Page MenuHomeFreeBSD

markj (Mark Johnston)
User

Projects

User Details

User Since
Mar 12 2014, 1:00 AM (270 w, 6 d)

Recent Activity

Today

markj added inline comments to D20300: Introducing pdopen..
Tue, May 21, 4:07 AM
markj committed rS348027: MFC r347570:.
MFC r347570:
Tue, May 21, 1:27 AM
markj added a reviewer for D20327: Don't reset memory attributes when mapping physical addresses for ACPI.: alc.
Tue, May 21, 1:00 AM

Yesterday

markj committed rS348011: Hook DEFINE_IFUNC.9 up to the build..
Hook DEFINE_IFUNC.9 up to the build.
Mon, May 20, 9:23 PM
markj committed rS348004: kcov depends on eventhandler.h..
kcov depends on eventhandler.h.
Mon, May 20, 7:14 PM
markj committed rS348003: Add a man page for DEFINE_IFUNC..
Add a man page for DEFINE_IFUNC.
Mon, May 20, 7:12 PM
markj closed D20310: Add a DEFINE_IFUNC man page..
Mon, May 20, 7:12 PM
markj committed rS348002: Typo..
Typo.
Mon, May 20, 7:09 PM
markj added a reviewer for D20317: Netdump: coalesce writes less than buffer size for more efficient use of network: cem.
Mon, May 20, 6:42 PM
markj added a comment to D20317: Netdump: coalesce writes less than buffer size for more efficient use of network.

I think this is probably fine - we are lying to the upper layers about having successfully written a block when in fact we're only caching it, but that shouldn't cause problems. You might consider doing this in the minidump code instead, since the coalescing will help disk dumps as well. But I don't mind having this be netdump-specific for now.

Mon, May 20, 6:42 PM

Sun, May 19

markj accepted D20316: ksyms: Fixup symbols for powerpc in the kernel, not just modules.
Sun, May 19, 6:07 PM
markj added inline comments to D20316: ksyms: Fixup symbols for powerpc in the kernel, not just modules.
Sun, May 19, 3:25 PM
markj updated the diff for D20310: Add a DEFINE_IFUNC man page..

Remove unintended line break.

Sun, May 19, 3:20 PM
markj added inline comments to D20316: ksyms: Fixup symbols for powerpc in the kernel, not just modules.
Sun, May 19, 3:08 PM

Sat, May 18

markj accepted D20304: NDFREE(): Fix unlocking for LOCKPARENT|LOCKLEAF and ndp->ni_dvp == ndp->ni_vp.
Sat, May 18, 9:51 PM
markj updated the diff for D20310: Add a DEFINE_IFUNC man page..

Try to make the wording around relocations more generic.

Sat, May 18, 9:44 PM
markj committed rP501981: Update to netdumpd-20190518..
Update to netdumpd-20190518.
Sat, May 18, 9:38 PM
markj added inline comments to D20310: Add a DEFINE_IFUNC man page..
Sat, May 18, 9:08 PM
markj updated the diff for D20310: Add a DEFINE_IFUNC man page..

Address feedback.

Sat, May 18, 9:07 PM
markj updated the diff for D20310: Add a DEFINE_IFUNC man page..

Typo.

Sat, May 18, 8:16 PM
markj updated the diff for D20310: Add a DEFINE_IFUNC man page..

Address feedback.

Sat, May 18, 8:15 PM
markj added a comment to D20310: Add a DEFINE_IFUNC man page..
In D20310#437745, @kib wrote:

Why do we need this page ?

Sat, May 18, 8:15 PM
markj added reviewers for D20310: Add a DEFINE_IFUNC man page.: emaste, kib, manpages.
Sat, May 18, 6:45 PM
markj created D20310: Add a DEFINE_IFUNC man page..
Sat, May 18, 6:45 PM
markj added a comment to D20304: NDFREE(): Fix unlocking for LOCKPARENT|LOCKLEAF and ndp->ni_dvp == ndp->ni_vp.

So if ni_dvp == ni_vp, does the caller hold two references on vp or one? It was not clear to me from some quick reading of the code.

