Re: Panic testing ZFS raidz expansion feature.

[Maurizio Vairani]

Hi Alexander,
thank you very much, with your patch it works.
--
Maurizio

Il giorno ven 10 gen 2025 alle ore 16:08 Alexander Motin 
ha scritto:

> Hi Maurizio,
>
> I was able to reproduce it on a real hardware.  The critical points are
> that vdev should have ashift > 9 and it should be a debug build.  This
> patch fixes it for me: https://github.com/openzfs/zfs/pull/16942 .
>
> On 10.01.2025 03:33, Maurizio Vairani wrote:
> > I'm using vm-bhyve-1.5.0_1. The vm-byve package is configured correctly.
> >
> > # vm iso
> >
> https://download.freebsd.org/snapshots/amd64/amd64/ISO-IMAGES/15.0/FreeBSD-15.0-CURRENT-amd64-20250109-4c89d59e0cda-274641-bootonly.iso
> <
> https://download.freebsd.org/snapshots/amd64/amd64/ISO-IMAGES/15.0/FreeBSD-15.0-CURRENT-amd64-20250109-4c89d59e0cda-274641-bootonly.iso
> >
> > # vm create fbsd15-13
> > # cd /mnt/pool2tb/vm/fbsd15-13
> > In fbsd15-13.conf:
> > memory=2G
> > # vm install fbsd15-13
> > FreeBSD-15.0-CURRENT-amd64-20250109-4c89d59e0cda-274641-bootonly.iso
> > # vm console fbsd15-13
> > Install the OS and switch off the VM.
> > # truncate -s 10G disk1.img
> > # truncate -s 10G disk2.img
> > # truncate -s 10G disk3.img
> > # truncate -s 10G disk4.img
> > lines added in fbsd15-13.conf:
> > disk1_type="virtio-blk"
> > disk1_name="disk1.img"
> > disk2_type="virtio-blk"
> > disk2_name="disk2.img"
> > disk3_type="virtio-blk"
> > disk3_name="disk3.img"
> > disk4_type="virtio-blk"
> > disk4_name="disk4.img"
> > # vm start fbsd15-13
> > # vm console fbsd15-13
> > In the VM:
> > # zpool create tank raidz1 vtbd1 vtbd2 vtbd3
> > # zpool attach tank raidz1-0 vtbd4
> > panic: VERIFY(vd == vd->vdev_top) failed
> >
> > cpuid = 0
> > time = 1736497428
> > KDB: stack backtrace:
> > db_trace_self_wrapper() at db_trace_self_wrapper+0x2b/frame
> > 0xfe005469f820
> > vpanic() at vpanic+0x136/frame 0xfe005469f950
> > spl_panic() at spl_panic+0x3a/frame 0xfe005469f9b0
> > zio_vdev_io_start() at zio_vdev_io_start+0x63a/frame 0xfe005469fa00
> > zio_nowait() at zio_nowait+0x10c/frame 0xfe005469fa40
> > vdev_check_boot_reserve() at vdev_check_boot_reserve+0x7a/frame
> > 0xfe005469fa70
> > spa_vdev_attach() at spa_vdev_attach+0x712/frame 0xfe005469faf0
> > zfs_ioc_vdev_attach() at zfs_ioc_vdev_attach+0x75/frame
> 0xfe005469fb30
> > zfsdev_ioctl_common() at zfsdev_ioctl_common+0x49a/frame
> 0xfe005469fbd0
> > zfsdev_ioctl() at zfsdev_ioctl+0xfb/frame 0xfe005469fc00
> > devfs_ioctl() at devfs_ioctl+0xd1/frame 0xfe005469fc50
> > vn_ioctl() at vn_ioctl+0xb6/frame 0xfe005469fcc0
> > devfs_ioctl_f() at devfs_ioctl_f+0x1e/frame 0xfe005469fce0
> > kern_ioctl() at kern_ioctl+0x286/frame 0xfe005469fd40
> > sys_ioctl() at sys_ioctl+0x12f/frame 0xfe005469fe00
> > amd64_syscall() at amd64_syscall+0x158/frame 0xfe005469ff30
> > fast_syscall_common() at fast_syscall_common+0xf8/frame
> 0xfe005469ff30
> > --- syscall (54, FreeBSD ELF64, ioctl), rip = 0x2f2054f58cca, rsp =
> > 0x2f2048e41cf8, rbp = 0x2f2048e41d60 ---
> > KDB: enter: panic
> > [ thread pid 999 tid 100184 ]
> > Stopped at  kdb_enter+0x33: movq$0,0x10538d2(%rip)
> > db>
>
> --
> Alexander Motin
>

HEADS UP: Release build system changes happening in 2025

[Colin Percival]

Hi all,

The FreeBSD Foundation, sponsored by the Sovereign Tech Agency, is funding
work this year which will be affecting the FreeBSD build process, and release
building in particular.  The main goal is to allow the entire release process
to run without special privileges (aka root); subsidiary goals include
improving the build process documentation so that release building also
doesn't require any special knowledge.  The goal is to have this work
completed before 15.0.

This should not affect users of FreeBSD release images -- the images will look
the same regardless of how they are built -- and the effect on people running
'make buildworld buildkernel installkernel installworld' should be minimal.

Organizations building products based on FreeBSD which involve changes to the
release process will probably need to set aside some time to merge this work
into their tree; in particular, if you have local patches to release/*, you
can expect to see merge conflicts at some point this year when you run 'git
rebase'.

If you have any questions or concerns about this work, please don't hesitate
to contact me (I might be able to help, or I might pass you along to the
Foundation people who are managing this project).

--
Colin Percival
FreeBSD Release Engineering Lead & EC2 platform maintainer
Founder, Tarsnap | www.tarsnap.com | Online backups for the truly paranoid

Re: WIFI compiling error: Can anyone fix it on main branch?

[Adrian Chadd]

I did temporarily break -head on some platforms for like 30 minutes like a
week ago? Is everything ok now?

What was the error?

-adrian


On Tue, 7 Jan 2025 at 04:17, Dave Cottlehuber  wrote:

> On Tue, 7 Jan 2025, at 09:39, Chen, Alvin W wrote:
> > Thanks!
> >
> > Internal Use - Confidential
>
> Hi Alvin,
>
> Can you share your HEAD commit, & an error message?
>
> 749b3b2c0629 works fine on my machine, its current as of 2 hours ago.
>
> A+
> Dave
>
>