Page MenuHomeFreeBSD

jhb (John Baldwin)
User

Projects (10)

User Details

User Since
Mar 11 2014, 8:46 PM (662 w, 21 h)

Recent Activity

Yesterday

jhb updated subscribers of D61661: [RFC] Introduce Solaris-style Door IPC subsystem (preview).

Hmmm, this is actually the second attempt at implementing this API (see D34097 authored by @bnovkov). The approach here seems quite a bit different however. What is your motivation for working on this?

Tue, Nov 17, 3:35 PM

Mon, Nov 16

jhb added a comment to D59652: trap_cfi(9): add man page.

To be clear, the whole stack is approved by me. I haven't done detailed review of the actual annotations themselves, but I'm fine if you merge them all with me as an "Approved by" assuming you've tested them (which you already describe doing both in comments and the commit logs). You might want to explicitly note you tested the amd64 ones with kgdb which might preempt some followup questions you might otherwise get.

Mon, Nov 16, 7:04 PM
jhb committed rG125b88f5aa91: vmm: release the lock when vrtc_snapshot() fails (authored by wanpengqian_gmail.com).
Mon, Nov 16, 7:01 PM
jhb committed rG018bd70c4ac9: vmm: do not let the RTC's periodic interrupt drift (authored by wanpengqian_gmail.com).
Mon, Nov 16, 7:01 PM
jhb closed D60232: vmm: release the lock when vrtc_snapshot() fails.
Mon, Nov 16, 7:01 PM
jhb closed D60233: vmm: do not let the RTC's periodic interrupt drift.
Mon, Nov 16, 7:01 PM
jhb accepted D60233: vmm: do not let the RTC's periodic interrupt drift.
Mon, Nov 16, 6:43 PM
jhb closed D60230: bhyve: e82545: deassert the interrupt when its last cause is masked.
Mon, Nov 16, 6:34 PM
jhb committed rG596a470a1baa: bhyve: e82545: deassert the interrupt when its last cause is masked (authored by wanpengqian_gmail.com).
Mon, Nov 16, 6:34 PM
jhb accepted D60230: bhyve: e82545: deassert the interrupt when its last cause is masked.
Mon, Nov 16, 4:51 PM
jhb added inline comments to D60228: bhyve: nvme: drop the completions of commands from before a reset.
Mon, Nov 16, 4:48 PM
jhb added a comment to D59652: trap_cfi(9): add man page.

In general I think this series is fine to merge, though my suggestions about the commit logs on the amd64 review I think are relevant for all the MD commit logs.

Mon, Nov 16, 4:41 PM
jhb added a comment to D59649: trap_cfi(9): introduce amd64 implementation.

Some suggestions on the commit log, maybe something like:

Mon, Nov 16, 4:39 PM
jhb accepted D59510: devel/freebsd-gcc*: also build libgcov.a.
Mon, Nov 16, 4:25 PM
jhb added a comment to D59959: opencrypto: avoid false KMSAN reports.

I think I agree with markj@. There aren't that many drivers you'd need to update for this? Ideally you'd mark the buffers after the assembly routines return using buffer/length passed to the assembly routines.

Mon, Nov 16, 4:22 PM
jhb added inline comments to D60671: Remove the i386 kernel and i386-only device drivers.
Mon, Nov 16, 3:21 PM

