Page MenuHomeFreeBSD

imp (Warner Losh)
User

Projects (8)

User Details

User Since
Jun 2 2014, 4:20 PM (650 w, 2 d)

Recent Activity

Today

imp accepted D61615: arm64: Fix EFI runtime lookups with efi_disable_vmap.

So apart from the commit message, this is OK. It does add an additional bogus va == pa requirement for runtime entries, but we already do it elsewhere (which is also bogus: I ran for years on a LinuxBoot arm64 machine w/o this requirement... But I've lost access to that environment so since I can't test it, I can't ask someone else to fix something they can't test)....

Wed, Nov 18, 3:08 AM · ARM, Contributor Reviews (src)
imp added a reviewer for D61640: x86: use AMD enumeration for IBRS, STIBP and SSBD: ali_mashtizadeh.com.
Wed, Nov 18, 3:02 AM

Sun, Nov 15

imp accepted D60672: sys: Remove PAE #ifdef's in drivers.
Sun, Nov 15, 4:21 AM

Sat, Nov 14

imp accepted D61532: uefisign: use ASN1_STRING_set0 instead of ASN1_STRING_set.
Sat, Nov 14, 5:43 AM

Fri, Nov 13

imp accepted D59141: loader.efi: Measure boot policy and loaded contents with a TPM.
Fri, Nov 13, 11:39 PM
imp accepted D61477: libsa: Fall back to DISCOVER promptly when INIT-REBOOT is refused.

This looks reasonable to me. I'd be interested in @mav 's take on it.

Fri, Nov 13, 7:52 PM
imp accepted D61418: msdosfs: bound slot offset in mbnambuf_write().

This looks good... but it might be good to have at least a reference FAT image with the mal-formed LFN entry for regression.

Fri, Nov 13, 4:39 PM

Thu, Nov 12

imp added a comment to D61476: loader.efi: Configure the boot interface when looking for an initmd.

I'll take a look at this. I also have a series that has a fix for this in it. I'll see which is better.

Thu, Nov 12, 10:41 PM
imp accepted D61456: Prune unused luaconf.h copies.

Yea, I forgot to remove these.

Thu, Nov 12, 6:38 AM
imp added a comment to D61175: liblua: Accept + and e in io.popen mode.

I think it is time to axe those too.

Thu, Nov 12, 3:13 AM
imp added a comment to D59856: pmc: Add native flamegraph generation.

codex also flags Three earlier defects also recur here: address-sign classification rejects valid low kernel addresses (line 154), unresolved frames are removed and entirely unresolved samples disappear (line 172), and L3-miss counting omits DRAM-served misses (lines 213–217).

Thu, Nov 12, 2:07 AM
imp added inline comments to D59855: pmc: Generate calltree based on pmc samples.
Thu, Nov 12, 1:56 AM
imp added inline comments to D58929: pmc: Analyze high latency memory accesses.
Thu, Nov 12, 1:42 AM
imp added a comment to D58928: pmc: Print IBS summary statistics.

codex flagged a couple spec vs code issues.

Thu, Nov 12, 1:35 AM
imp added a comment to D58927: pmc: Bargraphs and Bitfields and Histograms, Oh My!.

generally good (and cool functionality). Lots of nits.

Thu, Nov 12, 1:25 AM
imp added a comment to D58926: pmc: New annotate command.

So there's apparently a problem with --demangle

Thu, Nov 12, 1:08 AM
imp added a comment to D58925: pmc: Analyze branch mispredictions.

I know I'm being picky about the ordering here (and I should have flagged it a while ago), but there's a desire, when possible, to have each rev committed be bisectable. There's a proposal to allow 'loopback branches' to stack multiple commits and have only the last one be buildable, but that's not gone anywhere.

Thu, Nov 12, 1:00 AM
imp added inline comments to D58924: pmc: Display a program level breakdown of performance data.
Thu, Nov 12, 12:36 AM
imp added inline comments to D58923: pmc: Display system level breakdown of performance data.
Thu, Nov 12, 12:32 AM
imp added a comment to D61452: Pass the -b flag from the wrapper to the underlying filesystem checker.

Thi

Thu, Nov 12, 12:04 AM

Wed, Nov 11

imp accepted D59636: nvme: enforce the Abort Command Limit.
Wed, Nov 11, 5:11 PM

Tue, Nov 10

imp accepted D61416: pass(4): Wake select/kqueue waiters when an immediate CCB completes.
Tue, Nov 10, 11:07 PM
imp added a comment to D61091: aq: Add to sys/conf/NOTES.