Sat, May 18, 6:04 PM
markj committed rS347955: Update the DIAGNOSTIC-only vmem_check_sanity() after r347949..
Update the DIAGNOSTIC-only vmem_check_sanity() after r347949.
Sat, May 18, 2:19 PM
markj committed rS347950: Use M_NEXTFIT in memguard(9)..
Use M_NEXTFIT in memguard(9).
Sat, May 18, 2:02 AM
markj closed D17227: Use M_NEXTFIT in memguard..
Sat, May 18, 2:02 AM
markj committed rS347949: Implement the M_NEXTFIT allocation strategy for vmem(9)..
Implement the M_NEXTFIT allocation strategy for vmem(9).
Sat, May 18, 1:46 AM
markj closed D17226: Implement M_NEXTFIT..
Sat, May 18, 1:46 AM

Fri, May 17

markj committed rS347937: MFC r337716:.
MFC r337716:
Fri, May 17, 6:10 PM
markj committed rS347905: MFC r347569:.
MFC r347569:
Fri, May 17, 2:09 PM
markj committed rS347904: MFC r347564:.
MFC r347564:
Fri, May 17, 2:09 PM

Thu, May 16

markj committed rS347700: MFC r337715, r337751, r337754, r337758, r337813, r338354, r338687,.
MFC r337715, r337751, r337754, r337758, r337813, r338354, r338687,
Thu, May 16, 2:42 PM
markj accepted D20256: Simplify vm_reserv_break.
Thu, May 16, 1:30 PM
markj committed rS347691: Revert r347582 for now..
Revert r347582 for now.
Thu, May 16, 1:04 PM
markj accepted D20273: db show thread: avoid overlow in tick conversion.
Thu, May 16, 5:17 AM

Wed, May 15

markj accepted D19630: amd64 pmap: rework di removing global mutex.
Wed, May 15, 9:13 PM
markj committed rS347616: MFC r339046, r339073:.
MFC r339046, r339073:
Wed, May 15, 3:49 PM
markj added a comment to D17392: Add arm64 kernel support for -z ifunc-noplt.

I'm fine with breaking from the ABI in the kernel & modules as long as it's just for optimisation and we don't require the breakage to allow the kernel to still work after being built with, for example, gcc & linked with bfd.

Wed, May 15, 1:28 PM
markj added inline comments to D20256: Simplify vm_reserv_break.
Wed, May 15, 12:51 PM

Tue, May 14

markj committed rS347582: Close some races in multicast socket option handling..
Close some races in multicast socket option handling.
Tue, May 14, 9:31 PM
markj closed D20070: Fix mutual exclusion issues in multicast socket option handling..
Tue, May 14, 9:31 PM
markj added inline comments to D20256: Simplify vm_reserv_break.
Tue, May 14, 8:51 PM
markj added inline comments to D19630: amd64 pmap: rework di removing global mutex.
Tue, May 14, 8:16 PM
markj closed D19599: Add a UFS-specific implementation for lseek(SEEK_DATA)..

I committed this in r346932.

Tue, May 14, 7:19 PM
markj committed rS347570: Specify -z notext when building with -z ifunc-noplt..
Specify -z notext when building with -z ifunc-noplt.
Tue, May 14, 6:26 PM
markj closed D20260: Build the kernel with -z notext..
Tue, May 14, 6:26 PM
markj committed rS347569: Remove redundant -Wl uses from the kernel's LDFLAGS..
Remove redundant -Wl uses from the kernel's LDFLAGS.
Tue, May 14, 6:10 PM
markj added a comment to D20260: Build the kernel with -z notext..
In D20260#436661, @kib wrote:

BTW would it be better to use the -Wl,-z,notext -Wl,-z,ifunc-noplt syntax ?

Yes, or even -Wl,-z,notext,-z,ifunc-noplt. But at least the -z and its arg should be in the same -Wl, IMO.

