Page MenuHomeFreeBSD

seigo.tanimura_gmail.com (Seigo Tanimura)
User

Projects

User does not belong to any projects.

User Details

User Since
Feb 24 2024, 10:53 AM (13 w, 2 d)

Recent Activity

Today

seigo.tanimura_gmail.com added a comment to D45379: kern/vm: Use the dedicated bio zone for the swap bio..
In D45379#1035253, @imp wrote:

The other question I have: why is the swap pager going so nuts? Normally, back pressure keeps the source of I/Os from overwhelming the lower layers of the system (which is why allocations are failing and you've moved to a preallocation). Why isn't that the case here?

Tue, May 28, 5:08 AM

Yesterday

seigo.tanimura_gmail.com added a comment to D45379: kern/vm: Use the dedicated bio zone for the swap bio..
In D45379#1035237, @imp wrote:
root@pkgfactory2:~ # swapinfo
Device          1K-blocks     Used    Avail Capacity
/dev/nda1p1      67108820  7964000 59144820    12%
/dev/nda2p1      67108820  7987416 59121404    12%
Total           134217640 15951416 118266224    12%
root@pkgfactory2:~ # gpart show /dev/nda1
=>       40  134217648  nda1  GPT  (64G)
         40          8        - free -  (4.0K)
         48  134217640     1  freebsd-swap  (64G)

root@pkgfactory2:~ # gpart show /dev/nda2
=>       40  134217648  nda2  GPT  (64G)
         40          8        - free -  (4.0K)
         48  134217640     1  freebsd-swap  (64G)

Both of these drives are mispartitioned. For best performnace they should have at least 1MB if not more alignment for the first partition. Do we really need to add code to handle mis-aligned partition performance problems that are pilot error?

Mon, May 27, 5:39 PM
seigo.tanimura_gmail.com added a comment to D45215: kern/geom: Make the struct bio allocation more robust to a heavy load..

Poudriere-bulk(8) Test Results

  • The OS version has been updated to 14.1-RC1.
  • kern.geom.reserved_new_bios and vm.swap_reserved_new_bios are set to 65536, the default values.
  • The tendency of the results are the same as https://reviews.freebsd.org/D45215#1033442.
Mon, May 27, 5:10 PM
seigo.tanimura_gmail.com added a comment to D45379: kern/vm: Use the dedicated bio zone for the swap bio..

Prior to any analysis, I assume you're doing this to fix some memory allocation deadlocks in the swap path. Could you describe a concrete scenario where you experienced a problem that this patch solves? Were you swapping to some regular partition, an encrypted one, or some vnode or zvol?

Mon, May 27, 1:41 PM
seigo.tanimura_gmail.com requested review of D45381: kern/dev/xen: Use the per-blkback-device uma(9) zone for the bio..
Mon, May 27, 10:09 AM
seigo.tanimura_gmail.com requested review of D45380: kern/dev/nvme: Use the per-namespace uma(9) zone for the bio..
Mon, May 27, 10:08 AM
seigo.tanimura_gmail.com requested review of D45379: kern/vm: Use the dedicated bio zone for the swap bio..
Mon, May 27, 10:08 AM
seigo.tanimura_gmail.com requested review of D45378: kern/vm: Publish the name strings of the uma(9) zone..
Mon, May 27, 10:05 AM
seigo.tanimura_gmail.com updated the diff for D45215: kern/geom: Make the struct bio allocation more robust to a heavy load..

Diff Updates

  • The bio allocation accepts the caller-supplied uma(9) zone for both non-blocking (g_new_bio_uz()) and blocking (g_alloc_bio_uz()) allocations.
  • The bio cloners (g_clone_bio() and g_duplicate_bio()) inherits the uma(9) zone of the original bio.
  • One prerequisite and three usage diffs submitted separately.
    • Prerequisite:
      • (TBD) Publish the name strings of the uma(9) zone. Intended for the logging of the uma(9) zone names upon an allocation failure.
    • Usages:
      • (TBD) The VM swap pager.
      • (TBD) The nvme(4) namespace driver.
      • (TBD) The xen(4) blkback driver.
  • Optimization.
  • The new KPIs and loader tunables are covered in the man pages.
Mon, May 27, 10:03 AM
seigo.tanimura_gmail.com retitled D45215: kern/geom: Make the struct bio allocation more robust to a heavy load. from kern/geom: Reserve some bios for the non-blocking allocation. to kern/geom: Make the struct bio allocation more robust to a heavy load..
Mon, May 27, 9:59 AM