Sat, Nov 14

jhb requested review of D61575: nvmf_tcp: Avoid unsigned arithmetic overflow for data lengths in PDUs.
Sat, Nov 14, 5:40 PM
jhb requested review of D61574: nvmft: Add various TCP transport tests.
Sat, Nov 14, 5:40 PM
jhb requested review of D61573: nvmf_tcp: Handle data digest mismatches for received in-capsule data.
Sat, Nov 14, 5:40 PM
jhb requested review of D61571: nvmf_tcp: Ensure the PDU length is long enough for the header digest.
Sat, Nov 14, 5:40 PM
jhb requested review of D61570: nvmf_tcp: Avoid underflow when validating PDU header fields.
Sat, Nov 14, 5:40 PM
jhb requested review of D61572: nvmf_tcp: Return all of the header that has been received for PDU errors.
Sat, Nov 14, 5:40 PM
jhb requested review of D61569: libnvmf: Export some APIs useful for writing tests.
Sat, Nov 14, 5:40 PM
jhb requested review of D61568: nvmft: Switch to 100ms granularity for the KeepAlive timer.
Sat, Nov 14, 5:40 PM
jhb requested review of D61567: nvmft: When setting the KeepAlive timer, honor granularity from cdata.
Sat, Nov 14, 5:40 PM
jhb requested review of D61565: nvmft: Add various tests for getting and setting NVMe features.
Sat, Nov 14, 5:39 PM
jhb requested review of D61566: nvmft: Add a simple test for KeepAlive timeouts.
Sat, Nov 14, 5:39 PM
jhb requested review of D61564: nvmft: Shuffle a couple of fields for better structure packing.
Sat, Nov 14, 5:39 PM
jhb requested review of D61563: nvmft: Implement missing mandatory NVMe features.
Sat, Nov 14, 5:39 PM
jhb requested review of D61562: ctl: Support NVMe namespace-specific SET_FEATURES/GET_FEATURES commands.
Sat, Nov 14, 5:39 PM
jhb requested review of D61561: nvmft: Allow the keep alive timer to be managed as a feature.
Sat, Nov 14, 5:39 PM
jhb requested review of D61559: nvmft: Add additional error checking for SET_FEATURES commands.
Sat, Nov 14, 5:39 PM
jhb added a reviewer for D61553: ctld nvmf: Honor host-address and host-nqn restrictions: markj.
Sat, Nov 14, 5:39 PM
jhb requested review of D61560: nvmft: Initial support for the GET_FEATURES command.
Sat, Nov 14, 5:39 PM
jhb requested review of D61557: nvmft: Add an upper bound on the maximum transfer size (MDTS).
Sat, Nov 14, 5:39 PM
jhb requested review of D61558: fixup? nvmft: Add an upper bound on the maximum transfer size (MDTS).
Sat, Nov 14, 5:39 PM
jhb requested review of D61556: nvmft: Add some simple I/O tests.
Sat, Nov 14, 5:39 PM
jhb requested review of D61555: nvmft: Return the correct error if the buffer length for IDENTIFY is wrong.
Sat, Nov 14, 5:39 PM
jhb requested review of D61554: nvmfd: Always return a CQE with a suitable error if kernel handoff fails.
Sat, Nov 14, 5:39 PM
jhb requested review of D61553: ctld nvmf: Honor host-address and host-nqn restrictions.
Sat, Nov 14, 5:39 PM
jhb requested review of D61552: ctld nvmf: Return a more specific error for an unknown SubNQN.
Sat, Nov 14, 5:39 PM
jhb requested review of D61550: libnvmf: Add helper routines to return errors for CONNECT.
Sat, Nov 14, 5:39 PM
jhb requested review of D61551: ctld nvmf: Return a generic internal error response if queue handoff fails.
Sat, Nov 14, 5:39 PM
jhb updated the diff for D60477: nvmft: Handle transport errors on unaccepted qpairs.

Widen locking in nvmf_controller_error and remove overly-clever double-check

Sat, Nov 14, 5:37 PM
jhb updated the diff for D60671: Remove the i386 kernel and i386-only device drivers.

Folding in extra header trimming

Sat, Nov 14, 5:17 PM
jhb added a comment to D61349: libprocstat: Add static assertions for kinfo structure sizes.
In D61349#1410680, @jhb wrote:
In D61349#1406807, @kib wrote:

I like this, but I have a small request. I am curious if you agree with it.

Can we have a separate source file in libprocstat which only consists of assertions? This is from my experience with maintaining the asserts in kern_thread.c, where to get past compilation I have to fix the layout constants first, which was a secondary to get the code compiled.
It is less important for libprocstat since kinfo must not change often, but IMO it is worth it by being able to temporary remove the assert file from the compilation.

Anyway I am fine with the change as is as well.

Mmm, I would not mind that. For the kernel I generally can find some other object file, yes, though guessing which one is already built can be annoying. :)

Sat, Nov 14, 5:15 PM
jhb abandoned D60668: sys: Use armv7 instead of i386 for 32-bit KBI assertions.
Sat, Nov 14, 5:08 PM
jhb committed rG931a5d8866bc: acpi_pci_link: Handle IRQ values > 255.
Sat, Nov 14, 4:04 PM
jhb committed rG6c1b6f7cf814: acpi_pci_link: Factor out some duplicated code into a single copy.
Sat, Nov 14, 4:04 PM
jhb committed rGdf10f7b7500c: acpi_pci_link: Remove a helper variable from _PRS parsing.
Sat, Nov 14, 4:03 PM
jhb committed rG81dab94a5452: acpi_pci_link: Ignore tunables with negative values.
Sat, Nov 14, 4:03 PM
jhb added a comment to D61349: libprocstat: Add static assertions for kinfo structure sizes.
In D61349#1406807, @kib wrote:

I like this, but I have a small request. I am curious if you agree with it.

