Page MenuHomeFreeBSD

mav (Alexander Motin)
User

Projects

User Details

User Since
Jun 3 2014, 6:27 PM (318 w, 4 d)

Recent Activity

Thu, Jul 9

mav accepted D25168: geom(4): Look for aliases when searching for providers by name.

I haven't looked much on aliases, but this looks good to me.

Thu, Jul 9, 4:57 PM
mav added a comment to D25168: geom(4): Look for aliases when searching for providers by name.

On an unrelated note -- @mav I noticed that you have added the prefer non-withered providers logic in rS280685 but it's not obvious to me why we want to return an withered provider in g_provider_by_name(). Do you still remember the scenario that this would be useful for the caller?

Thu, Jul 9, 4:49 PM

Wed, Jul 8

mav committed rS363016: MFC r346386, r347890, r347896, r349895:.
MFC r346386, r347890, r347896, r349895:
Wed, Jul 8, 5:59 PM

Thu, Jul 2

mav committed rS362882: Release Ryan Moeller (freqlabs) from mentorship..
Release Ryan Moeller (freqlabs) from mentorship.
Thu, Jul 2, 1:10 AM
mav committed rS362881: MFC r362630: Fix few panics on NVMe's timing out initialization requests..
MFC r362630: Fix few panics on NVMe's timing out initialization requests.
Thu, Jul 2, 1:02 AM

Tue, Jun 30

mav committed rS362811: Decode APEI tables (BERT, EINJ, ERST, HEST)..
Decode APEI tables (BERT, EINJ, ERST, HEST).
Tue, Jun 30, 9:40 PM

Fri, Jun 26

mav committed rS362661: Add mostly dummy hw.pci.enable_aspm tunable..
Add mostly dummy hw.pci.enable_aspm tunable.
Fri, Jun 26, 7:55 PM

Thu, Jun 25

mav committed rS362630: Fix few panics on NVMe's timing out initialization requests..
Fix few panics on NVMe's timing out initialization requests.
Thu, Jun 25, 8:29 PM

Wed, Jun 24

mav committed rS362579: MFC r362337: Make polled request timeout less invasive..
MFC r362337: Make polled request timeout less invasive.
Wed, Jun 24, 1:49 PM
mav committed rS362578: MFC r362282: Fix admin qpair leak if detached during initial reset..
MFC r362282: Fix admin qpair leak if detached during initial reset.
Wed, Jun 24, 1:48 PM

Sat, Jun 20

mav accepted D25366: openzfs*: Bump ports again.
Sat, Jun 20, 2:04 PM

Fri, Jun 19

mav accepted D25352: openzfs*: Bump to latest master.
Fri, Jun 19, 12:51 AM
mav committed rS362357: MFC r362100: Fix config_intrhook leak on initial reset failure..
MFC r362100: Fix config_intrhook leak on initial reset failure.
Fri, Jun 19, 12:45 AM
mav committed rS362356: MFC r361835: Add bunch of HDA controller and codec IDs..
MFC r361835: Add bunch of HDA controller and codec IDs.
Fri, Jun 19, 12:44 AM
mav committed rS362355: MFC r361816: Limit AHCI to only one MSI if more is not needed..
MFC r361816: Limit AHCI to only one MSI if more is not needed.
Fri, Jun 19, 12:44 AM

Thu, Jun 18

mav committed rS362340: MFC r360504 (by imp): Style(9) nit: put function name at start of line..
MFC r360504 (by imp): Style(9) nit: put function name at start of line.
Thu, Jun 18, 8:25 PM
mav committed rS362339: MFC r360503 (by imp): Move / reword a comment..
MFC r360503 (by imp): Move / reword a comment.
Thu, Jun 18, 8:24 PM
mav committed rS362337: Make polled request timeout less invasive..
Make polled request timeout less invasive.
Thu, Jun 18, 7:16 PM
mav committed rS362282: Fix admin qpair leak if detached during initial reset..
Fix admin qpair leak if detached during initial reset.
Thu, Jun 18, 9:25 AM
mav requested changes to D25316: Generic interface driver for dmaengine devices.

I would be happy if we had some good abstraction for standalone DMA engines. Right now we are using IOAT engine in our products, but I'd be happy if there would be decent option from AMD side also. While this code is a good start, it is nowhere near something we could already use.

