jhb (John Baldwin)
User

Projects (6)

User Details

User Since
Mar 11 2014, 8:46 PM (192 w, 5 d)

Recent Activity

Thu, Nov 16

jhb committed rS325902: MFC 319517: Add a cross-reference to sysdecode_socket_protocol(3)..
MFC 319517: Add a cross-reference to sysdecode_socket_protocol(3).
Thu, Nov 16, 7:07 PM
jhb committed rS325900: MFC 325039: Rework pass through changes in r305485 to be safer..
MFC 325039: Rework pass through changes in r305485 to be safer.
Thu, Nov 16, 6:22 PM
jhb committed rS325885: MFC 324993:.
MFC 324993:
Thu, Nov 16, 6:56 AM

Wed, Nov 15

jhb committed rS325856: MFC 323588: Recognize NT_PTLWPINFO and NT_ARM_VFP in FreeBSD ELF cores..
MFC 323588: Recognize NT_PTLWPINFO and NT_ARM_VFP in FreeBSD ELF cores.
Wed, Nov 15, 3:24 PM
jhb committed rS325837: MFC 323584: Add a NT_ARM_VFP ELF core note to hold VFP registers for each….
MFC 323584: Add a NT_ARM_VFP ELF core note to hold VFP registers for each…
Wed, Nov 15, 2:04 AM

Tue, Nov 14

jhb committed rS325835: Use #if instead of #ifdef for __BSD_VISIBLE tests..
Use #if instead of #ifdef for __BSD_VISIBLE tests.
Tue, Nov 14, 11:50 PM
jhb closed D12977: Use #if instead of #ifdef for __BSD_VISIBLE tests..
Tue, Nov 14, 11:50 PM
jhb committed rS325831: MFC 323581,323582,323583: Add ptrace operations for VFP registers..
MFC 323581,323582,323583: Add ptrace operations for VFP registers.
Tue, Nov 14, 10:18 PM
jhb committed rS325816: Appease old GCC by disabling .cfi_sections for GCC 4.x..
Appease old GCC by disabling .cfi_sections for GCC 4.x.
Tue, Nov 14, 5:16 PM
jhb committed rS325810: MFC 323580,323933,323934,324814,324817: Enable AT_HWCAP on arm..
MFC 323580,323933,323934,324814,324817: Enable AT_HWCAP on arm.
Tue, Nov 14, 4:03 PM

Mon, Nov 13

jhb committed rD51199: Promote FreeBSD/mips from Tier-3 to Tier-2..
Promote FreeBSD/mips from Tier-3 to Tier-2.
Mon, Nov 13, 9:41 PM
jhb committed rS325778: Move loop to clear TDB_SUSPEND into PT_DETACH case..
Move loop to clear TDB_SUSPEND into PT_DETACH case.
Mon, Nov 13, 9:22 PM
jhb added a comment to D12927: Only pass -F flag to linker if it is supported.

If you get a build error now that is probably just as well as a .error. I'd be tempted to avoid adding a dedicated .error for an odd edge case if the build already fails.

Mon, Nov 13, 9:20 PM
jhb added a comment to D13042: Add two new tunables / sysctls to controll reboot after panic.
In D13042#271835, @imp wrote:
In D13042#271831, @jhb wrote:

This is probably fine. You could also perhaps have a variant of kib@'s suggestion that takes a string: kern.panic_shutdown = {"reboot", "poweroff", "power-cycle"}. Internally you would store it as a simple int / enum, but use a SYSCTL_PROC to map strings to the internal values. This also means you can't set both RB_POWEROFF and RB_POWERCYCLE (which the current patch lets you do) for which the behavior isn't clear (which one has precedence, etc.?)

As implemented, POWERCYCLE takes precedence over POWEROFF which takes precedence over HALT which takes precedence over AUTOBOOT, unless autoboot fails, in which case we halt. It isn't documented that this should be the case, and I only implemented it this way because I wanted the ability to have the thing that power cycles fail and then have it reboot or possibly power off. Though the system call can handle all these possibilities, I'm not sure the variations on halt/reboot/shutdown can specify this much nuance (though the fallback to reboot or halt is possible, not sure that you can power cycle with a fallback to power off.

