Activity
From 2022-05-28 to 2022-06-26
2022-06-26
-
11:55 PM Bug #14759 (In Progress): sd open/close semaphore is unfortunate
-
-
-
- The network cache accelerator (NCA) is an idea whose time has come, and gone.
The idea is to offload serving web p... -
- The kssl framework was intended to facilitate in-kernel offloading of the SSL operations required for the NCA (networ...
-
- FYI: I have rebooted under KVM / qemu on CentOS 8 numerous times in the past week, no problem.
-
- Yes please!
I almost wonder whether some of the drivers (like smrt) should be using this flag that aren't. -
- If you could collect the stack backtrace from this panic, it would really help:
The way to do that (assuming you'v... -
-
- Note that DEVSLEEP is a deeper sleep than SLUMBER.
Today in illumos we do not even support SLUMBER. -
- New devices support the notion of a new DEVSLEEP state in AHCI. However, our current code is ill equipped to deal wi...
-
- Some of our internal commentary:
All
Comments
Work Log
History
Activity
Ascending order - Click to sort in ... -
- The relevant fix I made for this is here:...
-
- Our QA team discovered a panic condition when tearing down a VNIC, if the operation is aborted.
Essentially, they ... -
- A little more information.
ATA-8/ACS makes the NAA identifier *mandatory*.
However, if you look at SATA-1.0, it... -
- Here's the delta I implemented for RackTop. Note that this code assumes that the device is a "modern" SATA drive. P...
-
- The LUN device IDs are generally available on modern SATA devices (required by T13 specifications), but our SATL does...
-
- One more thing. The driver that I have out for RTI was also explicitly tested for timeout, abort, and reset handling...
-
- It is indeed using sd.
This driver has been tested in the following environments:
CentOS 8 based KVM using libvirt
...
2022-06-25
-
08:45 PM Bug #14753 (Closed): convert zpool-features.7 to use mdoc
- "git commit ea962d11118b10579c946c4ac15559148ddf3cf8":https://github.com/illumos/illumos-gate/commit/ea962d11118b1057...
-
08:58 AM Feature #14763 (In Progress): bhyve upstream sync 2022 August
- Sync upstream activity in FreeBSD's bhyve into illumos.
The following list shows the commits brought over as part ...
2022-06-24
-
- "git commit d203f83453f8fc67147d6661b4c202c221ed0411":https://github.com/illumos/illumos-gate/commit/d203f83453f8fc67...
-
04:50 AM Feature #14712: Replace find_elf with something faster
- Since the question might come up, doing a very simple comparison of run times between the two, a run of each version ...
-
- "git commit 29621f011bb9ea4f2d9ea887e971bbb3910ee931":https://github.com/illumos/illumos-gate/commit/29621f011bb9ea4f...
-
01:23 AM Bug #14635: bhyve should expose additional vcpu state
- The typical battery of test guests booted and ran fine on these bits (atop AMD and Intel gear).
-
- With the change in place, the existing bhyve test suite still passes (on AMD and Intel):...
-
- The added functionality in this change was primarily tested with out-of-gate software, namely Propolis with some draf...
2022-06-23
-
- "git commit d515dd7754a14758624ee9b1330197cdb6a47c49":https://github.com/illumos/illumos-gate/commit/d515dd7754a14758...
-
- When testing with #14635 (and dev propolis), if @vmm_allow_state_writes@ was written to non-zero prior to attempting ...
-
- On a system running bits built from the change, the bhyve test suite still passes (on both AMD and Intel):...
-
- This change is part of a series (including #14635) of work to enable save/restore and eventually live migration in bh...
-
03:33 PM Bug #14745: ZFS should handle unknown/invalid vdev devids gracefully
- Testing: I've verified that pools still import after this change, which they unsurprisingly do. Moving to a newer ker...
-
02:13 PM Bug #14762 (New): fio and FDINFO need to understand each other
- While stress-testing #14685 I discovered another race, thanks to a DEBUG kernel:...
-
12:57 PM Bug #14761: diskinfo(8) doesn't see all NVMe devices immediately after boot
- try to run 'devfsadm -C' before 'diskinfo' and i'll see all drives.
blkdev not implemented power manager and if driv... -
- I have a test system here with 24 NVMe devices, only a handful of which I'm actually using. Running diskinfo immediat...
-
06:25 AM Bug #14760 (In Progress): Reduce ZIO io_lock contention on sorted scrub
- Update from OpenZFS:...
-
-
-
- The sd driver serializes all open/close operations such that any attempts to open the drive must block while any othe...
2022-06-22
-
10:35 PM Bug #14758: spurious presence detect change notification when enabling PCIe hotplug interrupt for the first time
- Rearranging @pciehpc_enable_intr@ to enable all the notification sources first, then clear the pending notifications,...
