Page MenuHomeFreeBSD

jhb (John Baldwin)
User

Projects (8)

User Details

User Since
Mar 11 2014, 8:46 PM (369 w, 4 d)

Recent Activity

Yesterday

jhb added a comment to D29688: iscsi: Kick threads out of iscsi_ioctl() during unload..

To reproduce, just kldunload iscsi while iscsid is running. The kldunload will hang in "devdrn" inside of destroy_dev().

Sat, Apr 10, 12:02 AM
jhb requested review of D29688: iscsi: Kick threads out of iscsi_ioctl() during unload..
Sat, Apr 10, 12:00 AM

Fri, Apr 9

jhb accepted D29687: x86: add x86_clear_dbregs() and ensure all ABI exec set_regs method clear debug registers file.
Fri, Apr 9, 11:52 PM
jhb committed R10:86e352c934e5: Fix a typo in a comment: frame -> framework. (authored by jhb).
Fix a typo in a comment: frame -> framework.
Fri, Apr 9, 11:12 PM
jhb committed R10:6a06b00a0d1f: nlmrsa: Remove this deprecated driver. (authored by jhb).
nlmrsa: Remove this deprecated driver.
Fri, Apr 9, 11:12 PM
jhb added a comment to D29663: De-prioritize network driver ithreads to mitigate livelock.

My first thought was much closer to Gleb's comment. That is, I think the mistake is probably to be lumping callouts in with other soft interrupts. I think probably what we want is something about like this:

Fri, Apr 9, 11:02 PM
jhb accepted D29676: Rename struct device to struct _device.
Fri, Apr 9, 10:48 PM
jhb accepted D29652: relnotes: note some more userland removals.
Fri, Apr 9, 10:45 PM
jhb accepted D29583: Drop 'Set to' from the beginning of most src.conf(5) knobs.
Fri, Apr 9, 10:43 PM
jhb accepted D29653: relnotes: note some more kernel/driver removals.
Fri, Apr 9, 10:43 PM
jhb updated the diff for D29297: cxgbe: Add counters for iSCSI PDUs transmitted via TOE..
  • Add octets stat and move to per queue.
Fri, Apr 9, 6:21 PM
jhb retitled D29297: cxgbe: Add counters for iSCSI PDUs transmitted via TOE. from Add a counter of iSCSI PDUs transmitted via TOE. to cxgbe: Add counters for iSCSI PDUs transmitted via TOE..
Fri, Apr 9, 6:12 PM

Thu, Apr 8

jhb requested review of D29662: cxgbe: Ignore doomed virtual interfaces when updating the clip table..
Thu, Apr 8, 11:40 PM
jhb added inline comments to D29652: relnotes: note some more userland removals.
Thu, Apr 8, 7:22 PM
jhb added inline comments to D29648: relnotes: fix-up some section oddities.
Thu, Apr 8, 7:21 PM
jhb added a comment to D29651: relnotes: add a section with some RISC-V changes.

Note that the notes only document changes from the most recent release (in this case 12.2) to 13.0, not 12.0 to 13.0, so some of these aren't relevant. I wonder if the OpenSBI support made it into 12.2? If not, it warrants a mention I think. Hmm, 12.2 did include support for booting via OpenSBI v0.7.

Thu, Apr 8, 7:18 PM
jhb accepted D29654: relnotes: add some kernel/driver additions.
Thu, Apr 8, 7:07 PM
jhb added inline comments to D29653: relnotes: note some more kernel/driver removals.
Thu, Apr 8, 7:06 PM
jhb added a comment to D29652: relnotes: note some more userland removals.

Previously Ed was consistent in suggesting "obsolete" to refer to old GPLv2 toolchain bits.

Thu, Apr 8, 7:05 PM
jhb added inline comments to D29607: byhve: add option to specify IP address for gdb.
Thu, Apr 8, 6:59 PM
jhb added inline comments to D29543: license: Create a license guideline document.
Thu, Apr 8, 6:52 PM
jhb added a comment to D29580: Enable -Wdate-time warning by default with Clang.