Having said that, I'm cool with an enum thing, but didn't recall that SYSCTL_PROC settings actually could be tuneables. That would make it easier to use, perhaps (assuming we spell power-off and power-cycle either both with or both without - and allow both forms), but this is a rare enough case I worry about too much polish getting in the way. So I'm inclined to push back a little to see how important you think your proposed interface is....

Mon, Nov 13, 9:17 PM
jhb closed D13073: Pull the PT_ATTACH case out of the 'sendsig:' block..
Mon, Nov 13, 9:09 PM
jhb committed rS325777: Pull the PT_ATTACH case out of the 'sendsig:' block..
Pull the PT_ATTACH case out of the 'sendsig:' block.
Mon, Nov 13, 9:09 PM
jhb created D13073: Pull the PT_ATTACH case out of the 'sendsig:' block..
Mon, Nov 13, 8:14 PM
jhb committed rS325771: Only clear a pending thread event if one is pending..
Only clear a pending thread event if one is pending.
Mon, Nov 13, 7:59 PM
jhb closed D12837: Only clear a pending thread event if one is pending..
Mon, Nov 13, 7:59 PM
jhb added inline comments to D12837: Only clear a pending thread event if one is pending..
Mon, Nov 13, 7:51 PM
jhb added a comment to D12837: Only clear a pending thread event if one is pending..

I will probably try to "uninline" the PT_ATTACH case as a followup after this. After the last bits of refactoring I think it will probably be worth it.

Mon, Nov 13, 7:22 PM
jhb updated the diff for D12837: Only clear a pending thread event if one is pending..
  • Consolidate kern_psignal cases for PT_ATTACH.
Mon, Nov 13, 7:12 PM
jhb updated the diff for D12837: Only clear a pending thread event if one is pending..
  • Move P_STOPPED_TRACE section up.
  • Require P_STOPPED_TRACE for most actions.
  • Clarify some comments.
Mon, Nov 13, 7:00 PM
jhb added a comment to D13042: Add two new tunables / sysctls to controll reboot after panic.

This is probably fine. You could also perhaps have a variant of kib@'s suggestion that takes a string: kern.panic_shutdown = {"reboot", "poweroff", "power-cycle"}. Internally you would store it as a simple int / enum, but use a SYSCTL_PROC to map strings to the internal values. This also means you can't set both RB_POWEROFF and RB_POWERCYCLE (which the current patch lets you do) for which the behavior isn't clear (which one has precedence, etc.?)

Mon, Nov 13, 6:53 PM
jhb accepted D12925: Only build the libmagic build-tools if MK_FILE != no.
Mon, Nov 13, 6:48 PM
jhb added a comment to D12928: Fix new compiler warnings in binutils.

I would probably be inclined to just add -Wno-foo to WARNFLAGS.clang with a suitable compiler version check instead if this generates equivalent code for non-CHERI. (For CHERI I think we can just always build ld.bfd as a plain MIPS binary while we still need ld.bfd).

Mon, Nov 13, 6:45 PM
jhb added a comment to D12927: Only pass -F flag to linker if it is supported.

Hmm, lib/Makefile only recurses into libdl if the linker supports filters (that was in the original commit that added libdl: rS320872), so this shouldn't be needed? The libdl library only works as a filter, so building a non-filter version will just build a broken one I think.

Mon, Nov 13, 6:40 PM
jhb added a comment to D12926: Don't build share/syscons in build-tools stage if MK_SYSCONS == "no".

Most likely this is correct, just adding Bryan for a sanity check review.

Mon, Nov 13, 6:35 PM
jhb added a reviewer for D12926: Don't build share/syscons in build-tools stage if MK_SYSCONS == "no": bdrewery.
Mon, Nov 13, 6:35 PM
jhb added a comment to D12916: Allow setting the initial LDFLAGS for the cross environment with XLDFLAGS.

FWIW, I used -Wno-unused-command-line-argument in XCFLAGS as my workaround for using -fuse-ld. I'll defer to Bryan on how this should work. One question I have is does this DTRT if a Makefile modifies LDFLAGS? I almost wonder if what we want is to add CCLD and CXXLD variables that are used instead of CC and CXX when linking libraries and programs and have CCLD set to "XCC XCFLAGS XLDFLAGS" and CXXLD set to "XCXX XCXXFLAGS XLDFLAGS" in CROSSENV? (CCLD would default to CC and CXXLD would default to CXX)

Mon, Nov 13, 6:34 PM
jhb added a comment to D12912: Don't fail the build due to clang integer constant range warnings.