Yea. Life happened. I'll try to build this tonight

Tue, Nov 10, 9:22 PM
imp accepted D61406: nvmecontrol: Range-check the format command's field values.
Tue, Nov 10, 9:21 PM
imp requested changes to D61401: sys: add __null_terminated and bound the hash.h helpers.

I'm not sure tagging every string in the kernel with __null_terminated is viable. Can you explain this a little more?

Tue, Nov 10, 8:17 PM
imp accepted D61396: sysvsem: Plug another place where we leak __sem_base to userspace.
Tue, Nov 10, 7:52 PM
imp accepted D61404: conf: add GENERIC-BOUNDS_SAFETY kernel configs.
Tue, Nov 10, 6:56 PM
imp accepted D61146: snmp_hostres(3): strengthen work on the devd(8) socket.

Oh! Right! this is what you needed all that for.

Tue, Nov 10, 3:46 PM
imp accepted D61356: nvmecontrol: Mask the fields in the Format NVM command.

Thanks ken! Looks great now I think.

Tue, Nov 10, 3:38 PM
imp accepted D61344: devd: define _PATH_DEVD_SOCKET in <paths.h>.
Tue, Nov 10, 7:16 AM
imp accepted D61345: unistd.h: sort sysconf(3) arguments.
Tue, Nov 10, 7:15 AM
imp accepted D61346: devd: expose maximum record size via sysconf(3).
Tue, Nov 10, 7:15 AM
imp added a comment to D61356: nvmecontrol: Mask the fields in the Format NVM command.

Code wise, this looks good.

Tue, Nov 10, 6:54 AM
imp requested changes to D61356: nvmecontrol: Mask the fields in the Format NVM command.

Co-authored-by is reserved for humans. Use assisted-by instead.

Tue, Nov 10, 6:54 AM
imp accepted D61350: kdump: Use types and constants from amd64 linux32 for Linux on i386.
Tue, Nov 10, 2:14 AM
imp accepted D61348: sys: Enable kinfo structure size assertions for all architectures.
Tue, Nov 10, 1:39 AM
imp accepted D61297: kobj: Make the containers of pointers to base classes constant.
Tue, Nov 10, 1:24 AM

Sun, Nov 8

imp accepted D61301: asmc: Use c99 initializers with for-loop counters to improve scoping.
Sun, Nov 8, 6:42 AM
imp accepted D61302: asmc: Fix style(9) issues.
Sun, Nov 8, 6:41 AM

Fri, Nov 6

imp accepted D59854: mps(4): add support for hostboot.
Fri, Nov 6, 11:24 PM
imp accepted D61276: sys/kobj.h: Convert macro DEFINE_CLASS to designated initializers.
Fri, Nov 6, 11:22 PM
imp accepted D61277: camdd: Fix the worker startup race.
Fri, Nov 6, 11:15 PM
imp accepted D61245: libsecureboot restore default trust anchors.
Fri, Nov 6, 12:23 AM

Thu, Nov 5

imp accepted D61225: libfetch: Use monotonic clock for timeouts.
Thu, Nov 5, 10:48 PM
imp accepted D61191: uart: Add MediaTek UART driver.

This looks good, but I didn't verify all the details of clocks and baud generation.

Thu, Nov 5, 9:41 PM
imp accepted D59155: nvme: program the arbitration burst from Identify RAB.

I'm not going to be able to do the netflix testing after all, so go for it. If there's problem, then they can report them on upgrade.

Thu, Nov 5, 9:27 PM

Wed, Nov 4

imp accepted D61007: libsecureboot: add support for more signatures.

This looks fairly clean. I'm not seeing anything really big. I'd chunk it smaller, but after going back and forth, I'm just going to mention it.

Wed, Nov 4, 10:40 PM
imp accepted D61145: loader: Handle unavailable VBE during mode lookup.
Wed, Nov 4, 5:03 PM
imp added a comment to D60971: move SEC_TO_NSEC() out of pf code.

IMHO, more review is needed here. This is a system header that is installed in /usr/include. We should check if other common APIs exist or not for this operation. Does it match existing operations that we already define there?

Wed, Nov 4, 4:58 PM
imp added a comment to D60972: introduce MSEC_TO_NSEC().