Wed, May 22

seigo.tanimura_gmail.com added a comment to D45215: kern/geom: Make the struct bio allocation more robust to a heavy load..

Updated Test Results with Fix

Wed, May 22, 3:39 AM

Tue, May 21

seigo.tanimura_gmail.com updated the diff for D45215: kern/geom: Make the struct bio allocation more robust to a heavy load..

One problem looked over regarding M_USE_RESERVE:

Tue, May 21, 8:30 AM

Sun, May 19

seigo.tanimura_gmail.com added a comment to D45215: kern/geom: Make the struct bio allocation more robust to a heavy load..

Patch for Bio Alloation Stats
{F84091654}

Sun, May 19, 7:20 AM
seigo.tanimura_gmail.com updated the diff for D45215: kern/geom: Make the struct bio allocation more robust to a heavy load..

Simplify the change accordingly to the reviews up to https://reviews.freebsd.org/D45215#1031442.

Sun, May 19, 5:45 AM
seigo.tanimura_gmail.com added a comment to D45215: kern/geom: Make the struct bio allocation more robust to a heavy load..
In D45215#1031630, @imp wrote:

What does nvmecontrol identify ndaX say here? What's the optimal I/O boundary? And maybe we should just have a knob to disable trying to use it.

Sun, May 19, 5:30 AM

Thu, May 16

seigo.tanimura_gmail.com added a comment to D45215: kern/geom: Make the struct bio allocation more robust to a heavy load..

Key findings out of quick codewalk over g_new_bio() callers
sys/dev/nvme/nvme_ns.c, the nvme(4) namespace driver, calls g_new_bio() in the heaviest manner.

  • A lot of child bios are allocated in a burst.
  • The child bios are always allocated in the non-blocking way by g_new_bio(), regardless from the bio commands.
  • The child bios are allocated and freed within the nvme(4) namespace driver.
Thu, May 16, 6:43 PM
seigo.tanimura_gmail.com added inline comments to D45215: kern/geom: Make the struct bio allocation more robust to a heavy load..
Thu, May 16, 10:17 AM
seigo.tanimura_gmail.com added inline comments to D45215: kern/geom: Make the struct bio allocation more robust to a heavy load..
Thu, May 16, 8:48 AM
seigo.tanimura_gmail.com added a comment to D45215: kern/geom: Make the struct bio allocation more robust to a heavy load..

PR: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=279021

Thu, May 16, 7:58 AM
seigo.tanimura_gmail.com requested review of D45215: kern/geom: Make the struct bio allocation more robust to a heavy load..
Thu, May 16, 7:58 AM

Apr 9 2024

seigo.tanimura_gmail.com added a comment to D44170: kern/openzfs: Apply the essential part of FreeBSD-EN-23:18.openzfs. (stable/13).

This change is to be applied to stable/13 only (obviously). So maybe I'm not understanding your question?

Apr 9 2024, 11:22 AM
seigo.tanimura_gmail.com updated the diff for D44176: kern/openzfs: Regulate the ZFS ARC pruning process precisely. (releng/13.2).

Follow style(9).

Apr 9 2024, 10:10 AM
seigo.tanimura_gmail.com updated the diff for D44175: kern/openzfs: Regulate the ZFS ARC pruning process precisely. (stable/13).

Follow style(9).

Apr 9 2024, 9:56 AM
seigo.tanimura_gmail.com updated the diff for D44174: kern/openzfs: Regulate the ZFS ARC pruning process precisely. (stable/14).

Follow style(9).

