imp (Warner Losh)
User

Projects

User Details

User Since
Jun 2 2014, 4:20 PM (147 w, 3 d)

Recent Activity

Wed, Mar 29

imp committed rS316171: xfsread inlined uses more space, so remove the inline tag. This.
xfsread inlined uses more space, so remove the inline tag. This
Wed, Mar 29, 6:36 PM

Tue, Mar 28

imp committed rS316100: Remove -fno-guess-branch-probability and -fno-unit-at-a-time..
Remove -fno-guess-branch-probability and -fno-unit-at-a-time.
Tue, Mar 28, 6:09 PM
imp committed rS316079: Simply retire the sedification of the boot2.s file. It's been obsolete.
Simply retire the sedification of the boot2.s file. It's been obsolete
Tue, Mar 28, 7:58 AM

Mon, Mar 27

imp committed rS316064: Fix build with path names with 'align' or 'nop' in them..
Fix build with path names with 'align' or 'nop' in them.
Mon, Mar 27, 10:53 PM

Fri, Mar 24

imp committed rS315918: Add 'device iic' to bring in userland I2C driver..
Add 'device iic' to bring in userland I2C driver.
Fri, Mar 24, 10:33 PM
imp committed rS315901: Use a more stream-lined version of fix_value..
Use a more stream-lined version of fix_value.
Fri, Mar 24, 1:46 PM

Thu, Mar 23

imp closed D6286: Implement quote escaping. String values may now contain " if you it is preceded by \. by committing rS315773: Implement quote escaping. String values may now contain " if you.
Thu, Mar 23, 2:37 AM
imp committed rS315773: Implement quote escaping. String values may now contain " if you.
Implement quote escaping. String values may now contain " if you
Thu, Mar 23, 2:37 AM
imp committed rS315770: Define StrCmp in a funky was to be bug-compatible with EDK2 code..
Define StrCmp in a funky was to be bug-compatible with EDK2 code.
Thu, Mar 23, 2:31 AM
imp committed rS315771: Fix a coverity-discovered NULL pointer dereference..
Fix a coverity-discovered NULL pointer dereference.
Thu, Mar 23, 2:31 AM

Wed, Mar 22

imp committed rS315740: Simplify the code a little..
Simplify the code a little.
Wed, Mar 22, 8:52 PM
imp committed rS315735: Implement moving SD..
Implement moving SD.
Wed, Mar 22, 7:19 PM
imp committed rS315733: Impelemnt ttys onifexists in init..
Impelemnt ttys onifexists in init.
Wed, Mar 22, 7:00 PM
imp closed D10037: Impelemnt ttys onifexists in init. by committing rS315733: Impelemnt ttys onifexists in init..
Wed, Mar 22, 7:00 PM
imp added inline comments to D10085: Explicitly pass ABI flags for MIPS..
Wed, Mar 22, 6:59 PM
imp accepted D10096: usr.sbin.makefs.makefs_cd9660_tests.F_flag fails on a kernel missing 'device cd'.

Same thing 4 times... Maybe a routine would be better?

Wed, Mar 22, 4:44 PM
imp accepted D10085: Explicitly pass ABI flags for MIPS..

I generally like this change. We tried very hard to require the system compiler to do the right thing for two reasons back in the day. First, we wanted to make sure a naked cc could just build things so ports would work better. Second, there were several tricky, hard-coded at the time, bits in different builds that got the wrong ABI tags and so wouldn't link (hack.so?) that were a pita to track down... If the latter has been solved, the former does remain. However, I think it's better that we're explicit since it enables a wider range of compiler to build the base system, and the problem of using those compilers with ports is orthogonal. We shouldn't require it any longer since it is standing in the way of making things better. I regret that we decided to do what we did back in the day since it has held back the external-toolchain work on mips a bit.

Wed, Mar 22, 3:59 AM

Mon, Mar 20

imp accepted D9970: libcam: NULL out freed `ccb.cdm.matches` and `ccb.cdm.patterns` pointers to avoid double frees.

Thanks for making the suggested changes...

Mon, Mar 20, 3:39 PM

Fri, Mar 17

imp added a comment to D10041: Stop building assym.o in modules.