-
- While chasing #14747, I noticed that the race was caused by a hotplug event occurring as soon as the hotplug interrup...
-
- This caught my recently, as I was referencing this manual page.
-
- I no longer have access to this hardware, and honestly I've lost interest in this bug. I'm unassigning myself.
-
-
- It turns out that you really don't want to multipath SES devices.
I'm rejecting this. -
-
- It turns out that this was actually misguided.
We shipped a product for about a year with this support, but it tur... -
- If BSR-14195 integrates, then it will close this issue as well. I obviously encountered this same issue a long long ...
-
- "git commit 3b5f2d22219c7c9f6926c804c8fa13b60d9e8a63":https://github.com/illumos/illumos-gate/commit/3b5f2d22219c7c9f...
-
- To test this, a nightly build was run with the @-t@ option to use the usr/src/tools code (i.e. the new find_elf). The...
2022-06-21
-
- Like bhyve, viona presets a Private kernel interface which is nonetheless used by out-of-gate consumers (risking brea...
-
- Currently, there is no interface to query the "current avail index" and "used index" values from viona's in-kernel ri...
-
- Changing the bug summary/subject to reflect what experiments have shown.
-
- Ease up future updates from OpenZFS.
-
- Get us as close to OpenZFS version as possible.
Testing done: inspect with man command.
2022-06-20
-
- Update from OpenZFS (commit fc5200aa9b345972fc1b99869b03a373090b84c7)...
-
- "git commit ae5a8bed14db6c16225cac733ea042c27e242d18":https://github.com/illumos/illumos-gate/commit/ae5a8bed14db6c16...
2022-06-18
-
04:36 AM Feature #14750: pfiles on a core file could print socket information
- The reason for the difference in information is that the core file does not currently capture this kind of informatio...
-
- "git commit a196c3ffe35a9a1589267b660ce6c2ff2ccab5e9":https://github.com/illumos/illumos-gate/commit/a196c3ffe35a9a15...
2022-06-17
-
05:17 PM Feature #14750 (New): pfiles on a core file could print socket information
- I've got a core file from a process with a bunch of open sockets. `pfiles` on the process reports a lot more informa...
-
- The changes in the associated gerrit move the biggest unused static functions inside @#ifdef@ blocks that relate to t...
-
- As noted in #14422, one way in which the artifacts produced by gcc10 differ from those produced by gcc7 is that gcc10...
2022-06-16
-
08:26 PM Bug #13920 (Closed): support for hiding the cursor on the console
-
- We have been running with this change (combined with 13921 and 13922) in RackTop's shipping product for around a year...
-
- "git commit c43f6c5551067e7546f6e78df8e37d63bc255192":https://github.com/illumos/illumos-gate/commit/c43f6c5551067e75...
-
- See 13922 for testing notes.
-
- "git commit c43f6c5551067e7546f6e78df8e37d63bc255192":https://github.com/illumos/illumos-gate/commit/c43f6c5551067e75...
-
- See 13920 for testing notes.
-
- "git commit c43f6c5551067e7546f6e78df8e37d63bc255192":https://github.com/illumos/illumos-gate/commit/c43f6c5551067e75...
-
04:02 PM Bug #14594: psrinfo(1M) may fail to emit socket type
- With this change, the same machine reports:...
-
- Port following changes to loader:...
-
- To test this I went through and confirmed that we enumerated all of our PCIe devices the same way as we did previousl...
-
- For the moment, what I've done is to go ahead and basically done (3) right now which doesn't quite solve the user exp...
2022-06-15
-
- While doing bring up on our platform, we discovered the following panic. The following is the analysis that was provi...
-
09:39 PM Bug #14746 (New): ascii(7) cross references old /usr/pub/ascii not /usr/share/lib/pub/ascii
- Exactly as the synopsis, ...
-
-
- Wasn't this pushed in commit baf9a8500e8914f180ede682c95a37266e1a4e37 ? If it was, please close this bug.
-
-
- I fixed this and added the recent fields that were in the PCIe gen 5 spec here. With this fixed here's what you now s...
-
- "git commit ec6a85f5d3beb28d36eda5ef325b7410d7f3cfa3":https://github.com/illumos/illumos-gate/commit/ec6a85f5d3beb28d...
-
- Additionally, we are putting this into SmartOS 20220616 given our immediate MNX Cloud machines are either tested with...
-
- With #14686 we'll introduce new devid types for NVMe devices. This works nice and well even for root pools on NVMe: Z...
-
- This gives me some incentive to see about getting the bhyve vioscsi support wired up too, if only for testing and the...