Somewhere in .9 since they are kernel only.
I'd look at tvtohz.9 as a model maybe?
I'd call the man page SEC_TO_NSEC.9 and have MLINKS for the other routines.
Shouldn't be a big add, if you need help, give a yell.

Wed, Nov 4, 4:52 PM
imp accepted D60988: sdiob: copy vendor/device IDs from card common CIS.
Wed, Nov 4, 7:21 AM

Tue, Nov 3

imp requested changes to D60972: introduce MSEC_TO_NSEC().

So I approved the last one, but not this one. We need man pages for these things.

Tue, Nov 3, 9:53 PM
imp added inline comments to D61161: netlink: Pre-register the missing devctl event systems.
Tue, Nov 3, 9:00 PM
imp accepted D61164: bnxt: avoid unintialized use of media_type in bnxt_add_media_types().
Tue, Nov 3, 8:53 PM

Mon, Nov 2

imp accepted D61022: cam: Fix the mmc_data mapping length for XPT_MMC_IO.

woof! Way better.

Mon, Nov 2, 8:47 PM
imp accepted D61021: cam: Fix XPT_MMC_IO buffer unmap order in cam_periph_unmapmem().
Mon, Nov 2, 8:47 PM
imp accepted D60888: dwmmc: Add HS200 support.
Mon, Nov 2, 8:46 PM
imp accepted D61077: mmccam: Wait for the power to settle when probing a card.

I'm going to click OK here... but this can lead to 1ms or 2ms delays in processing other CAM devices since it's a busy wait. For MMC, that's likely OK. They don't usually have other devices on the system that are so performance critical that a stall like this would matter... However, as usage shifts, we may need to revisit one day. For now I think this is OK, with my mentioning the caveat in case people trace some future performance problem back to this commit and then to this review.

Mon, Nov 2, 8:45 PM
imp accepted D61111: ufshci: detect completions with the completion notification register.
Mon, Nov 2, 8:40 PM
imp accepted D61112: ufshci: add debug aids for driving the recovery path.
Mon, Nov 2, 8:40 PM
imp accepted D61113: ufshci: claim a slot before completing it.
Mon, Nov 2, 8:39 PM
imp accepted D61114: ufshci: keep a slot for admin requests.

Smart. NVMe didn't used to need anything like this, but it might now with apple hardware support.

Mon, Nov 2, 8:39 PM
imp accepted D61115: ufshci: tear down the controller in a safe order.
Mon, Nov 2, 8:38 PM
imp accepted D61116: ufshci: reclaim in-flight requests across a reset in SDB mode.
Mon, Nov 2, 8:38 PM
imp accepted D61117: ufshci: requeue requests that a reset aborted.

Similar comments: UFS mirrors NVMe closely in its setup, and nvme does this.

Mon, Nov 2, 8:37 PM
imp accepted D61118: ufshci: hold back I/O during a reset.

mirrors what nvme controller does. The different SIMs are different in how they handle this, but I'm partial to this method, and I think for the world of UFS, it's simplicity is better than other SIMs that need added complexity to deal with some things failing and some not.

Mon, Nov 2, 8:37 PM
imp accepted D61119: ufshci: fail the controller when a polled command never completes.
Mon, Nov 2, 8:35 PM
imp accepted D61120: ufshci: tell CAM when the controller is gone for good.
Mon, Nov 2, 8:35 PM
imp accepted D61121: ufshci: try the link startup more than once.
Mon, Nov 2, 8:34 PM
imp accepted D61122: ufshci: report the SCSI status and sense to CAM.

This looks right to me....

Mon, Nov 2, 8:33 PM
imp added a comment to D61106: arm64: Build and install kernel.bin with OCIBHYVE.

After reading through everything else, I'm not sure why we can't load the kernel.elf into low memory. But that's code on arm64 since we have no userboot.so there. I'm less worried about this than with my first comment, but I'd be interested to see the boot time differences.

Mon, Nov 2, 7:18 PM
imp added a comment to D61102: bhyve: Allow an arm64 kernel as the boot ROM.

JFYI I've been working towards something similar for amd64, though paused for now. There, bhyve needs to know which boot protocol to use, so I added a bootflavour option that one can specify on the command line; this lets one explicitly select from a small number of protocols, one of which would be "directly boot this amd64 kernel".

Here I guess that's not quite necessary, though I still don't love that we have to taste the bootrom image to see if it's a kernel. Do you think it'd be worth introducing that here?