Once upon a time this was needed to create the assym.h file that encoded all the struct sizes and offsets the .S files used. No clue why it is here though....

Fri, Mar 17, 8:22 PM

Thu, Mar 16

imp added inline comments to D10037: Impelemnt ttys onifexists in init..
Thu, Mar 16, 8:50 PM
imp added inline comments to D10037: Impelemnt ttys onifexists in init..
Thu, Mar 16, 8:40 PM
imp created D10037: Impelemnt ttys onifexists in init..
Thu, Mar 16, 8:09 PM

Tue, Mar 14

imp accepted D9996: Fix a memory leak in ccdconfig.
Tue, Mar 14, 4:56 PM
imp accepted D9577: Add the possibility to use another DTC.

Looks like all of Bryan's feedback has been answered, and that answers the questions I had.

Tue, Mar 14, 4:01 PM · ARM

Mon, Mar 13

imp added inline comments to D9970: libcam: NULL out freed `ccb.cdm.matches` and `ccb.cdm.patterns` pointers to avoid double frees.
Mon, Mar 13, 5:13 PM
imp added a comment to D9970: libcam: NULL out freed `ccb.cdm.matches` and `ccb.cdm.patterns` pointers to avoid double frees.
In D9970#206301, @ngie wrote:
In D9970#206077, @imp wrote:

I like setting persistent things to NULL after free, but setting local variables to NULL after free on the 'exit' path of the function is actually harmful: Either the compiler will generate code that wastes cycles in what could be a critical path, or it will optimize it away and maybe issue a warning.... Neither outcome is good, and there's no benefit to doing this.

I forgot to reply to this portion of the comment: could you please provide an example of this? I know that running free/munmap on a large collection of memory pages can eat up a lot of time unnecessarily on exit, but I'm not quite sure I understand the negative effects you're mentioning above.

Mon, Mar 13, 4:46 PM
imp accepted D9968: lib/libcam/cam_cdbparse.3: fix manpage warnings.

Generally looks good, modulo some of the text. But it's not a text cleanup, per se, so I'm good with this going in.

Mon, Mar 13, 4:32 PM

Sun, Mar 12

imp committed rS315175: Convert gnu to using SRCTOP.
Convert gnu to using SRCTOP
Sun, Mar 12, 6:59 PM
imp committed rS315174: Make rescue use SRCTOP.
Make rescue use SRCTOP
Sun, Mar 12, 6:59 PM
imp committed rS315172: Convert include over to SRCTOP.
Convert include over to SRCTOP
Sun, Mar 12, 6:59 PM
imp committed rS315173: Fix two CURDIR references in comments that should be SRCTOP.
Fix two CURDIR references in comments that should be SRCTOP
Sun, Mar 12, 6:59 PM
imp committed rS315171: Move /etc/ to SRCTOP.
Move /etc/ to SRCTOP
Sun, Mar 12, 6:59 PM
imp committed rS315170: Adopt SRCTOP in usr.bin.
Adopt SRCTOP in usr.bin
Sun, Mar 12, 6:59 PM
imp closed D9932: Prefer ${SRCTOP}/foo over ${.CURDIR}/../../foo and ${SRCTOP}/usr.bin/foo over ${.CURDIR}/../foo for paths in Makefiles. by committing rS315170: Adopt SRCTOP in usr.bin.
Sun, Mar 12, 6:59 PM
imp committed rS315166: Export the actual LID state via sysctl. This allows more complex.
Export the actual LID state via sysctl. This allows more complex
Sun, Mar 12, 6:00 PM
imp added a comment to D9970: libcam: NULL out freed `ccb.cdm.matches` and `ccb.cdm.patterns` pointers to avoid double frees.

I like setting persistent things to NULL after free, but setting local variables to NULL after free on the 'exit' path of the function is actually harmful: Either the compiler will generate code that wastes cycles in what could be a critical path, or it will optimize it away and maybe issue a warning.... Neither outcome is good, and there's no benefit to doing this.

Sun, Mar 12, 4:37 PM
imp added a comment to D9969: lib/libcam/cam.3: note that cam_freeccb(3) with ccb == NULL is a no-op.

