Page MenuHomeFreeBSD

mav (Alexander Motin)
User

Projects

User Details

User Since
Jun 3 2014, 6:27 PM (358 w, 1 d)

Recent Activity

Yesterday

mav committed R10:6c4ea5bb59ac: Add IDs for ASMedia ASM116x PCIe 3.0 AHCI controllers. (authored by mav).
Add IDs for ASMedia ASM116x PCIe 3.0 AHCI controllers.
Wed, Apr 14, 1:04 AM
mav committed R10:3bec9180c9e7: Add IDs for ASMedia ASM116x PCIe 3.0 AHCI controllers. (authored by mav).
Add IDs for ASMedia ASM116x PCIe 3.0 AHCI controllers.
Wed, Apr 14, 1:03 AM
mav committed R10:bc6d368bb8a3: pms: handle maximum size IO with any alignment (authored by rlibby).
pms: handle maximum size IO with any alignment
Wed, Apr 14, 12:48 AM

Tue, Apr 13

mav committed R10:e49d3eb40324: Fix race in case of device destruction. (authored by mav).
Fix race in case of device destruction.
Tue, Apr 13, 3:26 PM

Sun, Apr 11

mav added inline comments to D29688: iscsi: Kick threads out of iscsi_ioctl() during unload..
Sun, Apr 11, 1:49 AM

Sat, Apr 10

mav accepted D29685: uma: Introduce per-domain reclamation functions.

Looks good to me.

Sat, Apr 10, 5:16 PM
mav accepted D29688: iscsi: Kick threads out of iscsi_ioctl() during unload..
Sat, Apr 10, 4:28 PM

Fri, Apr 9

mav added a comment to D29663: De-prioritize network driver ithreads to mitigate livelock.

My first thought about the high network priority was about serial ports without any hardware buffering and flow control, extremely sensitive to interrupt latency, but uart seems to use TTY priority, at least now. Not sure what else could be so time-critical in networking before, but sure there could be NICs with very few RX buffers. Typical disk hardware though should have much higher latency tolerance, since the amount of traffic is usually predetermined being initiated from the system, that is why I guess its priority was set lower.

Fri, Apr 9, 3:14 AM

Thu, Apr 8

mav accepted D29104: domainset: Define additional global policies.
Thu, Apr 8, 9:11 PM
mav added inline comments to D29104: domainset: Define additional global policies.
Thu, Apr 8, 7:16 PM
mav added inline comments to D29105: uma: Implement a non-strict first-touch policy for slab allocation.
Thu, Apr 8, 6:41 PM

Wed, Apr 7

mav committed R10:5a8d32b53b91: Add IDs for ASMedia ASM116x PCIe 3.0 AHCI controllers. (authored by mav).
Add IDs for ASMedia ASM116x PCIe 3.0 AHCI controllers.
Wed, Apr 7, 7:10 PM
mav committed R10:88fe518973b7: Move time math out of disabled interrupts sections. (authored by mav).
Move time math out of disabled interrupts sections.
Wed, Apr 7, 12:35 AM
mav committed R10:0d745e16f2d7: Move time math out of disabled interrupts sections. (authored by mav).
Move time math out of disabled interrupts sections.
Wed, Apr 7, 12:35 AM
mav committed R10:40074783d818: Do not read timer extra time when MWAIT is used. (authored by mav).
Do not read timer extra time when MWAIT is used.
Wed, Apr 7, 12:32 AM
mav committed R10:c3afa8d1884f: Change mwait_bm_avoidance use to match Linux. (authored by mav).
Change mwait_bm_avoidance use to match Linux.
Wed, Apr 7, 12:32 AM
mav committed R10:e6e4f79a6c4d: Do not read timer extra time when MWAIT is used. (authored by mav).
Do not read timer extra time when MWAIT is used.
Wed, Apr 7, 12:31 AM
mav committed R10:9c27e7141b74: Change mwait_bm_avoidance use to match Linux. (authored by mav).
Change mwait_bm_avoidance use to match Linux.
Wed, Apr 7, 12:31 AM

Tue, Apr 6

mav committed R10:ac503c194cd8: Introduce "soft" serseq variant. (authored by mav).
Introduce "soft" serseq variant.
Tue, Apr 6, 9:29 PM

Mon, Apr 5

mav added a comment to D29587: rc: replace NETWORK with NETWORKING.