Looks like GCC 6 supports -Wdate-time so I think you can just add this to CWARNFLAGS.

Thu, Apr 8, 6:32 PM
jhb added inline comments to D29583: Drop 'Set to' from the beginning of most src.conf(5) knobs.
Thu, Apr 8, 6:30 PM
jhb accepted D29579: cxgbe/tom: Fix potential leak in t4_aiotx_process_job..
Thu, Apr 8, 6:26 PM
jhb added a comment to D29540: sendfile: Fix error initialization in sendfile_getobj().

BTW, confirmed this fixed the error Chelsio's QA was seeing internally.

Thu, Apr 8, 6:20 PM

Fri, Apr 2

jhb added inline comments to D29543: license: Create a license guideline document.
Fri, Apr 2, 9:16 PM

Thu, Apr 1

jhb committed R10:c86de1dab8e6: cryptocheck: Expand the set of sizes tested by -z. (authored by jhb).
cryptocheck: Expand the set of sizes tested by -z.
Thu, Apr 1, 10:50 PM
jhb committed R10:d2e076c37b09: ossl: Don't encryt/decrypt too much data for chacha20. (authored by jhb).
ossl: Don't encryt/decrypt too much data for chacha20.
Thu, Apr 1, 10:50 PM
jhb closed D29518: cryptocheck: Expand the set of sizes tested by -z..
Thu, Apr 1, 10:49 PM
jhb closed D29517: ossl: Don't encryt/decrypt too much data for chacha20..
Thu, Apr 1, 10:49 PM
jhb added inline comments to D29518: cryptocheck: Expand the set of sizes tested by -z..
Thu, Apr 1, 10:46 PM
jhb updated the diff for D29518: cryptocheck: Expand the set of sizes tested by -z..

Altered step in second block to 8.

Thu, Apr 1, 9:59 PM
jhb accepted D29529: x86: remove gcov kernel support//config(8): remove support for -p.

Thu, Apr 1, 9:23 PM
jhb accepted D29527: cxgbe(4): Always use the per-VI callout to read interface stats..
Thu, Apr 1, 9:21 PM
jhb added a comment to D29540: sendfile: Fix error initialization in sendfile_getobj().

I'm just waiting to see if the folks who reported this to me can confirm this fixes the issue they were seeing. I've been unable to reproduce it myself locally on my own test boxes. :-/

Thu, Apr 1, 8:55 PM
jhb added a comment to D29524: tcp: Shouldn't drain empty sbuf.

I would probably instead have changed sbuf_drain() to just return instead of asserting. Asserting made sense when it was a private implementation function, as a public API I think just returning instead of asserting is better behavior.

Thu, Apr 1, 5:38 PM
jhb added a comment to D29529: x86: remove gcov kernel support//config(8): remove support for -p.

Is there some more glue in sys/conf we can axe as well? Specifically, all the places that check PROFLEVEL?

Thu, Apr 1, 5:36 PM
jhb added inline comments to D29527: cxgbe(4): Always use the per-VI callout to read interface stats..
Thu, Apr 1, 5:30 PM
jhb added reviewers for D29538: bhyve: Move the gdb_active check to gdb_cpu_suspend().: elenamihailescu22_gmail.com, darius.mihaim_gmail.com.
Thu, Apr 1, 4:48 PM
jhb requested review of D29538: bhyve: Move the gdb_active check to gdb_cpu_suspend()..
Thu, Apr 1, 4:47 PM

Wed, Mar 31

jhb added inline comments to D29518: cryptocheck: Expand the set of sizes tested by -z..
Wed, Mar 31, 10:00 PM
jhb added a comment to D29517: ossl: Don't encryt/decrypt too much data for chacha20..

Drew had a panic in production where the relevant mbuf had a cipher text payload of 0x1b7 bytes in length. I was able to reproduce with a size of 112 (one full block plus enough of a partial block that the poly1305 hash equaled a second full block). I've tested this fix with both sizes. The following review for cryptocheck expands the sizes covered to handle this case, but it in a slightly more general way.