Apr 9 2024, 9:52 AM
seigo.tanimura_gmail.com updated the diff for D44173: kern/openzfs: Regulate the ZFS ARC pruning process precisely..

Follow style(9).

Apr 9 2024, 9:22 AM
seigo.tanimura_gmail.com added a comment to D44170: kern/openzfs: Apply the essential part of FreeBSD-EN-23:18.openzfs. (stable/13).

Are you going to apply this change to stable/13 only, or "backport" it to main first?

Apr 9 2024, 5:02 AM

Mar 26 2024

seigo.tanimura_gmail.com abandoned D44218: kern/nullfs: Add the sysctl to control the vnode caching upon mounting nullfs. (stable/13).

Merged to stable/13 as d119f5a194.

Mar 26 2024, 7:34 AM

Mar 5 2024

seigo.tanimura_gmail.com updated the diff for D44218: kern/nullfs: Add the sysctl to control the vnode caching upon mounting nullfs. (stable/13).

Reflect https://reviews.freebsd.org/D44217#1008900.

Mar 5 2024, 1:42 PM
seigo.tanimura_gmail.com updated the diff for D44217: kern/nullfs: Add the sysctl to control the vnode caching upon mounting nullfs..

Reflect https://reviews.freebsd.org/D44217#1008900.

Mar 5 2024, 1:41 PM
seigo.tanimura_gmail.com added a reviewer for D44217: kern/nullfs: Add the sysctl to control the vnode caching upon mounting nullfs.: kib.
Mar 5 2024, 1:39 PM
seigo.tanimura_gmail.com abandoned D44178: kern/nullfs: Implement the recycling on the nullfs nodes. (stable/13).

Superseded by https://reviews.freebsd.org/D44218.

Mar 5 2024, 10:47 AM
seigo.tanimura_gmail.com abandoned D44177: kern/nullfs: Implement the recycling on the nullfs nodes..

Superseded by https://reviews.freebsd.org/D44217.

Mar 5 2024, 10:47 AM
seigo.tanimura_gmail.com updated the diff for D44218: kern/nullfs: Add the sysctl to control the vnode caching upon mounting nullfs. (stable/13).

Reflect https://reviews.freebsd.org/D44217#1008785 .

Mar 5 2024, 10:42 AM
seigo.tanimura_gmail.com updated the diff for D44217: kern/nullfs: Add the sysctl to control the vnode caching upon mounting nullfs..

Reflect https://reviews.freebsd.org/D44217#1008785 .

Mar 5 2024, 10:39 AM
seigo.tanimura_gmail.com added a comment to D44177: kern/nullfs: Implement the recycling on the nullfs nodes..
In D44177#1008409, @kib wrote:

nullfs should be served, and is served by normal vnlru thread. If you need something special, -o nocache is the solution.

Then is it OK to make the nullfs lower vnode caching configurable via sysctl in the system-wide way?

The applications that mount the nullfs filesystems dynamically can then enjoy the solution without modifying the applications.

Mar 5 2024, 6:44 AM
seigo.tanimura_gmail.com added a comment to D44217: kern/nullfs: Add the sysctl to control the vnode caching upon mounting nullfs..

This fix has been confirmed to achieve the same effect as https://reviews.freebsd.org/D44177.

Mar 5 2024, 6:43 AM
seigo.tanimura_gmail.com updated subscribers of D44217: kern/nullfs: Add the sysctl to control the vnode caching upon mounting nullfs..
Mar 5 2024, 6:37 AM
seigo.tanimura_gmail.com added a comment to D44218: kern/nullfs: Add the sysctl to control the vnode caching upon mounting nullfs. (stable/13).

Unlike https://reviews.freebsd.org/D44178, this fix does not depend on the ZFS fix and can be merged independently from it.

Mar 5 2024, 6:37 AM
seigo.tanimura_gmail.com added a comment to D44217: kern/nullfs: Add the sysctl to control the vnode caching upon mounting nullfs..

Unlike https://reviews.freebsd.org/D44177, this fix does not depend on the ZFS fix and can be merged independently from it.