Mon, Nov 2, 7:01 PM
imp added inline comments to D61102: bhyve: Allow an arm64 kernel as the boot ROM.
Mon, Nov 2, 6:54 PM
imp accepted D61110: pci: restore resizable BAR sizes in pci_cfg_restore().
Mon, Nov 2, 6:06 AM
imp accepted D61091: aq: Add to sys/conf/NOTES.

If it builds, I'll push it in. But NOTES is used by all the architectures, so it will take me a while to test.

Mon, Nov 2, 12:49 AM
imp accepted D60973: sdhci_pci: Add Intel Sunrise Point-LP SDXC controller.
Mon, Nov 2, 12:47 AM
imp requested changes to D61106: arm64: Build and install kernel.bin with OCIBHYVE.
Mon, Nov 2, 12:12 AM
imp added a comment to D61106: arm64: Build and install kernel.bin with OCIBHYVE.

Doesn't kernel.elf still load in this case? We've been shying away from kernel.bin in general, so I'd like to understand the need. The ELF loading code is small, especially when we don't have any drivers.

Mon, Nov 2, 12:12 AM
imp accepted D61105: release: Package the OCIBHYVE kernel as FreeBSD-kernel-ocibhyve.
Mon, Nov 2, 12:10 AM
imp requested changes to D61104: amd64: Add OCIBHYVE, the twin of the arm64 bhyve guest kernel.
Mon, Nov 2, 12:09 AM
imp requested changes to D61103: arm64: Add OCIBHYVE, a kernel for bhyve guests without a loader.
Mon, Nov 2, 12:07 AM

Sat, Oct 31

imp committed R11:a4ebe3c85250: sysutils/intel-pcm: Resign as maintianer.
Sat, Oct 31, 7:00 AM

Fri, Oct 30

imp accepted D60971: move SEC_TO_NSEC() out of pf code.
Fri, Oct 30, 5:29 PM
imp accepted D61023: mmc: Fix clock handling in mmc-pwrseq-simple set_power().
Fri, Oct 30, 5:15 PM
imp accepted D60884: mmccam: Add HS200 support for eMMC.

The code looks good, and make things safer. Not sure I understand the comment I flagged, but that's likely nothing.

Fri, Oct 30, 5:14 PM
imp accepted D61012: sdhci: Do not reprogram an unchanged clock.
Fri, Oct 30, 5:08 PM
imp accepted D60888: dwmmc: Add HS200 support.

My comments are generally small for this, so this is likely fine as is.
The tuning algorithm for the rockport I didn't audit beyond looking for obvious mistakes, everything else I put under the microscope.

Fri, Oct 30, 5:07 PM
imp accepted D61024: mmc: Add a power-up settling delay in mmc_fdt_set_power().
Fri, Oct 30, 4:46 PM
imp accepted D61009: bootpd: clamp the DHCP reply length to the packet buffer.
Fri, Oct 30, 5:26 AM

Thu, Oct 22

imp accepted D60964: mpr, mps: sleep 1 ms for the doorbell ack whatever hz is.

This is a proper migration to a newer API.

Thu, Oct 22, 4:44 AM

Wed, Oct 21

imp accepted D60939: libc: Add a man page caveat about the `w` modifier.
Wed, Oct 21, 11:27 PM
imp accepted D59533: camcontrol: Add support for NVMe namespace Identify.

Oopps! I thought I'd accepted this before now. Done.

Wed, Oct 21, 9:32 PM
imp added inline comments to D60884: mmccam: Add HS200 support for eMMC.
Wed, Oct 21, 9:21 PM
imp accepted D60921: mmccam: Add a helper to change the host settings.
Wed, Oct 21, 9:20 PM
imp accepted D59531: nvme: Add reservation, format and features command building support.
Wed, Oct 21, 9:18 PM
imp closed D60850: 2026Q3 status: storage stack improvements.

Landed w/o proper review markings, my bad.

Wed, Oct 21, 9:07 PM
imp closed D60849: 2026Q3 status: boot loader improvements.

Oh, But I pushed w/o the reviewed stuff. My bad.

Wed, Oct 21, 9:07 PM
imp added a comment to D60849: 2026Q3 status: boot loader improvements.

updated and pushed! Thanks for the feedback!

Wed, Oct 21, 9:06 PM
imp committed R9:0a189197d8e5: 2026Q3 status: storage stack improvements.
Wed, Oct 21, 9:03 PM
imp committed R9:85ec78b84693: 2026Q3 status: boot loader improvements.
Wed, Oct 21, 9:03 PM