Thu, Jun 18, 9:18 AM
mav added a reviewer for D25316: Generic interface driver for dmaengine devices: cem.
Thu, Jun 18, 9:18 AM

Sat, Jun 13

mav added a reviewer for D25260: ichsmb_pci: convert to pci_device_table / add PCI_PNP_INFO: imp.
Sat, Jun 13, 5:33 PM
mav accepted D25260: ichsmb_pci: convert to pci_device_table / add PCI_PNP_INFO.

Makes sense to me, but I am not very familiar with the PNP_INFO code.

Sat, Jun 13, 5:32 PM

Fri, Jun 12

mav committed rS362105: MFC r351516: Announce PCI Segment Groups supported to PCI host _OSC..
MFC r351516: Announce PCI Segment Groups supported to PCI host _OSC.
Fri, Jun 12, 3:38 PM
mav committed rS362100: Fix config_intrhook leak on initial reset failure..
Fix config_intrhook leak on initial reset failure.
Fri, Jun 12, 2:14 PM

Jun 9 2020

mav committed rS361953: MFC r361630: Remove session locking from cfiscsi_pdu_update_cmdsn()..
MFC r361630: Remove session locking from cfiscsi_pdu_update_cmdsn().
Jun 9 2020, 2:04 AM
mav committed rS361952: MFC r361625: Report STATUS_QUEUED/SENT in `ctladm dumpooa` output..
MFC r361625: Report STATUS_QUEUED/SENT in `ctladm dumpooa` output.
Jun 9 2020, 2:01 AM
mav committed rS361951: MFC r361616: Move EXPDATASN/R2TSN from PDU to CTL_PRIV_FRONTEND..
MFC r361616: Move EXPDATASN/R2TSN from PDU to CTL_PRIV_FRONTEND.
Jun 9 2020, 2:01 AM
mav committed rS361950: MFC r361609: Remove PDU_TOTAL_TRANSFER_LEN() macro..
MFC r361609: Remove PDU_TOTAL_TRANSFER_LEN() macro.
Jun 9 2020, 2:00 AM

Jun 8 2020

mav committed rS361939: Implement zero-copy iSCSI target transmission/read..
Implement zero-copy iSCSI target transmission/read.
Jun 8 2020, 8:54 PM

Jun 7 2020

mav accepted D25157: ahci_generic: add quirk for NXP0004 (NXP Layerscape LX2160A).

If its indeed known to be broken -- OK.

Jun 7 2020, 10:54 PM · arm64

Jun 5 2020

mav accepted D25155: CISS: Support >48 JBOD drives, fix SES enumeration, no panic on unplug, sysctl tunables &more verbosity....

I am not very close to ciss, only fixed couple issues there, but it looks OK to me, except the initiator_id comment and somewhat weird way to detect max target number.

Jun 5 2020, 7:40 PM
mav committed rS361835: Add bunch of HDA controller and codec IDs..
Add bunch of HDA controller and codec IDs.
Jun 5 2020, 3:07 PM
mav committed rS361816: Limit AHCI to only one MSI if more is not needed..
Limit AHCI to only one MSI if more is not needed.
Jun 5 2020, 2:21 AM

Jun 4 2020

mav committed rS361795: MFC r354637 (by royger): xen: fix dispatching of NMIs.
MFC r354637 (by royger): xen: fix dispatching of NMIs
Jun 4 2020, 5:08 PM
mav committed rS361794: MFC r354638: teach db_nextframe/x86 about [X]xen_intr_upcall interrupt handler.
MFC r354638: teach db_nextframe/x86 about [X]xen_intr_upcall interrupt handler
Jun 4 2020, 5:02 PM
mav added inline comments to D25128: hdac_intr_handler: keep working until global interrupt status clears.
Jun 4 2020, 1:54 PM
mav added a comment to D25128: hdac_intr_handler: keep working until global interrupt status clears.

Make sure you test both MSI and legacy interrupts, because they may work very different. I don't have objections for the patch if it helps some broken hardware, but comparing to AHCI, the last generates another interrupt when interrupt status register is cleared while there is some more work left, avoiding another expensive register read. Second read should not be a huge deal for sound, but still it is not free.

Jun 4 2020, 1:36 PM

Jun 3 2020

mav accepted D25122: openzfs*: Update to latest master.
Jun 3 2020, 6:45 PM
mav accepted D25120: Avoid trying to toggle TSO twice.