Can we have a separate source file in libprocstat which only consists of assertions? This is from my experience with maintaining the asserts in kern_thread.c, where to get past compilation I have to fix the layout constants first, which was a secondary to get the code compiled.
It is less important for libprocstat since kinfo must not change often, but IMO it is worth it by being able to temporary remove the assert file from the compilation.

Anyway I am fine with the change as is as well.

Sat, Nov 14, 3:32 PM

Thu, Nov 12

jhb committed rG02c0eefc4e0b: nvmft: Document all CTL_NVMF status codes.
Thu, Nov 12, 3:45 PM
jhb committed rGbfdce403f4a6: nvme: Typo ranage -> range.
Thu, Nov 12, 3:45 PM
jhb committed rG9a99fa3932dd: ctld: Fix namespace IDs for exported NVMe namespaces part 2.
Thu, Nov 12, 3:45 PM
jhb committed rGcc33771bf771: ctld: Use expand_unsigned for parsing unsigned NVMe option values.
Thu, Nov 12, 3:45 PM
jhb abandoned D60534: nvmft: Add a test for CONNECT with an invalid record format.

Replaced by D61468

Thu, Nov 12, 2:49 PM
jhb requested review of D61468: nvmft: Add tests for various invalid fields in CONNECT commands.
Thu, Nov 12, 2:46 PM
jhb updated the diff for D60469: libnvmf: Improve handling of the CATTR field in CONNECT commands.

Update for libnvmf now owning sockets and various other refactors for later tests

Thu, Nov 12, 2:44 PM
jhb updated the diff for D60479: nvmft: Add various tests for fetching log pages.

Update for libnvmf now owning sockets and various other refactors for later tests

Thu, Nov 12, 2:44 PM
jhb updated the diff for D60478: nvmft: Add a test for an I/O queue with a byte-swapped queue ID.

Update for libnvmf now owning sockets and various other refactors for later tests

Thu, Nov 12, 2:44 PM
jhb updated the diff for D60473: nvmft: Add tests for various shutdown/disconnect scenarios.

Update for libnvmf now owning sockets

Thu, Nov 12, 2:41 PM
jhb updated the diff for D60472: nvmft: Cancel any existing terminate task when enabling a controller.

Hold lock across terminate task scheduling

Thu, Nov 12, 2:41 PM
jhb requested review of D61467: libnvmf: Change qpairs to own underlying transport connections.
Thu, Nov 12, 2:40 PM

Mon, Nov 9

jhb retitled D60473: nvmft: Add tests for various shutdown/disconnect scenarios from nvmft: Add tests for various various shutdown/disconnect scenarios
Mon, Nov 9, 9:37 PM
jhb retitled D60472: nvmft: Cancel any existing terminate task when enabling a controller from nvmft: Cancel and drain any existing terminate task when enabling a controller
Mon, Nov 9, 9:17 PM
jhb added inline comments to D60472: nvmft: Cancel any existing terminate task when enabling a controller.
Mon, Nov 9, 7:35 PM
jhb added a comment to D60671: Remove the i386 kernel and i386-only device drivers.

After trimming some more bits from libkvm in userspace, I've done another round of trimming in sys/i386/include. For now you can find the "new" updates in a fixup commit here: https://github.com/freebsd/freebsd-src/commit/825e61d8154dd9913d2dc92fa9f74f7fa060dca2 The interesting changes perhaps are the removals from <machine/cpufunc.h> as potentially some of those could be used in userspace in ports maybe (but not likely).

Mon, Nov 9, 6:42 PM
jhb added inline comments to D60671: Remove the i386 kernel and i386-only device drivers.
Mon, Nov 9, 6:38 PM
jhb requested review of D61352: libkvm: Remove a fallback for pre-5.0 i386 kernels.
Mon, Nov 9, 6:35 PM
jhb requested review of D61351: libkvm: Drop static assertions for the i386 backend.
Mon, Nov 9, 6:35 PM
jhb requested review of D61350: kdump: Use types and constants from amd64 linux32 for Linux on i386.
Mon, Nov 9, 6:33 PM
jhb added a comment to D60668: sys: Use armv7 instead of i386 for 32-bit KBI assertions.
In D60668#1406320, @jhb wrote:

See D61348 and D61349 for changes to enable the kinfo assertions on all architectures and to add them into libprocstat in userspace which covers the -m32 case during lib32 builds.