I see in NETWORKING: "# PROVIDE: NETWORKING NETWORK", so it does not look broken to me, but it looks weird.

Mon, Apr 5, 4:06 PM
mav committed R10:5a898b2b78ce: Set PCIe device's Max_Payload_Size to match PCIe root's. (authored by mav).
Set PCIe device's Max_Payload_Size to match PCIe root's.
Mon, Apr 5, 2:41 PM
mav accepted D29578: rc: make ctld depend on NETWORKING.
Mon, Apr 5, 4:12 AM

Wed, Mar 24

mav committed R10:d00a723a2a5c: nvme: Replace potentially long DELAY() with pause(). (authored by mav).
nvme: Replace potentially long DELAY() with pause().
Wed, Mar 24, 1:27 AM
mav committed R10:ed407c92e43f: nvme: Replace potentially long DELAY() with pause(). (authored by mav).
nvme: Replace potentially long DELAY() with pause().
Wed, Mar 24, 1:26 AM

Mon, Mar 22

mav committed R10:0cb06bcc54d9: Include ccompile.h after opt_global.h. (authored by mav).
Include ccompile.h after opt_global.h.
Mon, Mar 22, 12:54 AM
mav added inline comments to D28583: nvme: Use shared timeout rather than timeout per transaction.
Mon, Mar 22, 12:50 AM

Wed, Mar 17

mav committed R10:4fbbe523653b: nvme: Replace potentially long DELAY() with pause(). (authored by mav).
nvme: Replace potentially long DELAY() with pause().
Wed, Mar 17, 2:37 PM
mav closed D29309: nvme: Replace potentially long DELAY()'s with pause()'s.
Wed, Mar 17, 2:36 PM
mav added inline comments to D29309: nvme: Replace potentially long DELAY()'s with pause()'s.
Wed, Mar 17, 5:06 AM
mav requested review of D29309: nvme: Replace potentially long DELAY()'s with pause()'s.
Wed, Mar 17, 2:45 AM