Adding dim@ who normally manages updates to llvm in the base. Dimitry, Alex is building CheriBSD with a top-of-tree-ish clang. I'm not sure if you are currently building FreeBSD with clang 6 and have hit the same build failures already? Does disabling errors for this warning seem sensible to you?

Mon, Nov 13, 6:28 PM
jhb added a reviewer for D12912: Don't fail the build due to clang integer constant range warnings: dim.
Mon, Nov 13, 6:26 PM
jhb added inline comments to D12837: Only clear a pending thread event if one is pending..
Mon, Nov 13, 6:25 PM
jhb accepted D12910: Remove -fuse-ld= when creating _LDFLAGS from LDFLAGS.
Mon, Nov 13, 6:24 PM

Fri, Nov 10

jhb accepted D13015: Axe xlint..

I would love to remove all the '#ifdef lint' bits in the future.

Fri, Nov 10, 7:13 PM
jhb committed rS325624: Some fixups to the CFI directives for PLT stub entry points..
Some fixups to the CFI directives for PLT stub entry points.
Fri, Nov 10, 1:17 AM
jhb committed rS325623: Use NESTED() instead of LEAF() for rtld_start..
Use NESTED() instead of LEAF() for rtld_start.
Fri, Nov 10, 1:14 AM

Wed, Nov 8

jhb accepted D12796: Zero uninitialized part of struct ptrace_lwpinfo..
Wed, Nov 8, 10:10 PM

Tue, Nov 7

jhb committed rS325519: Wrap to 80 columns. No functional change..
Wrap to 80 columns. No functional change.
Tue, Nov 7, 6:11 PM
jhb committed rP453644: MFH: r452019.
MFH: r452019
Tue, Nov 7, 3:52 PM
jhb added inline comments to D12785: New gets_s(3) stdio function.
Tue, Nov 7, 3:51 PM
jhb added inline comments to D12796: Zero uninitialized part of struct ptrace_lwpinfo..
Tue, Nov 7, 3:51 PM
jhb accepted D12279: libkvm: add kvm_walk_pages API..

This should be fine. I think the diff in phab is just a diff relative to the previous rather than the full diff FWIW.

Tue, Nov 7, 3:51 PM
jhb created D12977: Use #if instead of #ifdef for __BSD_VISIBLE tests..
Tue, Nov 7, 3:51 PM

Mon, Oct 30

jhb created D12837: Only clear a pending thread event if one is pending..
Mon, Oct 30, 5:24 PM

Fri, Oct 27

jhb closed D12661: Rework pass through changes in r305485 to be safer..
Fri, Oct 27, 2:57 PM
jhb committed rS325039: Rework pass through changes in r305485 to be safer..
Rework pass through changes in r305485 to be safer.
Fri, Oct 27, 2:57 PM
jhb committed rD51149: Document FreeBSD versions 1200044 through 1200052..
Document FreeBSD versions 1200044 through 1200052.
Fri, Oct 27, 4:02 AM
jhb committed rS325028: Discard the correct thread event reported for a ptrace stop..
Discard the correct thread event reported for a ptrace stop.
Fri, Oct 27, 3:16 AM
jhb closed D12794: Discard the correct thread event reported for a ptrace stop..
Fri, Oct 27, 3:16 AM

Thu, Oct 26

jhb added inline comments to D12794: Discard the correct thread event reported for a ptrace stop..
Thu, Oct 26, 4:21 PM
jhb updated the diff for D12794: Discard the correct thread event reported for a ptrace stop..
  • Remove debugging.
Thu, Oct 26, 2:39 PM
jhb added inline comments to D12794: Discard the correct thread event reported for a ptrace stop..
Thu, Oct 26, 2:24 PM
jhb accepted D12765: newfs: warn if newer than kernel.
Thu, Oct 26, 10:44 AM
jhb created D12794: Discard the correct thread event reported for a ptrace stop..
Thu, Oct 26, 9:58 AM