It makes sense to me on a brief look, but please request review from the driver maintainers.

Jun 3 2020, 4:32 PM

Jun 2 2020

mav committed rS361736: MFC 361400: Do not try to fill socket send buffer to the last byte..
MFC 361400: Do not try to fill socket send buffer to the last byte.
Jun 2 2020, 8:43 PM
mav committed rS361735: MFC 361400: Do not try to fill socket send buffer to the last byte..
MFC 361400: Do not try to fill socket send buffer to the last byte.
Jun 2 2020, 8:42 PM
mav committed rS361734: MFC r361604: Make struct ctl_be_lun first element of struct ctl_be_*_lun..
MFC r361604: Make struct ctl_be_lun first element of struct ctl_be_*_lun.
Jun 2 2020, 8:40 PM
mav committed rS361733: MFC r361590: Remove ctl_free_beio() LUN and ctl_io dependencies..
MFC r361590: Remove ctl_free_beio() LUN and ctl_io dependencies.
Jun 2 2020, 8:39 PM
mav committed rS361732: MFC r361536: Properly check kern_sg_entries for S/G list..
MFC r361536: Properly check kern_sg_entries for S/G list.
Jun 2 2020, 8:39 PM
mav committed rS361731: MFC r361536: Properly check kern_sg_entries for S/G list..
MFC r361536: Properly check kern_sg_entries for S/G list.
Jun 2 2020, 8:38 PM
mav committed rS361730: MFC r361502: Do not remove upcall if we haven't yet..
MFC r361502: Do not remove upcall if we haven't yet.
Jun 2 2020, 8:37 PM
mav committed rS361729: MFC r361502, r361509: Fix fallout of r319722 in CTL HA..
MFC r361502, r361509: Fix fallout of r319722 in CTL HA.
Jun 2 2020, 8:37 PM

Jun 1 2020

mav accepted D25077: tmpfs: Preserve alignment of struct fid fields.
Jun 1 2020, 1:49 AM

May 31 2020

mav added reviewers for D25077: tmpfs: Preserve alignment of struct fid fields: kib, mjg.
May 31 2020, 3:12 PM
mav requested changes to D25077: tmpfs: Preserve alignment of struct fid fields.

While its a good catch to find the overflow, your solution will not work on architectures with strict alignment. The problem is that struct tmpfs_fid_data require 64-bit alignment due to its fields, while fid_data offset within struct fid is 32 bits. So either you need some alignment-agnostic access to the fid_data, such as bcopy(), or split field(s) to 32-bit parts to fit into the struct fid alignment, or struct fid needs some padding to align fid_data to 64bits.

May 31 2020, 3:11 PM
mav accepted D25074: ctld: Fix leaks in conf_apply()..

It should be OK if desided, but I don't think there is a leak, since the whole old configuration is freed after application complete.

May 31 2020, 2:04 PM

May 29 2020

mav committed rS361630: Remove session locking from cfiscsi_pdu_update_cmdsn()..
Remove session locking from cfiscsi_pdu_update_cmdsn().
May 29 2020, 5:53 PM
mav committed rS361625: Report STATUS_QUEUED/SENT in `ctladm dumpooa` output..
Report STATUS_QUEUED/SENT in `ctladm dumpooa` output.
May 29 2020, 1:08 PM
mav committed rS361616: Move EXPDATASN/R2TSN from PDU to CTL_PRIV_FRONTEND..
Move EXPDATASN/R2TSN from PDU to CTL_PRIV_FRONTEND.
May 29 2020, 2:33 AM

May 28 2020

mav committed rS361609: Remove PDU_TOTAL_TRANSFER_LEN() macro..
Remove PDU_TOTAL_TRANSFER_LEN() macro.
May 28 2020, 11:56 PM
mav committed rS361604: Make struct ctl_be_lun first element of struct ctl_be_*_lun..
Make struct ctl_be_lun first element of struct ctl_be_*_lun.
May 28 2020, 9:30 PM
mav committed rS361590: Remove ctl_free_beio() LUN and ctl_io dependencies..
Remove ctl_free_beio() LUN and ctl_io dependencies.
May 28 2020, 6:12 PM