Forget what I said about anything being missing, other than my ability to read for comprehension.

Sun, Mar 12, 4:32 PM
imp accepted D9969: lib/libcam/cam.3: note that cam_freeccb(3) with ccb == NULL is a no-op.

Note that the text doesn't match the description. The description is missing the word 'NULL' I think.

Sun, Mar 12, 4:31 PM

Fri, Mar 10

imp added a comment to D9915: Extensive use of reallocarray(3) in userland..

I'm agnostic to the change.
I get why they are doing malloc(a * b) -> reallocarray(NULL, a, b), but wonder if it would be better to have a wider API that's mallocarray(a,b) if you are looking for syntactic goodness along with your overflow protection....

Fri, Mar 10, 5:06 PM
imp updated the diff for D9932: Prefer ${SRCTOP}/foo over ${.CURDIR}/../../foo and ${SRCTOP}/usr.bin/foo over ${.CURDIR}/../foo for paths in Makefiles..
  • Move /etc/ to SRCTOP
  • Convert include over to SRCTOP
  • Fix two CURDIR references in comments that should be SRCTOP
  • Make rescue use SRCTOP
  • Fold with usr.bin
  • Convert gnu/lib to using SRCTOP
  • gnu/usr.bin SRCTOP
Fri, Mar 10, 6:25 AM
imp added a comment to D9932: Prefer ${SRCTOP}/foo over ${.CURDIR}/../../foo and ${SRCTOP}/usr.bin/foo over ${.CURDIR}/../foo for paths in Makefiles..

I think I overstate it with "everywhere: I think "in selected places" might be a better term. We use the gmake equivalent $(abspath $(somepath)) a lot and it makes things a lot more readable. As for the speed of it, whatever gmake does doesnt seem to take so long as I've never noticed a slowdown.

Fri, Mar 10, 4:35 AM
imp added a comment to D9932: Prefer ${SRCTOP}/foo over ${.CURDIR}/../../foo and ${SRCTOP}/usr.bin/foo over ${.CURDIR}/../foo for paths in Makefiles..

just my $0.02

Fri, Mar 10, 3:04 AM

Thu, Mar 9

imp updated the summary of D9932: Prefer ${SRCTOP}/foo over ${.CURDIR}/../../foo and ${SRCTOP}/usr.bin/foo over ${.CURDIR}/../foo for paths in Makefiles..
Thu, Mar 9, 6:06 PM
imp created D9932: Prefer ${SRCTOP}/foo over ${.CURDIR}/../../foo and ${SRCTOP}/usr.bin/foo over ${.CURDIR}/../foo for paths in Makefiles..
Thu, Mar 9, 6:05 PM
imp accepted D9927: Fix typo in nvme(4).
Thu, Mar 9, 4:56 AM
imp accepted D9907: Port i.MX6 to PLATFORM_SMP.

This looks good to my eye. One nit.

Thu, Mar 9, 4:55 AM
imp committed rS314928: efidp manipulates UEFI Device Paths in various ways. At the moment, it.
efidp manipulates UEFI Device Paths in various ways. At the moment, it
Thu, Mar 9, 12:33 AM
imp committed rS314926: Bring in EDK2 routines for printing and parsing device paths..
Bring in EDK2 routines for printing and parsing device paths.
Thu, Mar 9, 12:31 AM
imp committed rS314927: Finish implementing -d/--device/--device-path flag to print variable.
Finish implementing -d/--device/--device-path flag to print variable
Thu, Mar 9, 12:31 AM
imp committed rS314925: Share UCS2/UTF8 routines between boot loader and userland..
Share UCS2/UTF8 routines between boot loader and userland.
Thu, Mar 9, 12:24 AM

Wed, Mar 8

imp added a comment to D7149: Fix gcc warnings.

Peter: This seems like a common case (idiom) that we should maybe solve without a huge number of #ifdef INVARIANTS, maybe by a new macro that does the right thing in the !INVARIANTS case?