Mon, Nov 9, 6:33 PM
jhb added a comment to D60668: sys: Use armv7 instead of i386 for 32-bit KBI assertions.

See D61348 and D61349 for changes to enable the kinfo assertions on all architectures and to add them into libprocstat in userspace which covers the -m32 case during lib32 builds.

Mon, Nov 9, 6:30 PM
jhb requested review of D61349: libprocstat: Add static assertions for kinfo structure sizes.
Mon, Nov 9, 6:28 PM
jhb requested review of D61348: sys: Enable kinfo structure size assertions for all architectures.
Mon, Nov 9, 6:28 PM
jhb requested review of D61347: kdump: Fix some cpp conditionals.
Mon, Nov 9, 6:27 PM
jhb committed rGf5949447047d: krb5: Don't include libsys in LIBADD.
Mon, Nov 9, 6:26 PM
jhb closed D60724: krb5: Don't include libsys in LIBADD.
Mon, Nov 9, 6:26 PM
jhb closed D60723: pam_krb5: Append to CFLAGS instead of replacing.
Mon, Nov 9, 6:26 PM
jhb committed rG9e99355a33f2: pam_krb5: Append to CFLAGS instead of replacing.
Mon, Nov 9, 6:26 PM
jhb committed rG17a42690bbd7: .github: Update actions version to 7.
Mon, Nov 9, 6:22 PM
jhb accepted D61309: acpi_pci_link: Preemptively avoid buffer overflow.

I guess this is ok to do as a separate commit, but I think it makes more sense to fold this into the 3rd commit as it isn't meaningful until then. Also, I think in your last commit, it's probably a bug that you don't expand pos_irq to a full u_int since _PRS can also use extended IRQ types.

Mon, Nov 9, 4:32 PM
jhb added inline comments to D61311: acpi_pci_link: Support IRQs > 255.
Mon, Nov 9, 4:24 PM
jhb added inline comments to D61310: acpi_pci_link: Treat negative IRQs as invalid.
Mon, Nov 9, 4:14 PM

Mon, Nov 2

jhb accepted D61035: linker_set: Remove obsolete macros.
Mon, Nov 2, 11:18 PM

Tue, Oct 20

jhb added inline comments to D60477: nvmft: Handle transport errors on unaccepted qpairs.
Tue, Oct 20, 3:21 PM
jhb added inline comments to D60671: Remove the i386 kernel and i386-only device drivers.
Tue, Oct 20, 2:41 PM

Oct 19 2026

jhb added inline comments to D60477: nvmft: Handle transport errors on unaccepted qpairs.
Oct 19 2026, 5:20 PM
jhb added inline comments to D60477: nvmft: Handle transport errors on unaccepted qpairs.
Oct 19 2026, 5:19 PM
jhb added inline comments to D60807: bxe: Use __SIZEOF_LONG__ to compute BITS_PER_LONG.
Oct 19 2026, 5:10 PM

Oct 17 2026

jhb updated the summary of D60674: isa: Retire the ISAPNP option and related code.
Oct 17 2026, 4:26 PM
jhb updated the summary of D60673: Retire the pnpinfo(8) utility.
Oct 17 2026, 4:25 PM
jhb requested review of D60809: x86 lapic: Assume amd64 values when setting IDT entries.
Oct 17 2026, 4:24 PM
jhb requested review of D60808: amd64: Close gap between the TLB shootdown and bitmap IPIs.
Oct 17 2026, 4:24 PM
jhb requested review of D60807: bxe: Use __SIZEOF_LONG__ to compute BITS_PER_LONG.
Oct 17 2026, 4:22 PM
jhb requested review of D60806: padlock: Remove architecture #ifdefs.
Oct 17 2026, 4:22 PM
jhb added inline comments to D60671: Remove the i386 kernel and i386-only device drivers.
Oct 17 2026, 4:17 PM
jhb added inline comments to D60671: Remove the i386 kernel and i386-only device drivers.
Oct 17 2026, 3:58 PM
jhb added inline comments to D60671: Remove the i386 kernel and i386-only device drivers.
Oct 17 2026, 3:56 PM