Wed, Mar 31, 5:30 PM
jhb requested review of D29517: ossl: Don't encryt/decrypt too much data for chacha20..
Wed, Mar 31, 5:26 PM
jhb requested review of D29518: cryptocheck: Expand the set of sizes tested by -z..
Wed, Mar 31, 5:26 PM
jhb accepted D29483: tcp: drain tcp_hostcache_list in between per-bucket locks.
Wed, Mar 31, 3:29 PM
jhb accepted D29481: Export sbuf_drain to orchestrate lock and drain action.
Wed, Mar 31, 3:29 PM

Tue, Mar 30

jhb accepted D29496: x86: clear %db registers in new process.
Tue, Mar 30, 10:58 PM
jhb accepted D29503: cxgbe/t4_tom: restore socket's protosw before entering TIME_WAIT..
Tue, Mar 30, 10:57 PM

Mon, Mar 29

jhb committed R10:77ffc9399dad: Stop installing kernel-only crypto headers to /usr/include/crypto. (authored by jhb).
Stop installing kernel-only crypto headers to /usr/include/crypto.
Mon, Mar 29, 11:24 PM
jhb committed R10:325427f71d37: xnb: Don't pass SIOC{ADD,DEL}MULTI to ifmedia_ioctl(). (authored by jhb).
xnb: Don't pass SIOC{ADD,DEL}MULTI to ifmedia_ioctl().
Mon, Mar 29, 10:47 PM
jhb committed R10:f9d590884ae3: Rename linux_set_upcall_kse() to linux_set_upcall(). (authored by jhb).
Rename linux_set_upcall_kse() to linux_set_upcall().
Mon, Mar 29, 10:46 PM
jhb committed R10:10438a42f166: x86: Reduce code duplication in cpu_fork() and cpu_copy_thread(). (authored by jhb).
x86: Reduce code duplication in cpu_fork() and cpu_copy_thread().
Mon, Mar 29, 10:46 PM
jhb committed R10:c6b96921a819: aarch64: Clear TLS registers during exec(). (authored by jhb).
aarch64: Clear TLS registers during exec().
Mon, Mar 29, 10:35 PM
jhb committed R10:d8debda0d0a3: x86: Update some stale comments in cpu_fork() and cpu_copy_thread(). (authored by jhb).
x86: Update some stale comments in cpu_fork() and cpu_copy_thread().
Mon, Mar 29, 10:14 PM
jhb committed R10:1a325aa20403: Remove unused wrappers around kproc_create() and kproc_exit(). (authored by jhb).
Remove unused wrappers around kproc_create() and kproc_exit().
Mon, Mar 29, 10:13 PM
jhb committed R10:053eb17635e1: Set TDP_KTHREAD before calling cpu_fork() and cpu_copy_thread(). (authored by jhb).
Set TDP_KTHREAD before calling cpu_fork() and cpu_copy_thread().
Mon, Mar 29, 10:13 PM
jhb committed R10:60ae7f745da0: x86: Always use clean FPU and segment base state for new kthreads. (authored by jhb).
x86: Always use clean FPU and segment base state for new kthreads.
Mon, Mar 29, 10:13 PM
jhb committed R10:cc6d86f2ff9c: Don't pass RFPROC to kproc_create(), it is redundant. (authored by jhb).
Don't pass RFPROC to kproc_create(), it is redundant.
Mon, Mar 29, 10:13 PM
jhb committed R10:1ebb5d7920fe: x86: Copy the FPU/XSAVE state from the creating thread to new threads. (authored by jhb).
x86: Copy the FPU/XSAVE state from the creating thread to new threads.
Mon, Mar 29, 10:13 PM
jhb committed R10:57d7992b50ba: amd64: Only update fsbase/gsbase in pcb for curthread. (authored by jhb).
amd64: Only update fsbase/gsbase in pcb for curthread.
Mon, Mar 29, 10:13 PM
jhb committed R10:77dfa7cfafb7: amd64: Cleanups to setting TLS registers for Linux binaries. (authored by jhb).
amd64: Cleanups to setting TLS registers for Linux binaries.
Mon, Mar 29, 10:13 PM
jhb added a comment to D29486: geom_vinum: add gone_in deprecation notice.