May 26 2020

mav committed rS361536: Properly check kern_sg_entries for S/G list..
Properly check kern_sg_entries for S/G list.
May 26 2020, 7:09 PM
mav added inline comments to D24877: pci: loosen PCIe hot-plug requirements.
May 26 2020, 4:57 PM
mav committed rS361509: Fix fallout of r319722 in CTL HA..
Fix fallout of r319722 in CTL HA.
May 26 2020, 3:08 PM
mav committed rS361502: Do not remove upcall if we haven't yet..
Do not remove upcall if we haven't yet.
May 26 2020, 1:57 PM

May 22 2020

mav committed rS361400: Do not try to fill socket send buffer to the last byte..
Do not try to fill socket send buffer to the last byte.
May 22 2020, 6:10 PM

May 19 2020

mav committed rS361256: MFC r360564: Cleanup LUN addition/removal..
MFC r360564: Cleanup LUN addition/removal.
May 19 2020, 2:42 PM
mav committed rS361255: MFC r360564: Cleanup LUN addition/removal..
MFC r360564: Cleanup LUN addition/removal.
May 19 2020, 2:32 PM
mav committed rS361252: MFC r359686: Fix panic on kern.cam.ctl.ha_role change after r333446..
MFC r359686: Fix panic on kern.cam.ctl.ha_role change after r333446.
May 19 2020, 1:52 PM

May 12 2020

mav accepted D24825: Make all the ata probe* and xpt* routines aprobe* and axpt* respectively..

Makes sense to me.

May 12 2020, 10:27 PM

May 10 2020

mav committed rS360864: MFC r360610: Add session locking in cfiscsi_ioctl_handoff()..
MFC r360610: Add session locking in cfiscsi_ioctl_handoff().
May 10 2020, 2:14 AM
mav committed rS360863: MFC r360610: Add session locking in cfiscsi_ioctl_handoff()..
MFC r360610: Add session locking in cfiscsi_ioctl_handoff().
May 10 2020, 2:14 AM

May 8 2020

mav accepted D24759: openzfs*: Update to latest master.
May 8 2020, 12:32 PM

May 7 2020

mav added a comment to D24428: smartpqi: Maintenance commit of Microchip smartpqi.

I've made only a very brief look, but here are some comments. Also it would be nice to reviewers if you would not mix formatting and real code changes.

May 7 2020, 2:23 PM

May 3 2020

mav committed rS360610: Add session locking in cfiscsi_ioctl_handoff()..
Add session locking in cfiscsi_ioctl_handoff().
May 3 2020, 4:15 PM
mav closed D24587: [iscsi] Fix initiator kernel panic if target ctl(4) port is disabled.
May 3 2020, 4:15 PM

May 2 2020

mav committed rS360564: Cleanup LUN addition/removal..
Cleanup LUN addition/removal.
May 2 2020, 4:55 PM

May 1 2020

mav added a comment to D24645: sched_ule: rate limit work stealing.

Don't we already have some of this via kern.sched.trysteal_limit ? I guess you are proposing adjusting trysteal_limit, and looking frequently at cores that share L2, and less frequently the farther out you go? I like that idea in general. But I'm not confident in my ability to implement or tune it.

May 1 2020, 2:12 PM
mav committed rS360534: MFC r323320 (by scottl): Fix intrhook release in MFI as well.
MFC r323320 (by scottl): Fix intrhook release in MFI as well
May 1 2020, 1:47 PM
mav added a comment to D24645: sched_ule: rate limit work stealing.

I was thinking about something like this many times. It would be nice to skip this often useless work. But each time I am stopped by worry that it may cause CPU under-utilization on a systems with load average about 1 per SMT thread. 10ms interval you mentioned sounds quite large to me, but honestly I can imagine some synchronous irregular load patterns where any chosen interval will cause a pessimization, dramatically increasing execution latency.

May 1 2020, 2:35 AM
mav committed rS360516: MFC r360269: Map family 0x5F (Denverton) to goldmont..
MFC r360269: Map family 0x5F (Denverton) to goldmont.
May 1 2020, 12:41 AM
mav committed rS360515: MFC r360266: Add family 0x5F (Denverton) to PMC_CPU_INTEL_ATOM_GOLDMONT..
MFC r360266: Add family 0x5F (Denverton) to PMC_CPU_INTEL_ATOM_GOLDMONT.
May 1 2020, 12:40 AM

Apr 30 2020

mav requested changes to D24587: [iscsi] Fix initiator kernel panic if target ctl(4) port is disabled.

I think it is correct, except the same should be added also 48 lines lower.