Tue, May 14, 5:44 PM
markj updated the summary of D20260: Build the kernel with -z notext..
Tue, May 14, 3:34 PM
markj created D20260: Build the kernel with -z notext..
Tue, May 14, 3:31 PM
markj committed rS347564: Fix formatting..
Fix formatting.
Tue, May 14, 3:20 PM
markj updated subscribers of D19247: Merge hold_count into wire_count..
In D19247#413465, @kib wrote:

Don't we need to split the kernel wire/user wire limits for this to be feasible, otherwise intensive io holding pages would deny user wiring requests.

Tue, May 14, 2:38 AM
markj updated the diff for D19247: Merge hold_count into wire_count..

Rebase.

Tue, May 14, 2:27 AM
markj committed rS347556: MFC r347227:.
MFC r347227:
Tue, May 14, 12:34 AM

Mon, May 13

markj abandoned D19390: Split kernel and user wire accounting..

A different approach was taken, see D19908.

Mon, May 13, 7:08 PM
markj abandoned D20065: Use system wiring in vmm(4)..
Mon, May 13, 7:07 PM
markj committed rD53012: Document OSVERSION 1300025..
Document OSVERSION 1300025.
Mon, May 13, 6:51 PM
markj committed rS347543: Add an UPDATING entry and bump __FreeBSD_version for r347532..
Add an UPDATING entry and bump __FreeBSD_version for r347532.
Mon, May 13, 6:48 PM
markj committed rS347541: Restore the pre-r347532 behaviour of ignoring wiring failures in mmap()..
Restore the pre-r347532 behaviour of ignoring wiring failures in mmap().
Mon, May 13, 6:40 PM
markj committed rS347536: Extend the libcap_sysctl tests..
Extend the libcap_sysctl tests.
Mon, May 13, 5:53 PM
markj closed D17856: Extend cap_sysctl tests..
Mon, May 13, 5:53 PM
markj committed rS347535: Convert the libcap_sysctl test cases to ATF..
Convert the libcap_sysctl test cases to ATF.
Mon, May 13, 5:51 PM
markj closed D17855: Convert existing cap_sysctl tests to ATF..
Mon, May 13, 5:51 PM
markj closed D17854: Add cap_sysctl() and cap_sysctlnametomib()..
Mon, May 13, 5:50 PM
markj committed rS347534: Add cap_sysctl(3) and cap_sysctlnametomib(3)..
Add cap_sysctl(3) and cap_sysctlnametomib(3).
Mon, May 13, 5:50 PM
markj closed D19908: Provide accounting of user-wired pages..
Mon, May 13, 4:39 PM
markj committed rS347532: Provide separate accounting for user-wired pages..
Provide separate accounting for user-wired pages.
Mon, May 13, 4:39 PM
markj added a comment to D19908: Provide accounting of user-wired pages..
In D19908#434693, @ngie wrote:

The user space and test changes all look great. I’ll defer to someone else (like kib@) about the VM changes.
Curious: why was the quantity changed from an int to a u_long quantity? Could a comment be added to note why the change is being made, and can it optionally be committed as a separate change, if the reasoning is for modern architecture scaling?

Mon, May 13, 4:22 PM
markj committed rS347515: Catch up with r347241..
Catch up with r347241.
Mon, May 13, 1:18 AM

Sun, May 12

markj updated subscribers of D20250: Build arm64 kernel with inline ffs, fls.

Perhaps @andrew, @manu, @greg_unrelenting.technology or one of the other ARM folks would be willing to test?

Sun, May 12, 11:56 PM
markj accepted D20225: HiFive Unleashed support.
Sun, May 12, 1:08 AM

Sat, May 11

markj added inline comments to D20225: HiFive Unleashed support.
Sat, May 11, 4:28 PM
markj accepted D20236: sysutils/devcpu-data: update to Intel github release 201900514.
Sat, May 11, 1:52 AM

Fri, May 10

markj accepted D20206: netdump: Ref the interface we're attached to.
Fri, May 10, 10:39 PM
markj added a comment to D20206: netdump: Ref the interface we're attached to.
In D20206#435984, @cem wrote:
  • Add and use netdump_enabled() predicate. We rely on sx_assert neutering

during kdb/panic to obviate the lock assertion during the dump-time code, which
generally does not take locks.