Wed, Oct 25

jhb committed rS324993: Add a test for sending a signal while stepping a thread via PT_STEP..
Add a test for sending a signal while stepping a thread via PT_STEP.
Wed, Oct 25, 5:23 PM

Oct 18 2017

jhb created D12724: Backport wpa fixes to stable/10..
Oct 18 2017, 9:16 PM
jhb committed rS324724: Remove CPU_HAVEFPU..
Remove CPU_HAVEFPU.
Oct 18 2017, 5:23 PM
jhb closed D12707: Remove CPU_HAVEFPU..
Oct 18 2017, 5:23 PM
jhb accepted D10513: Remove mbpool(9) now that it has no consumers..
Oct 18 2017, 12:11 AM

Oct 17 2017

jhb created D12707: Remove CPU_HAVEFPU..
Oct 17 2017, 10:25 PM
jhb committed rS324687: MFC 323579,323585: Add AT_HWCAP and AT_EHDRFLAGS on all platforms..
MFC 323579,323585: Add AT_HWCAP and AT_EHDRFLAGS on all platforms.
Oct 17 2017, 12:46 PM

Oct 16 2017

jhb accepted D12684: gdb kernel server: fixup Search:memory style.
Oct 16 2017, 6:20 PM

Oct 13 2017

jhb committed rS324612: MFC 324039: Don't defer wakeup()s for completed journal workitems..
MFC 324039: Don't defer wakeup()s for completed journal workitems.
Oct 13 2017, 10:41 PM
jhb committed rS324611: MFC 324073: Use UMA_ALIGNOF() for name cache UMA zones..
MFC 324073: Use UMA_ALIGNOF() for name cache UMA zones.
Oct 13 2017, 9:58 PM
jhb committed rS324602: MFC 324072: Add UMA_ALIGNOF()..
MFC 324072: Add UMA_ALIGNOF().
Oct 13 2017, 5:11 PM
jhb created D12661: Rework pass through changes in r305485 to be safer..
Oct 13 2017, 4:46 PM
jhb committed rP452019: Merge in updates for mipsn32 and 32-bit arm..
Merge in updates for mipsn32 and 32-bit arm.
Oct 13 2017, 4:32 PM
jhb closed D12630: Merge in updates for mipsn32 and 32-bit arm..
Oct 13 2017, 4:32 PM

Oct 10 2017

jhb updated the diff for D12630: Merge in updates for mipsn32 and 32-bit arm..
  • Use older regcache functions for 8.0.1.
Oct 10 2017, 8:34 PM

Oct 9 2017

jhb created D12630: Merge in updates for mipsn32 and 32-bit arm..
Oct 9 2017, 6:09 PM

Oct 4 2017

jhb added inline comments to D12572: Various pathconf() fixes..
Oct 4 2017, 8:36 PM
jhb committed rD51043: Note that pc98 is transitioning to Tier-4 in FreeBSD 12..
Note that pc98 is transitioning to Tier-4 in FreeBSD 12.
Oct 4 2017, 8:18 PM
jhb committed rD51042: Tweak Tier-1 description to not require Tinderbox by name..
Tweak Tier-1 description to not require Tinderbox by name.
Oct 4 2017, 8:12 PM

Oct 3 2017

jhb accepted D12571: Correct sense of crypt(3) NULL checks in init(8) and lock(1).
Oct 3 2017, 12:30 AM
jhb created D12572: Various pathconf() fixes..
Oct 3 2017, 12:03 AM

Oct 2 2017

jhb committed rS324228: Flesh out pathconf() on UDF..
Flesh out pathconf() on UDF.
Oct 2 2017, 11:31 PM
jhb committed rS324226: Return 64 for pathconf(_PC_FILESIZEBITS) on tmpfs..
Return 64 for pathconf(_PC_FILESIZEBITS) on tmpfs.
Oct 2 2017, 11:23 PM
jhb committed rS324224: Handle _PC_FILESIZEBITS and _PC_SYMLINK_MAX pathconf() requests in cd9660..
Handle _PC_FILESIZEBITS and _PC_SYMLINK_MAX pathconf() requests in cd9660.
Oct 2 2017, 11:12 PM
jhb added inline comments to D12569: Updates to Tiers..
Oct 2 2017, 7:24 PM
jhb committed rS324215: MFC 323994: Log signal number passed to PT_STEP requests in KTR_PTRACE traces..
MFC 323994: Log signal number passed to PT_STEP requests in KTR_PTRACE traces.
Oct 2 2017, 6:04 PM
jhb added inline comments to D12279: libkvm: add kvm_walk_pages API..
Oct 2 2017, 5:49 PM
jhb created D12569: Updates to Tiers..
Oct 2 2017, 5:34 PM