Mar 5 2024, 6:37 AM
seigo.tanimura_gmail.com added a comment to D44218: kern/nullfs: Add the sysctl to control the vnode caching upon mounting nullfs. (stable/13).

Target branches: stable/13, releng/13.3, releng/13.2
Backport of: https://reviews.freebsd.org/D44217
Revised diff of: https://reviews.freebsd.org/D44178

Mar 5 2024, 6:34 AM
seigo.tanimura_gmail.com added a comment to D44217: kern/nullfs: Add the sysctl to control the vnode caching upon mounting nullfs..

Target branches: main, stable/14, releng/14.0
Revised diff of: https://reviews.freebsd.org/D44177

Mar 5 2024, 6:33 AM
seigo.tanimura_gmail.com requested review of D44218: kern/nullfs: Add the sysctl to control the vnode caching upon mounting nullfs. (stable/13).
Mar 5 2024, 6:32 AM
seigo.tanimura_gmail.com requested review of D44217: kern/nullfs: Add the sysctl to control the vnode caching upon mounting nullfs..
Mar 5 2024, 6:31 AM

Mar 4 2024

seigo.tanimura_gmail.com added a comment to D44177: kern/nullfs: Implement the recycling on the nullfs nodes..
In D44177#1008409, @kib wrote:

nullfs should be served, and is served by normal vnlru thread. If you need something special, -o nocache is the solution.

Mar 4 2024, 7:08 AM

Mar 3 2024

seigo.tanimura_gmail.com added a comment to D44177: kern/nullfs: Implement the recycling on the nullfs nodes..
In D44177#1008050, @kib wrote:

Putting fuse and nfs aside.

I think the best route is to add some zfs fs property (I do not know what is the proper way) to set MNTK_NULL_NOCACHE on that fs.

Mar 3 2024, 9:09 AM
seigo.tanimura_gmail.com retitled D44170: kern/openzfs: Apply the essential part of FreeBSD-EN-23:18.openzfs. (stable/13) from kern/openzfs: Apply the essential part of FreeBSD-EN-23:18.openzfs. to kern/openzfs: Apply the essential part of FreeBSD-EN-23:18.openzfs. (stable/13).
Mar 3 2024, 12:17 AM

Mar 2 2024

seigo.tanimura_gmail.com added a comment to D44177: kern/nullfs: Implement the recycling on the nullfs nodes..

Thanks for your input.

Mar 2 2024, 8:47 AM

Mar 1 2024

seigo.tanimura_gmail.com retitled D44178: kern/nullfs: Implement the recycling on the nullfs nodes. (stable/13) from kern/nullfs: Implement the recycling on the nullfs nodes. to kern/nullfs: Implement the recycling on the nullfs nodes. (stable/13).
Mar 1 2024, 6:04 PM
seigo.tanimura_gmail.com retitled D44176: kern/openzfs: Regulate the ZFS ARC pruning process precisely. (releng/13.2) from kern/openzfs: Regulate the ZFS ARC pruning process precisely. to kern/openzfs: Regulate the ZFS ARC pruning process precisely. (releng/13.2).
Mar 1 2024, 5:58 PM
seigo.tanimura_gmail.com retitled D44175: kern/openzfs: Regulate the ZFS ARC pruning process precisely. (stable/13) from kern/openzfs: Regulate the ZFS ARC pruning process precisely. to kern/openzfs: Regulate the ZFS ARC pruning process precisely. (stable/13).
Mar 1 2024, 5:58 PM
seigo.tanimura_gmail.com retitled D44172: kern/vfs: Add the per-filesystem vnode counter to struct mount. (stable/13) from kern/vfs: Add the per-filesystem vnode counter to struct mount. to kern/vfs: Add the per-filesystem vnode counter to struct mount. (stable/13).
Mar 1 2024, 5:56 PM
seigo.tanimura_gmail.com retitled D44174: kern/openzfs: Regulate the ZFS ARC pruning process precisely. (stable/14) from kern/openzfs: Regulate the ZFS ARC pruning process precisely. to kern/openzfs: Regulate the ZFS ARC pruning process precisely. (stable/14).
Mar 1 2024, 5:52 PM
seigo.tanimura_gmail.com added a comment to D44178: kern/nullfs: Implement the recycling on the nullfs nodes. (stable/13).