-
03:32 AM Bug #14743: new vioscsi driver
- It would be good to make sure this works with Google Cloud (GCE) too, which was my original motivation for my vioscsi...
-
- There are currently two public drivers for virtio-scsi floating about -- one from Nexenta and another from Joshua Clu...
-
- "git commit 5c22bad5dcbd030d95aca36e1c2fb0f77b505e17":https://github.com/illumos/illumos-gate/commit/5c22bad5dcbd030d...
-
- Even FreeBSD adopted this from Linux. Let met quote FreeBSD's netinet/tcp.h:...
-
- "git commit e760f15095bdc9fa107e7c20ed2a5e4fb5865c1d":https://github.com/illumos/illumos-gate/commit/e760f15095bdc9fa...
-
- We lack support for AMD-VI (to enable PCI-passthru via the IOMMU) in bhyve, since it was not ported over with the res...
2022-06-14
-
- Testing this on my Intel machine (which supports VT-d), the @check_iommu@ test passes:...
-
- I tested this on and AMD system where AMD-vi is unimplemented, so IOMMU-related requests should fail. Using the test...
-
09:17 PM Bug #14730: When only site DC reboots, chaos ensues with LDAP and SMB
- Gordon Ross wrote in #note-2:
> Can you please attach the last parts of these files:
> /var/svc/log/system-idmap*.log... -
- Can you please attach the last parts of these files:
/var/svc/log/system-idmap*.log
/var/svc/log/network-smb-server... -
05:35 AM Bug #14730: When only site DC reboots, chaos ensues with LDAP and SMB
- This might be related to https://www.illumos.org/issues/13326, with plain LDAP on my setup everything also hangs if t...
-
- Intel CPUs affected by, and booted with, this patch include:
Intel(R) Core(TM) i3-6100U CPU 2.30GHz (Skylake)
Inte... -
- Update the Intel microcode to the latest one. Includes new Alder Lake platforms and:...
2022-06-13
-
- Recently, while developing a change to bhyve, we encountered a situation where a misbehaving cyclic (one that was rep...
2022-06-12
-
- Before fix:...
-
- There's an underflow in the calculation of new credits here:
usr/src/uts/common/fs/smbsrv/smb2_dispatch.c around 39... -
- Running smbtorture tests on a small VM with reduced max_credits (16) I saw many failures that were not in my "baselin...
-
10:53 AM Bug #14693: JMicron JMB585 AHCI SATA controller: System panics when trying to write data on pool
This time, and without doing any changes, the computer booted up with controller being detected as AHCI version fff...
2022-06-11
-
- Update from OpenZFS:
Since we use two B-trees q_exts_by_size and q_exts_by_addr, we should count 2x sizeof (range_...
2022-06-10
-
11:05 PM Bug #14735 (New): Users with a large group sidlist cause performance problems in SMB
- This issue was reported to us when a user copying 17MB of data across 733 files took over a minute just to 'find' the...
-
- "git commit 8130f8e19e97c2d371c7b8894112a375409fe34a":https://github.com/illumos/illumos-gate/commit/8130f8e19e97c2d3...
-
12:27 PM Feature #14547: it would be nice if 'prtconf -dp' worked
- After my changes:
$ prtconf -dp...
2022-06-09
-
- "git commit d4f59ae5164748e530cad42bd54b9b909d0c1ee8":https://github.com/illumos/illumos-gate/commit/d4f59ae5164748e5...
-
- The portions of this change related to vtd allocations were tested using the pending wad from #14732
-
- I tested this on both AMD and Intel hardware.
The test suite still passes:... -
- "git commit 61b20185b3a9f12c5f69672abe47b79dfb002cab":https://github.com/illumos/illumos-gate/commit/61b20185b3a9f12c...
-
- My question:
> While this fix does seem like an improvement to the code, do you have
> any idea about the interac... -
- "git commit 2514b110a74b7a0ce021feff362fe3c4c2352b43":https://github.com/illumos/illumos-gate/commit/2514b110a74b7a0c...
-
- The expansion ROM BAR is defined such that bits [31:11] contain bits [31:11] of the ROM address, so that a register c...
2022-06-08
-
- with grab_faults, we can try to print out the trace of function calls.
Without symbol table, we can not translate ad... -
- I ran the test suite on both AMD and Intel hardware, in addition to the normal battery of guest smoke tests. All of ...
2022-06-07
-
- With the proposed change in place, the bhyve tests still all pass:...
-
- VT-d's use of ACPI necessitates it as a module dependency for bhyve. Considering the IOMMU functionality is a fairly...
-
- The kernel potion of bhyve uses some shims for @malloc()@ and @free()@ from when it was ported from FreeBSD.
It woul...