For GELI I took the strategy of spamming a message each time a new volume was created with a deprecated algorithm via geli init, and a message each time an existing volume created a new provider (so if you booted with 5 GELI volumes and each used a deprecated algorithm, you got 5 warnings IIRC). The geli changes for this were in R10:5c420aae3b18027809507dc9142182d4290897bf. Unlike some other warnings I added for IPsec, I didn't bother with rate limiting.

Mon, Mar 29, 5:58 PM
jhb added a comment to D29486: geom_vinum: add gone_in deprecation notice.

Have you tried testing it? Maybe try to 'gvinum create' across two memory disks or ZVOLs to see if the warning works? You could maybe then do 'gvinum detach' and 'gvinum attach' to replicate boot-time probing. That is what I did when testing deprecation warnings for GELI algorithms I deprecated for 13. I don't have a good opinion on where the warning should go, and if it works in testing I'm fine with whatever works.

Mon, Mar 29, 5:44 PM
jhb added a comment to D29483: tcp: drain tcp_hostcache_list in between per-bucket locks.

FYI, the best way to indicate dependencies is to setup needed commits as "parent revisions" via "Edit Related Revisions" at the top-right (or use 'git arc' to upload changes together).

Mon, Mar 29, 5:40 PM
jhb committed R10:4d5460a720c5: bhyve: Enable virtio-scsi legacy config parsing. (authored by jhb).
bhyve: Enable virtio-scsi legacy config parsing.
Mon, Mar 29, 5:28 PM
jhb closed D29438: bhyve: Enable virtio-scsi legacy config parsing..
Mon, Mar 29, 5:28 PM
jhb accepted D29174: gdb: report specific stop reason for watchpoints.
Mon, Mar 29, 5:24 PM
jhb added a comment to D29424: gvinum: add deprecation notice.

Agreed, but that can be a separate review :)

Mon, Mar 29, 5:18 PM

Sat, Mar 27

jhb requested review of D29438: bhyve: Enable virtio-scsi legacy config parsing..
Sat, Mar 27, 12:28 AM

Fri, Mar 26

jhb committed R10:fe496dc02a9a: cxgbe: Make the TOE TLS stats per-queue instead of per-port. (authored by jhb).
cxgbe: Make the TOE TLS stats per-queue instead of per-port.
Fri, Mar 26, 10:21 PM
jhb committed R10:077ba6a845fa: cxgbe: Add a struct sge_ofld_txq type. (authored by jhb).
cxgbe: Add a struct sge_ofld_txq type.
Fri, Mar 26, 10:21 PM
jhb closed D29383: cxgbe: Make the TOE TLS stats per-queue instead of per-port..
Fri, Mar 26, 10:20 PM
jhb closed D29382: cxgbe: Add a struct sge_ofld_txq type..
Fri, Mar 26, 10:20 PM

Thu, Mar 25

jhb added inline comments to D29382: cxgbe: Add a struct sge_ofld_txq type..
Thu, Mar 25, 10:09 PM
jhb updated the diff for D29382: cxgbe: Add a struct sge_ofld_txq type..
  • Fix with RATELIMIT and malloc size.
Thu, Mar 25, 10:06 PM
jhb added a comment to D29424: gvinum: add deprecation notice.

Should we add a 'gone_in()' call somewhere in the kernel as well, e.g. when creating a new gvinum GEOM provider?

Thu, Mar 25, 8:55 PM
jhb accepted D29394: pci: enhance printf for leaked MSI[-X] vectors.
Thu, Mar 25, 8:53 PM
jhb added a comment to D29174: gdb: report specific stop reason for watchpoints.

I think this is probably fine. I would probably let the MD hook actually output whatever it wanted instead of requiring it to only do a single name and returning a value. That is I would just use gdb_tx_* directly in the new MD routine and not have the MI gdb_stop_reason intermediary, but I'm fine with it either way.

