re-import on reboot

2013-11-18 Thread Beeblebrox
>> Do have zfs_enable="YES" in rc.conf? Yes, and my ZFS root mounts without problem. Also in /boot/loader.conf: zfs_load="YES" opensolaris_load="YES" vfs.root.mountfrom="zfs:bsds" #- #_ZFS_PERFORMANCE #I have 4G of Ram vfs.zfs.prefetch_disable=0 #Ram 4GB => 512. Ram 8GB => value 1024 vfs.