Apr 30 2020, 8:17 PM
mav committed rS360500: MFC r333839 (by mmacy): getnextevent: put variable only used by KTR under ifdef….
MFC r333839 (by mmacy): getnextevent: put variable only used by KTR under ifdef…
Apr 30 2020, 5:51 PM

Apr 27 2020

mav committed rS360376: MFC r360123: Allow namespace-id specification where it makes sense..
MFC r360123: Allow namespace-id specification where it makes sense.
Apr 27 2020, 2:03 PM
mav committed rS360375: MFC r360117: Open device with O_RDONLY when command is non-invasive..
MFC r360117: Open device with O_RDONLY when command is non-invasive.
Apr 27 2020, 2:02 PM

Apr 24 2020

mav committed rS360279: MFC r323977 (by cem): g_resize_provider_event: Do not invoke orphan method twice.
MFC r323977 (by cem): g_resize_provider_event: Do not invoke orphan method twice
Apr 24 2020, 7:43 PM
mav committed rS360275: MFC r351514: Don't consider PCIe hot-plug command timeout fatal..
MFC r351514: Don't consider PCIe hot-plug command timeout fatal.
Apr 24 2020, 4:41 PM
mav committed rS360274: MFC r351514: Don't consider PCIe hot-plug command timeout fatal..
MFC r351514: Don't consider PCIe hot-plug command timeout fatal.
Apr 24 2020, 4:39 PM
mav committed rS360270: MFC r359622 (by mmacy): Update x86 counters.
MFC r359622 (by mmacy): Update x86 counters
Apr 24 2020, 4:19 PM
mav committed rS360269: Map family 0x5F (Denverton) to goldmont..
Map family 0x5F (Denverton) to goldmont.
Apr 24 2020, 4:05 PM
mav committed rS360266: Add family 0x5F (Denverton) to PMC_CPU_INTEL_ATOM_GOLDMONT..
Add family 0x5F (Denverton) to PMC_CPU_INTEL_ATOM_GOLDMONT.
Apr 24 2020, 3:09 PM
mav committed rS360265: MFC r359667 (by mw): Add hwpmc support for Intel Atom Goldmont microarchitecture.
MFC r359667 (by mw): Add hwpmc support for Intel Atom Goldmont microarchitecture
Apr 24 2020, 3:02 PM
mav committed rS360262: MFC r354470 (by gallatin): hwpmc : fix AMD perf counter MSR access.
MFC r354470 (by gallatin): hwpmc : fix AMD perf counter MSR access
Apr 24 2020, 2:48 PM

Apr 23 2020

mav added inline comments to D24549: Use a pthread mutex to protect the MSI and MSI-X state..
Apr 23 2020, 6:49 PM

Apr 21 2020

mav committed rS360156: MFC r348996 (by jtl):.
MFC r348996 (by jtl):
Apr 21 2020, 4:53 PM

Apr 20 2020

mav committed rS360128: MFC r348629 (by cem): daemon(8): Don't block SIGTERM during restart delay.
MFC r348629 (by cem): daemon(8): Don't block SIGTERM during restart delay
Apr 20 2020, 4:31 PM
mav committed rS360123: Allow namespace-id specification where it makes sense..
Allow namespace-id specification where it makes sense.
Apr 20 2020, 2:54 PM
mav committed rS360117: Open device with O_RDONLY when command is non-invasive..
Open device with O_RDONLY when command is non-invasive.
Apr 20 2020, 1:47 PM

Apr 19 2020

mav added a comment to D24480: Avoid panicing for commands that poll-out..
In D24480#538896, @imp wrote:

I'm thinking, though, that these should all just be normal timeouts that are short so we have the same issues everywhere and at least try to cancel the transaction before returning.

Apr 19 2020, 3:55 PM
mav added a comment to D24480: Avoid panicing for commands that poll-out..

I worry this is at least incomplete implementation or even wrong approach. The problem I see in what happen if/when the command still complete after the polling gave up, or when the timeout handler expire. I think at very least it will be a stack corruption.

Apr 19 2020, 3:21 PM

Apr 18 2020

mav accepted D24479: hdac: replace sketchy printf for unowned lock with a lock assertion.

This seems to be a very old piece of code. :)

Apr 18 2020, 1:04 AM

Apr 17 2020

mav accepted D24456: Don't try to copyout() to a kernel buffer..
Apr 17 2020, 4:25 PM