Mar 15 2021

mav committed R10:8ea35c45d520: Include ccompile.h after opt_global.h. (authored by mav).
Include ccompile.h after opt_global.h.
Mar 15 2021, 5:53 PM
mav committed R10:2cd7a99cebaf: Restore condition removed in df3747c6607b. (authored by mav).
Restore condition removed in df3747c6607b.
Mar 15 2021, 3:03 AM
mav committed R10:3ef86cd7c3cc: Make software iSCSI more configurable. (authored by mav).
Make software iSCSI more configurable.
Mar 15 2021, 3:03 AM
mav committed R10:6cd4542786f9: Move ic_check_send_space clear to the actual check. (authored by mav).
Move ic_check_send_space clear to the actual check.
Mar 15 2021, 3:03 AM
mav committed R10:cec95c506540: Replace STAILQ_SWAP() with simpler STAILQ_CONCAT(). (authored by mav).
Replace STAILQ_SWAP() with simpler STAILQ_CONCAT().
Mar 15 2021, 3:03 AM
mav committed R10:7b4859b4a26a: Micro-optimize OOA queue processing. (authored by mav).
Micro-optimize OOA queue processing.
Mar 15 2021, 3:03 AM
mav committed R10:3034c0dab755: Optimize TX coalescing by keeping pointer to last mbuf. (authored by mav).
Optimize TX coalescing by keeping pointer to last mbuf.
Mar 15 2021, 3:03 AM
mav committed R10:cb89ac5a1daa: Fix initiator panic after 6895f89fe54e. (authored by mav).
Fix initiator panic after 6895f89fe54e.
Mar 15 2021, 3:03 AM
mav committed R10:748feb192e2a: Optimize out few extra memory accesses. (authored by mav).
Optimize out few extra memory accesses.
Mar 15 2021, 3:03 AM
mav committed R10:6469aab051de: Coalesce socket reads in software iSCSI. (authored by mav).
Coalesce socket reads in software iSCSI.
Mar 15 2021, 3:03 AM
mav committed R10:cfd358d99732: Save context switch per I/O for iSCSI and IOCTL frontends. (authored by mav).
Save context switch per I/O for iSCSI and IOCTL frontends.
Mar 15 2021, 3:02 AM
mav committed R10:c4a81e647599: Refactor CTL datamove KPI. (authored by mav).
Refactor CTL datamove KPI.
Mar 15 2021, 3:02 AM
mav committed R10:4d5d50edc5ff: Fix build after 2c7dc6bae9fd. (authored by mav).
Fix build after 2c7dc6bae9fd.
Mar 15 2021, 3:02 AM
mav committed R10:2a99726fdc54: Microoptimize CTL I/O queues. (authored by mav).
Microoptimize CTL I/O queues.
Mar 15 2021, 3:02 AM
mav committed R10:15fe13c887d6: Move XPT_IMMEDIATE_NOTIFY handling out of periph lock. (authored by mav).
Move XPT_IMMEDIATE_NOTIFY handling out of periph lock.
Mar 15 2021, 3:02 AM
mav committed R10:aae8e02dfcb2: Move ic_check_send_space clear to the actual check. (authored by mav).
Move ic_check_send_space clear to the actual check.
Mar 15 2021, 2:41 AM
mav committed R10:8b0101da803d: Restore condition removed in df3747c6607b. (authored by mav).
Restore condition removed in df3747c6607b.
Mar 15 2021, 2:41 AM
mav committed R10:2f77e2810a9d: Replace STAILQ_SWAP() with simpler STAILQ_CONCAT(). (authored by mav).
Replace STAILQ_SWAP() with simpler STAILQ_CONCAT().
Mar 15 2021, 2:41 AM
mav committed R10:fb419e0ffb28: Fix initiator panic after 6895f89fe54e. (authored by mav).
Fix initiator panic after 6895f89fe54e.
Mar 15 2021, 2:41 AM
mav committed R10:6b409a8b32e0: Optimize TX coalescing by keeping pointer to last mbuf. (authored by mav).
Optimize TX coalescing by keeping pointer to last mbuf.
Mar 15 2021, 2:41 AM
mav committed R10:76d8d700f26c: Optimize out few extra memory accesses. (authored by mav).
Optimize out few extra memory accesses.
Mar 15 2021, 2:41 AM
mav committed R10:8db80adb47e3: Micro-optimize OOA queue processing. (authored by mav).
Micro-optimize OOA queue processing.
Mar 15 2021, 2:41 AM
mav committed R10:31d41a6a9b65: Coalesce socket reads in software iSCSI. (authored by mav).
Coalesce socket reads in software iSCSI.
Mar 15 2021, 2:40 AM
mav committed R10:26953f59a1a3: Fix build after 2c7dc6bae9fd. (authored by mav).
Fix build after 2c7dc6bae9fd.
Mar 15 2021, 2:40 AM
mav committed R10:a4bea2f2a65c: Refactor CTL datamove KPI. (authored by mav).
Refactor CTL datamove KPI.
Mar 15 2021, 2:40 AM
mav committed R10:d0b1f461e248: Microoptimize CTL I/O queues. (authored by mav).
Microoptimize CTL I/O queues.
Mar 15 2021, 2:40 AM
mav committed R10:6fb753b9fa19: Save context switch per I/O for iSCSI and IOCTL frontends. (authored by mav).
Save context switch per I/O for iSCSI and IOCTL frontends.
Mar 15 2021, 2:40 AM
mav committed R10:90ac5cb75e9f: Move XPT_IMMEDIATE_NOTIFY handling out of periph lock. (authored by mav).
Move XPT_IMMEDIATE_NOTIFY handling out of periph lock.
Mar 15 2021, 2:34 AM

Mar 11 2021

mav added a comment to D29222: cam: Move wakeup in xpt_done to before we drop the lock.
In D29222#654210, @imp wrote:

If we do the wakeup with the lock held, then that race isn't possible and the wakeup won't be missed.

As me and John just told, the wakeup will not be missed in any case, it can only be duplicated due to the race, since list inserting is done before that under the lock.

Mar 11 2021, 9:59 PM
mav added a comment to D29222: cam: Move wakeup in xpt_done to before we drop the lock.
In D29222#654204, @imp wrote:
In D29222#654202, @mav wrote:

I am not getting what is the point of bringing the wakeup() under the lock. Being out of the lock may cause extra wakeups, but they are not critical here.