Sep 30 2017

jhb committed rS324124: MFC 323631: Add an -a flag to getconf..
MFC 323631: Add an -a flag to getconf.
Sep 30 2017, 5:30 PM

Sep 29 2017

jhb added inline comments to D12279: libkvm: add kvm_walk_pages API..
Sep 29 2017, 9:07 PM

Sep 28 2017

jhb accepted D12525: devel/gdb: Updating to gdb 8.0.1.
Sep 28 2017, 6:30 PM

Sep 27 2017

jhb committed rS324073: Use UMA_ALIGNOF() for name cache UMA zones..
Use UMA_ALIGNOF() for name cache UMA zones.
Sep 27 2017, 11:19 PM
jhb committed rS324072: Add UMA_ALIGNOF()..
Add UMA_ALIGNOF().
Sep 27 2017, 11:15 PM

Sep 26 2017

jhb closed D12494: Don't defer wakeup()'s for completed journal workitems..
Sep 26 2017, 11:24 PM
jhb committed rS324039: Don't defer wakeup()s for completed journal workitems..
Don't defer wakeup()s for completed journal workitems.
Sep 26 2017, 11:24 PM
jhb accepted D12452: aesni(4): Add support for x86 SHA intrinsics.
Sep 26 2017, 10:07 PM

Sep 25 2017

jhb added inline comments to D12452: aesni(4): Add support for x86 SHA intrinsics.
Sep 25 2017, 8:54 PM
jhb committed rS323994: Log signal number passed to PT_STEP requests in KTR_PTRACE traces..
Log signal number passed to PT_STEP requests in KTR_PTRACE traces.
Sep 25 2017, 8:39 PM
jhb committed rS323993: Use tmpfs_print for tmpfs FIFOs..
Use tmpfs_print for tmpfs FIFOs.
Sep 25 2017, 8:26 PM
jhb added a comment to D12452: aesni(4): Add support for x86 SHA intrinsics.

If you'd like an alternate testing tool (one that can test AES-CBC chained with SHA for example), you can try 'cryptocheck' from github/bsdjhb/freebsd.git on the 'cryptocheck' branch. You can also run IPSec across a session using the ciphers for testing but that requires a bit more setup.

Sep 25 2017, 8:08 PM
jhb accepted D12451: crypto(9): Use a more specific error code when a capable driver is not found.

Thanks

Sep 25 2017, 8:02 PM
jhb accepted D12437: opencrypto: Loosen restriction on HMAC key sizes.
Sep 25 2017, 8:01 PM
jhb accepted D12120: Remove artificial restriction on lio_listio's operation count.
Sep 25 2017, 8:00 PM
jhb accepted D12437: opencrypto: Loosen restriction on HMAC key sizes.
In D12437#258495, @cem wrote:
In D12437#258465, @jhb wrote:

Unfortunately we can't just assume 'blocksize' because some algorithms (keyed SHA1, NULL, etc.) have 0 for the keysize. I just think 'max_keysize' is a bit of a misnomer because I consider the 'auth_xform' structures to provide values related to "what the algorithm uses" in which case 'keysize' is a single fixed-length field much like 'blocksize'. Handling "user provided keys" of different sizes is something that needs to happen at a slightly higher level similar to how one uses PKCS padding to deal with data that isn't an even multiple of a block. It's not the end of the world though. I looked and OpenSSL doesn't provide anything like 'keysize' in its EVP_MD structures, and HMAC() requires a key and assumes a block size (so for OpenSSL it seems you have to provide a bogus key for NULL HMAC which seems odd)

Ok. Would you like me to rename max_keysize back to keysize? Are there any other changes you'd like to see before this goes in?

Sep 25 2017, 6:33 PM
jhb created D12494: Don't defer wakeup()'s for completed journal workitems..
Sep 25 2017, 6:04 PM