Wed, Mar 8, 2:41 PM
imp committed rS314890: Copy needed include files from EDK2. This is a minimal set gleened.
Copy needed include files from EDK2. This is a minimal set gleened
Wed, Mar 8, 2:48 AM

Tue, Mar 7

imp committed rS314889: Avoid dereferencing unintialized elements in the error path..
Avoid dereferencing unintialized elements in the error path.
Tue, Mar 7, 11:06 PM
imp committed rS314888: cwd10 takes the low 32-bits and cwd11 takes the upper 32-bits of the.
cwd10 takes the low 32-bits and cwd11 takes the upper 32-bits of the
Tue, Mar 7, 11:03 PM
imp committed rS314884: Make multi-namespace nvme drives more robust..
Make multi-namespace nvme drives more robust.
Tue, Mar 7, 9:48 PM
imp committed rS314881: Record Tianocore EDK2 at 7babb4372e6a34cbbc54249b25056272a5a9924c from.
Record Tianocore EDK2 at 7babb4372e6a34cbbc54249b25056272a5a9924c from
Tue, Mar 7, 9:00 PM
imp committed rS314880: Move down to required dist directory for vendor tracking..
Move down to required dist directory for vendor tracking.
Tue, Mar 7, 8:58 PM
imp committed rS314879: Bring in snapshot of the MdePkg from Tianocore's EDK2 project at git.
Bring in snapshot of the MdePkg from Tianocore's EDK2 project at git
Tue, Mar 7, 8:54 PM
imp accepted D9829: Add a summary line of the ZFS Compressed ARC to top(1).
Tue, Mar 7, 4:47 PM

Sun, Mar 5

imp added a comment to D9903: Add C11 Appendix K function memset_s.

Yea, the _s functions are MS' NIH for other interfaces that had pre-dated their crazy by a lot.

Sun, Mar 5, 7:45 PM
imp added a comment to D9903: Add C11 Appendix K function memset_s.

The code and the tests look good.
There's no man page.

Sun, Mar 5, 7:27 PM

Fri, Mar 3

imp accepted D9879: Drop i486 from the default GENERIC i386 kernel..

This is fine, though the commit message could use some work.

Fri, Mar 3, 10:25 PM
imp committed rS314623: Make sure guid table is compiled before we use it..
Make sure guid table is compiled before we use it.
Fri, Mar 3, 8:23 PM
imp committed rS314621: If the guid specified can't be decoded as a GUID, try looking it up in.
If the guid specified can't be decoded as a GUID, try looking it up in
Fri, Mar 3, 8:23 PM
imp committed rS314619: Implement --guid/-g to print the known GUIDs as human readable. The.
Implement --guid/-g to print the known GUIDs as human readable. The
Fri, Mar 3, 8:23 PM
imp committed rS314618: Actually implement efivar -L.
Actually implement efivar -L
Fri, Mar 3, 8:23 PM
imp committed rS314620: Implement -d / --device-path to print the ascii representation of a variable….
Implement -d / --device-path to print the ascii representation of a variable…
Fri, Mar 3, 8:23 PM
imp committed rS314617: Print just the variables when we're not printing the values..
Print just the variables when we're not printing the values.
Fri, Mar 3, 8:23 PM
imp committed rS314615: Move uuid_table definition to efivar.h..
Move uuid_table definition to efivar.h.
Fri, Mar 3, 8:23 PM
imp committed rS314616: Only compile the known uuid table once..
Only compile the known uuid table once.
Fri, Mar 3, 8:23 PM

Wed, Mar 1

imp accepted D9841: Tidy up machine dependent Makefiles in libc..