Thu, Mar 25, 8:53 PM
jhb accepted D29390: resolv_test: Fix racy exit check, remove mutexes, and reduce output.
Thu, Mar 25, 8:49 PM
jhb added a comment to D29291: cxgbe(4): Allow a T6 adapter to switch between TOE and KTLS mode..

I should really add a TLS section to cxgbe(4) as well the talks about TOE TLS vs NIC TLS. It's on my todo list, just haven't done it yet :(

Thu, Mar 25, 4:53 PM
jhb accepted D29291: cxgbe(4): Allow a T6 adapter to switch between TOE and KTLS mode..

I would use "NIC TLS" in a few more messages to be consistent.

Thu, Mar 25, 4:52 PM

Wed, Mar 24

jhb committed R10:9f40a3be3d5d: bhyve hostbridge: Rename "device" property to "devid". (authored by jhb).
bhyve hostbridge: Rename "device" property to "devid".
Wed, Mar 24, 4:39 PM

Tue, Mar 23

jhb updated the diff for D29382: cxgbe: Add a struct sge_ofld_txq type..
  • Add alloc/free_ofld_txq() routines.
Tue, Mar 23, 9:54 PM
jhb updated the diff for D29383: cxgbe: Make the TOE TLS stats per-queue instead of per-port..
  • Update for alloc/free_txq_ofld().
Tue, Mar 23, 9:54 PM
jhb added a comment to D29291: cxgbe(4): Allow a T6 adapter to switch between TOE and KTLS mode..

Generally looks good to me. Nice to have this runtime togglelable now. I assume the minimum FW version we require now supports the new feature?

Tue, Mar 23, 9:23 PM
jhb added inline comments to D29390: resolv_test: Fix racy exit check, remove mutexes, and reduce output.
Tue, Mar 23, 8:04 PM
jhb added a comment to D29394: pci: enhance printf for leaked MSI[-X] vectors.

Hmm, we don't list the leaked resources for any other types so that's a bit inconsistent (e.g. we don't list the address or rid's of BARs).

Tue, Mar 23, 7:59 PM
jhb added inline comments to D29390: resolv_test: Fix racy exit check, remove mutexes, and reduce output.
Tue, Mar 23, 7:54 PM
jhb accepted D29381: Drop WARNS for ktrdump(8).
Tue, Mar 23, 7:45 PM

Mon, Mar 22

jhb committed R10:30e25f930445: nlmrsa: Mark deprecated for 14. (authored by jhb).
nlmrsa: Mark deprecated for 14.
Mon, Mar 22, 10:56 PM
jhb committed R10:6a1dbc8ba956: iscsi: Mark iSCSI CAM sims as non-pollable. (authored by jhb).
iscsi: Mark iSCSI CAM sims as non-pollable.
Mon, Mar 22, 10:53 PM
jhb committed R10:9df8f5a9c684: cam: Don't permit crashdumps on non-pollable devices. (authored by jhb).
cam: Don't permit crashdumps on non-pollable devices.
Mon, Mar 22, 10:52 PM
jhb committed R10:444aff36738a: cam: Permit non-pollable sims. (authored by jhb).
cam: Permit non-pollable sims.
Mon, Mar 22, 10:52 PM
jhb added inline comments to D29383: cxgbe: Make the TOE TLS stats per-queue instead of per-port..
Mon, Mar 22, 10:48 PM
jhb requested review of D29383: cxgbe: Make the TOE TLS stats per-queue instead of per-port..
Mon, Mar 22, 10:44 PM
jhb requested review of D29382: cxgbe: Add a struct sge_ofld_txq type..
Mon, Mar 22, 10:44 PM
jhb added inline comments to D29281: Relnotes/13.0: changed AES-NI identification by geli.
Mon, Mar 22, 9:08 PM
jhb added inline comments to D29381: Drop WARNS for ktrdump(8).
Mon, Mar 22, 7:19 PM
jhb added inline comments to D29378: pf: Handle unmapped mbufs when computing checksums.
Mon, Mar 22, 7:17 PM
jhb accepted D27625: truss: improved support for decoding compat32 arguments.

It's fine with me modulo the nits.

Mon, Mar 22, 7:13 PM