Fri, May 10, 9:56 PM
markj accepted D20206: netdump: Ref the interface we're attached to.
Fri, May 10, 9:20 PM
markj accepted D20233: netdump: Don't store sensitive key data we don't need.
Fri, May 10, 9:10 PM
markj committed rS347439: Atomically update the global gMsgId in libnetgraph..
Atomically update the global gMsgId in libnetgraph.
Fri, May 10, 4:44 PM
markj accepted D20206: netdump: Ref the interface we're attached to.
Fri, May 10, 4:06 PM
markj accepted D20227: Update stale comments in blist.h.

Approved.

Fri, May 10, 3:50 PM
markj added inline comments to D20226: Replace panic() with KASSERT in subr_blist.c.
Fri, May 10, 3:49 PM
markj committed rS347423: Avoid returning a NULL pointer from the Intel hw PRNG ifunc resolver..
Avoid returning a NULL pointer from the Intel hw PRNG ifunc resolver.
Fri, May 10, 4:28 AM
markj closed D20218: Do not return NULL from the Intel hw PRNG ifunc resolver..
Fri, May 10, 4:28 AM

Thu, May 9

markj accepted D20209: RISC-V PMAP: Ignore setting physical memory attributes.
Thu, May 9, 11:12 PM
markj added a reviewer for D20218: Do not return NULL from the Intel hw PRNG ifunc resolver.: secteam.
Thu, May 9, 10:29 PM
markj added inline comments to D20218: Do not return NULL from the Intel hw PRNG ifunc resolver..
Thu, May 9, 10:24 PM
markj added a comment to D20218: Do not return NULL from the Intel hw PRNG ifunc resolver..
In D20218#435623, @cem wrote:

And the linker has no problem with the NULL pointer at ifunc resolution time, it's later when subsystems like DTrace want to examine kernel symbols and trip over a symbol with non-zero size and a value of 0; it's not quite right of me to say that the kernel linker doesn't "handle" this case.

Sure, I guess it's really dtrace that doesn't like zero symbols. We could monkey patch this or we could change dtrace to just ignore ifuncs that resolve to zero. I don't know if there are other consumers than dtrace that might stumble over the same obstacle, though it seems like maybe a pain point for further adoption of ifuncs.

Thu, May 9, 10:16 PM
markj added inline comments to D20218: Do not return NULL from the Intel hw PRNG ifunc resolver..
Thu, May 9, 10:04 PM
markj updated the diff for D20218: Do not return NULL from the Intel hw PRNG ifunc resolver..

Go back to panic().

Thu, May 9, 10:04 PM
markj updated the diff for D20218: Do not return NULL from the Intel hw PRNG ifunc resolver..

Use a better panic message and switch to KASSERT().

Thu, May 9, 8:59 PM
markj added a comment to D20218: Do not return NULL from the Intel hw PRNG ifunc resolver..
In D20218#435593, @cem wrote:

Huh, I figured linker failure would prevent MOD_LOAD and thus it wouldn't matter. I had thought about providing a 3rd panic function like this but didn't know if it was required and couldn't think of a reason it might be. Guess I was mistaken :-).

Thu, May 9, 8:59 PM
markj added reviewers for D20218: Do not return NULL from the Intel hw PRNG ifunc resolver.: cem, kib.
Thu, May 9, 8:44 PM
markj created D20218: Do not return NULL from the Intel hw PRNG ifunc resolver..
Thu, May 9, 8:44 PM

Wed, May 8

markj accepted D20201: Put other CPUs into tight loop when updating microcode from loaded OS..
Wed, May 8, 7:48 PM
markj added inline comments to D20201: Put other CPUs into tight loop when updating microcode from loaded OS..
Wed, May 8, 7:18 PM
markj committed rS347353: MFC r346960:.
MFC r346960:
Wed, May 8, 4:07 PM
markj committed rS347349: MFC r346960:.
MFC r346960:
Wed, May 8, 4:04 PM
markj committed rS347348: MFC r346888:.
MFC r346888:
Wed, May 8, 4:02 PM