lgtm

Wed, Mar 1, 2:40 PM
imp committed rS314481: Add kernel support for the TP-LINK MR3040..
Add kernel support for the TP-LINK MR3040.
Wed, Mar 1, 5:54 AM
imp committed rS314480: Back out r314471. In https://reviews.freebsd.org/D1858 it was clear.
Back out r314471. In https://reviews.freebsd.org/D1858 it was clear
Wed, Mar 1, 5:38 AM
imp committed rS314477: Fixed TCO v3 tick convert.
Fixed TCO v3 tick convert
Wed, Mar 1, 5:19 AM
imp committed rS314473: Create README.md file for viewing on github..
Create README.md file for viewing on github.
Wed, Mar 1, 5:05 AM
imp committed rS314471: Fix VNET - DAD detected duplicate IPv6 address.
Fix VNET - DAD detected duplicate IPv6 address
Wed, Mar 1, 4:47 AM
imp committed rS314469: Update tw_cl_share.h to allow 255 LUNs.
Update tw_cl_share.h to allow 255 LUNs
Wed, Mar 1, 4:41 AM
imp committed rS314467: Make wsp process a single touchpad tap and interpret it as a.
Make wsp process a single touchpad tap and interpret it as a
Wed, Mar 1, 4:28 AM
imp committed rS314449: Revert prior commit to restore the files mangled by my "fixing" merge.
Revert prior commit to restore the files mangled by my "fixing" merge
Wed, Mar 1, 2:11 AM
imp committed rS314448: strstr.c was inadvertently blasted with a copy of isa_nmi.c. Revert.
strstr.c was inadvertently blasted with a copy of isa_nmi.c. Revert
Wed, Mar 1, 2:08 AM

Feb 28 2017

imp committed rS314436: Renumber copyright clause 4.
Renumber copyright clause 4
Feb 28 2017, 11:44 PM
imp committed rS314428: Add Ubiquiti Rocket M support.
Add Ubiquiti Rocket M support
Feb 28 2017, 10:50 PM
imp committed rS314426: This should have been K3771_INIT..
This should have been K3771_INIT.
Feb 28 2017, 10:23 PM
imp committed rS314423: Fix arge0 mdio bus.
Fix arge0 mdio bus
Feb 28 2017, 9:39 PM
imp committed rS314422: Update PICOSTATION_M2HP.
Update PICOSTATION_M2HP
Feb 28 2017, 9:39 PM
imp committed rS314408: Fix include/arpa/nameser_compat.h by adding T_DNAME definition..
Fix include/arpa/nameser_compat.h by adding T_DNAME definition.
Feb 28 2017, 8:34 PM
imp committed rS314407: Add support for Vodafone/Huawei K3771..
Add support for Vodafone/Huawei K3771.
Feb 28 2017, 8:25 PM

Feb 27 2017

imp accepted D9601: Expand sbufs to be able to output directly to the log+console interface.

I can't think of a better name....

Feb 27 2017, 11:22 PM
imp added inline comments to D9828: Fix a panic during boot while executing the vt_timer() callout.
Feb 27 2017, 10:28 PM

Feb 26 2017

imp committed rS314317: Include pcib_private.h for prototypes..
Include pcib_private.h for prototypes.
Feb 26 2017, 9:33 PM
imp added inline comments to D6286: Implement quote escaping. String values may now contain " if you it is preceded by \..
Feb 26 2017, 7:31 AM
imp retitled D9803: Don't call _OSC for a bit we've previously successfully acquired. from to Don't call _OSC for a bit we've previously successfully acquired..
Feb 26 2017, 7:17 AM
imp updated the diff for D6286: Implement quote escaping. String values may now contain " if you it is preceded by \..

Fix issues raised in the review.

Feb 26 2017, 7:03 AM
imp abandoned D8263: Start u-boot framework.

This has been committed, but in different form.

Feb 26 2017, 6:35 AM

Feb 25 2017

imp committed rS314283: Move inclusion of opt_printf.h around so that we can compile all the.
Move inclusion of opt_printf.h around so that we can compile all the
Feb 25 2017, 10:11 PM
imp closed D8285: Add seatbelt for upgrades.

This was comitted like forever ago.

Feb 25 2017, 9:34 PM
imp closed D9717: make PCI_HP use pcib_request_feature.

Date: Sat Feb 25 06:11:59 2017
New Revision: 314250
URL: https://svnweb.freebsd.org/changeset/base/314250

Feb 25 2017, 4:11 PM
imp accepted D9717: make PCI_HP use pcib_request_feature.
Feb 25 2017, 4:11 PM
imp added a comment to D9717: make PCI_HP use pcib_request_feature.

Made suggested changes.

Feb 25 2017, 4:10 PM
imp closed D9716: Rename pci_pcie_intr to pcie_pcie_intr_hotplug since it only handles hotplug events.

Log:

Rename pci_pcie_intr to pci_pcie_intr_hotplug.
Feb 25 2017, 4:08 PM