2022-06-06
-
-
- There are actually a couple problems here.
1: Checking for max_connections should happen earlier, before we try to al... -
- Once a month at one of our sites a DC reboots for routine updates. During this time, the ldap/client service has iss...
-
- Change tested via full gate build and manual inspection of the rendered man page using @mandoc@
2022-06-05
-
08:26 AM Bug #14322: Remote OmniOS SMB user/group management via Windows computermanagement not working
- > ::smbuser -v
SMB user information (fffffe11a3599b48):
UID: 1
SSNID: 1ee6373549f
State: 1 (LOGGED_ON)
Flags: 0x...
2022-06-04
-
- This happens when the SMB server has created too many sessions, and runs into max_connections.
-
- Can you get the output of "::smbuser -v" for that user?
In particular, I'd like to see the user flags. -
- An interesting symptom of this bug, which I guess I should preserve for the record, is that if you have an object fil...
2022-06-03
-
06:57 PM Bug #14729 (In Progress): ipmi: port exponential backoff enhancement from FreeBSD
- https://github.com/freebsd/freebsd-src/commit/1524298754d00f6a98d1dc6df7945baeb8f24d17
Implementing the exponentia... -
- I spotted this when reading the man page....
-
- "git commit feea3b2dd6c0c3b59dfb1ab0f992b01906567d90":https://github.com/illumos/illumos-gate/commit/feea3b2dd6c0c3b5...
-
12:53 AM Bug #14721: getopt(1) erroneously says scheduled for removal.
- Fix integrated.
-
- "git commit 3281964baf91e4061898a5490ee91e126e81116c":https://github.com/illumos/illumos-gate/commit/3281964baf91e406...
-
- This adds support for an initial part of the AMD unified memory controller. This has the plumbing for support for Zen...
-
- "git commit e309284527479df5fbac1270f2abd4a739f1ab72":https://github.com/illumos/illumos-gate/commit/e309284527479df5...
-
- It turns out that this is happening because of 3 bugs.
The original, important, bug is that we merge sections whic...
2022-06-02
-
- Testing notes: `git pbchk` is clean, and I diff'ed the nroff'ed output from the new and old man pages and the only di...
-
- NVMe spec 1.2 added a namespace GUID (NGUID) to the namespace identification as an addition or replacement to the EUI...
-
- Ryan Goodfellow wrote in #note-3:
> h3. Testing Notes
>
> Here are the original repro steps with the patch applie... -
05:38 AM Bug #14724: failed delete of IP address made it undeletable
- h3. Testing Notes
Here are the original repro steps with the patch applied. The partially deleted behavior no long... -
- After reading through the delete code path for IPv6 addresses, here is what I think is going on.
@i_ipadm_delete_i...
2022-06-01
-
11:46 PM Bug #14722: ld should keep group members in separate output sections
- No rush!
-
- I'm pretty sure I've got to the bottom of this -- we're shooting ourselves in the foot. I just have to decide on whi...
-
- > Do you have the input objects that produced ruby-glommed.o?
I ran ... -
- Ignore comment #2, I got myself all twisted around.
So we're seeing this group:... -
- I _think_ this is a typo introduced in 2008 where by...
-
- I think this is because we're setting the comdat flag on a NULL input section descriptor. I'll give it a proper look.
-
- Hello! I ran into a linker crash while trying to build a development version of
Ruby. I'm on OmniOS and the linker l... -
- Thanks for the report. I'll take a swing at this.
-
- Ryan Goodfellow found the following steps to reproduce:...
-
- Apparently the ZFS tests need truncate(1) which is part of GNU coreutils, the package should contain a dependency to ...
-
- getopt(1) includes this paragraph:...
-
- I'll have to dig, but I may have reproduction programs too.
-
- There are full sensors that do not report thresholds, and fac_prov_ipmi should NOT consider this an error, and should...
-
- Encountered a situation where the ldapsearch command was crashing. Re-running with @UMEM_DEBUG=default@ provided a us...
2022-05-31
-
09:07 PM Bug #14041: Dell R650 Xeon Gold 5318Y not booting
- I have tested this patch via the following method:
This was tested via adding the changes to Tintri's omnios based i... -
- I'd recommend looking for stacks in portfs in this dump (`::stacks -m portfs`) as well.
-
- We've seen this once or twice in SmartOS in very racy conditions, where timer_delete() vs. close(evport_fd) conspire....
-
- An interesting crash while running some ordinary stuff user-level stuff.
This is a VMware guest, in case that might ...
2022-05-30
-
- From https://reviews.freebsd.org/D35328...
2022-05-29
-
- The bhyve accelerated virtio net (viona) driver does not currently support the virtio net control queue or being swit...
Also available in: Atom