Target branches: stable/13, releng/13.3, releng/13.2
Backport of: https://reviews.freebsd.org/D44177

Mar 1 2024, 5:44 PM
seigo.tanimura_gmail.com requested review of D44178: kern/nullfs: Implement the recycling on the nullfs nodes. (stable/13).
Mar 1 2024, 5:43 PM
seigo.tanimura_gmail.com added a comment to D44177: kern/nullfs: Implement the recycling on the nullfs nodes..

Target branches: main, stable/14, releng/14.0

Mar 1 2024, 5:42 PM
seigo.tanimura_gmail.com requested review of D44177: kern/nullfs: Implement the recycling on the nullfs nodes..
Mar 1 2024, 5:41 PM
seigo.tanimura_gmail.com added a comment to D44176: kern/openzfs: Regulate the ZFS ARC pruning process precisely. (releng/13.2).

Target branch: releng/13.2
Backport of: https://reviews.freebsd.org/D44173

Mar 1 2024, 5:40 PM
seigo.tanimura_gmail.com requested review of D44176: kern/openzfs: Regulate the ZFS ARC pruning process precisely. (releng/13.2).
Mar 1 2024, 5:39 PM
seigo.tanimura_gmail.com added a comment to D44175: kern/openzfs: Regulate the ZFS ARC pruning process precisely. (stable/13).

Target branches: stable/13, releng/13.3
Backport of: https://reviews.freebsd.org/D44173

Mar 1 2024, 5:37 PM
seigo.tanimura_gmail.com requested review of D44175: kern/openzfs: Regulate the ZFS ARC pruning process precisely. (stable/13).
Mar 1 2024, 5:36 PM
seigo.tanimura_gmail.com added a comment to D44174: kern/openzfs: Regulate the ZFS ARC pruning process precisely. (stable/14).

Target branches: stable/14, releng/14.0
Backport of: https://reviews.freebsd.org/D44173

Mar 1 2024, 5:34 PM
seigo.tanimura_gmail.com requested review of D44174: kern/openzfs: Regulate the ZFS ARC pruning process precisely. (stable/14).
Mar 1 2024, 5:32 PM
seigo.tanimura_gmail.com added a comment to D44173: kern/openzfs: Regulate the ZFS ARC pruning process precisely..

Target branch: main

Mar 1 2024, 5:31 PM
seigo.tanimura_gmail.com requested review of D44173: kern/openzfs: Regulate the ZFS ARC pruning process precisely..
Mar 1 2024, 5:30 PM
seigo.tanimura_gmail.com added a comment to D44172: kern/vfs: Add the per-filesystem vnode counter to struct mount. (stable/13).

Target branches: stable/13, releng/13.3, releng/13.2
Backport of: https://reviews.freebsd.org/D44171

Mar 1 2024, 5:28 PM
seigo.tanimura_gmail.com requested review of D44172: kern/vfs: Add the per-filesystem vnode counter to struct mount. (stable/13).
Mar 1 2024, 5:26 PM
seigo.tanimura_gmail.com added a comment to D44171: kern/vfs: Add the per-filesystem vnode counter to struct mount..

Target branches: main, stable/14, releng/14.0

Mar 1 2024, 5:26 PM
seigo.tanimura_gmail.com requested review of D44171: kern/vfs: Add the per-filesystem vnode counter to struct mount..
Mar 1 2024, 5:24 PM
seigo.tanimura_gmail.com added a comment to D44170: kern/openzfs: Apply the essential part of FreeBSD-EN-23:18.openzfs. (stable/13).

Target branches: stable/13, releng/13.3, releng/13.2

Mar 1 2024, 5:22 PM
seigo.tanimura_gmail.com requested review of D44170: kern/openzfs: Apply the essential part of FreeBSD-EN-23:18.openzfs. (stable/13).
Mar 1 2024, 5:20 PM