Doing the wakeup outside of the lock races the waker. If you do it inside of the lock, it won't race.

Mar 11 2021, 6:45 PM
mav added a comment to D29222: cam: Move wakeup in xpt_done to before we drop the lock.

I am not getting what is the point of bringing the wakeup() under the lock. Being out of the lock may cause extra wakeups, but they are not critical here.

Mar 11 2021, 6:28 PM
mav accepted D29006: nvme: use config_intrhook_drain to avoid removable card races.
Mar 11 2021, 2:51 PM
mav accepted D29005: config_intrhook: provide config_intrhook_drain.
Mar 11 2021, 2:45 PM

Mar 10 2021

mav committed R10:2cee045b4d62: Move time math out of disabled interrupts sections. (authored by mav).
Move time math out of disabled interrupts sections.
Mar 10 2021, 6:53 PM

Mar 9 2021

mav committed R10:4058265d605d: The list of ports in configuration path shall be protected by locks, (authored by glebius).
The list of ports in configuration path shall be protected by locks,
Mar 9 2021, 10:40 PM
mav committed R10:84e5f9455a18: Do not exit ctl_be_block_worker() prematurely. (authored by mav).
Do not exit ctl_be_block_worker() prematurely.
Mar 9 2021, 3:16 AM
mav committed R10:c120ef9e141e: Do not exit ctl_be_block_worker() prematurely. (authored by mav).
Do not exit ctl_be_block_worker() prematurely.
Mar 9 2021, 2:48 AM
mav committed R10:a3d416f3daaf: Do not exit ctl_be_block_worker() prematurely. (authored by mav).
Do not exit ctl_be_block_worker() prematurely.
Mar 9 2021, 2:48 AM

Mar 8 2021

mav committed R10:075e4807df3e: Do not read timer extra time when MWAIT is used. (authored by mav).
Do not read timer extra time when MWAIT is used.
Mar 8 2021, 11:55 PM
mav committed R10:455219675dbd: Change mwait_bm_avoidance use to match Linux. (authored by mav).
Change mwait_bm_avoidance use to match Linux.
Mar 8 2021, 11:28 PM
mav added a comment to D29101: wait for device mounts in zpool and dumpon.

So the recommendation for zpool is to use the first version of the patch that call root_hold_wait before the zpool import, correct?

Mar 8 2021, 7:56 PM
mav accepted D29004: config_intrhook: Move from TAILQ to STAILQ and padding.

Still LGTM.

Mar 8 2021, 6:07 PM
mav added a comment to D29101: wait for device mounts in zpool and dumpon.

For root on ZFS kernel automatically calls vfs_mountroot_wait() before doing anything. I think it is the only safe way. Attempt to import pool before all disks are detected may end up in importing pool at older transaction group or using old, removed now from pool devices. So I would insert root_hold_wait call at the beginning of zpool_start().

Mar 8 2021, 6:03 PM

Mar 7 2021

mav committed R10:bff56d9cff25: Fix diroffdiroff, probably copy/paste bug. (authored by mav).
Fix diroffdiroff, probably copy/paste bug.
Mar 7 2021, 12:26 AM
mav committed R10:05083436a6e5: Fix diroffdiroff, probably copy/paste bug. (authored by mav).
Fix diroffdiroff, probably copy/paste bug.
Mar 7 2021, 12:26 AM

Mar 6 2021

mav committed R10:6ed39db2573b: Do not exit ctl_be_block_worker() prematurely. (authored by mav).
Do not exit ctl_be_block_worker() prematurely.
Mar 6 2021, 3:47 AM

Mar 5 2021

mav committed R10:3f926f5aed36: Remove pointless lun->be_lun checks. (authored by mav).
Remove pointless lun->be_lun checks.
Mar 5 2021, 12:54 AM
mav committed R10:5ea5c72f5226: Remove pointless lun->be_lun checks. (authored by mav).
Remove pointless lun->be_lun checks.
Mar 5 2021, 12:54 AM

Mar 4 2021

mav committed R10:e549b1bc5373: amdtemp(4): Add missing Family 17h models (authored by cem).
amdtemp(4): Add missing Family 17h models
Mar 4 2021, 3:15 PM
mav committed R10:c7d9dcb97ab7: amdsmn(4), amdtemp(4): add support for Family 19h (Zen 3) (authored by cem).
amdsmn(4), amdtemp(4): add support for Family 19h (Zen 3)
Mar 4 2021, 3:15 PM
mav committed R10:89f349bab6d6: amdtemp(4), amdsmn(4): Attach to Ryzen 4000 APU (Zen 2, "Renoir") (authored by cem).
amdtemp(4), amdsmn(4): Attach to Ryzen 4000 APU (Zen 2, "Renoir")
Mar 4 2021, 3:15 PM
mav committed R10:40c0c3ef1f7d: amdtemp(4): Remove dead code that snuck in with r357190 (authored by cem).
amdtemp(4): Remove dead code that snuck in with r357190
Mar 4 2021, 3:15 PM
mav committed R10:4b8f69d7805b: amdtemp(4): Add support for Family 17h CCD sensors (authored by cem).
amdtemp(4): Add support for Family 17h CCD sensors
Mar 4 2021, 3:15 PM
mav committed R10:ba47eae9819a: amdtemp(4): Refactor shared temperature calculation logic (authored by cem).
amdtemp(4): Refactor shared temperature calculation logic
Mar 4 2021, 3:15 PM
mav committed R10:3e1da9a9f85b: Make DataSN counter of solicited Data-Out local. (authored by mav).
Make DataSN counter of solicited Data-Out local.
Mar 4 2021, 2:41 PM
mav committed R10:a00bf7d9bba1: Make DataSN counter of solicited Data-Out local. (authored by mav).
Make DataSN counter of solicited Data-Out local.
Mar 4 2021, 2:40 PM
mav accepted D29004: config_intrhook: Move from TAILQ to STAILQ and padding.

Should probably be merged with D29005.

Mar 4 2021, 2:37 PM
mav accepted D29006: nvme: use config_intrhook_drain to avoid removable card races.
Mar 4 2021, 2:34 PM
mav accepted D29005: config_intrhook: provide config_intrhook_drain.
Mar 4 2021, 2:34 PM

Mar 3 2021

mav added a comment to D28583: nvme: Use shared timeout rather than timeout per transaction.

I like the general direction. I also was thinking about something like this to avoid callout operations on every I/O.

Mar 3 2021, 10:54 PM
mav added inline comments to D28674: cam: use small CCBs for SCSI and ATA IO.
Mar 3 2021, 10:08 PM
mav accepted D28672: cam: drop 'struct ccb_eng_exec' and 'struct ccb_eng_inq'..

I have no specific objections, but I suppose those were added for something.

Mar 3 2021, 9:55 PM
mav added inline comments to D29018: camcontrol: depop command.
Mar 3 2021, 9:51 PM
mav added inline comments to D29017: libcam: Define depop structures and introduce scsi_wrap.
Mar 3 2021, 9:42 PM
mav accepted D29006: nvme: use config_intrhook_drain to avoid removable card races.

Other than the commend above I think it may work.

Mar 3 2021, 9:23 PM
mav added inline comments to D29006: nvme: use config_intrhook_drain to avoid removable card races.
Mar 3 2021, 9:21 PM
mav committed R10:afc3e54eeee6: Move ic_check_send_space clear to the actual check. (authored by mav).
Move ic_check_send_space clear to the actual check.
Mar 3 2021, 8:30 PM
mav committed R10:aff9b9ee894e: Restore condition removed in df3747c6607b. (authored by mav).
Restore condition removed in df3747c6607b.
Mar 3 2021, 5:07 PM

Mar 2 2021

mav committed R10:df3747c6607b: Replace STAILQ_SWAP() with simpler STAILQ_CONCAT(). (authored by mav).
Replace STAILQ_SWAP() with simpler STAILQ_CONCAT().
Mar 2 2021, 11:49 PM
mav committed R10:06e9c710998b: Fix initiator panic after 6895f89fe54e. (authored by mav).
Fix initiator panic after 6895f89fe54e.
Mar 2 2021, 9:12 PM