Page MenuHomeFreeBSD

cem (Conrad Meyer)
User

Projects

User Details

User Since
Jul 9 2015, 9:56 PM (201 w, 4 d)

Recent Activity

Today

cem added a comment to D20318: random(4): deduplicate explicit_bzero() in harvest.

Thanks for taking a look, Xin!

Tue, May 21, 2:42 AM
cem updated the test plan for D20329: net/dpdk: Fix build after r347984.
Tue, May 21, 1:55 AM
cem added a reviewer for D20329: net/dpdk: Fix build after r347984: Ports Committers.

Need ports committer blessing to fix.

Tue, May 21, 1:49 AM
cem updated the diff for D20329: net/dpdk: Fix build after r347984.

Now with patches 'svn add'ed

Tue, May 21, 1:46 AM
cem created D20329: net/dpdk: Fix build after r347984.
Tue, May 21, 1:45 AM
cem committed rS348026: Include eventhandler.h in more compilation units.
Include eventhandler.h in more compilation units
Tue, May 21, 1:19 AM
cem committed rS348022: Add two missing eventhandler.h headers.
Add two missing eventhandler.h headers
Tue, May 21, 12:04 AM

Yesterday

cem added a comment to D20317: Netdump: coalesce writes less than buffer size for more efficient use of network.

The changes look straightforward and correct to me, modulo stylistic nits. I agree with everything Mark has already said.

Mon, May 20, 7:24 PM
cem committed rS347984: Extract eventfilter declarations to sys/_eventfilter.h.
Extract eventfilter declarations to sys/_eventfilter.h
Mon, May 20, 12:39 AM

Sun, May 19

cem added a watcher for secteam: cem.
Sun, May 19, 9:32 PM
cem added inline comments to D20312: random(4): Generalize algorithm-independent APIs.
Sun, May 19, 9:31 PM
cem added a reviewer for D20312: random(4): Generalize algorithm-independent APIs: secteam.
Sun, May 19, 9:23 PM
cem added a reviewer for D20313: EXPERIMENTAL Fortuna: allow increased concurrency: secteam.

Need secteam approval.

Sun, May 19, 9:22 PM
cem updated the diff for D20313: EXPERIMENTAL Fortuna: allow increased concurrency.

Add default-off (for now) knob to enable/disable the feature

Sun, May 19, 9:21 PM
cem created D20318: random(4): deduplicate explicit_bzero() in harvest.
Sun, May 19, 8:49 PM
cem retitled D20312: random(4): Generalize algorithm-independent APIs from random(4): Reduce locking and memcpy overhead to random(4): Generalize algorithm-independent APIs.
Sun, May 19, 7:37 PM
cem updated the diff for D20312: random(4): Generalize algorithm-independent APIs.
  • Add comment to genrandom justifying not rekeying Chacha (caveat, this whole

function is removed in the following differential, but I did try to flesh out a
similar comment in the relevant place there)

Sun, May 19, 7:36 PM
cem accepted D20305: Set a specific value in vm_guest for older detection methods.

Looks great, thanks for doing this!

Sun, May 19, 6:32 PM
cem added a comment to D20313: EXPERIMENTAL Fortuna: allow increased concurrency.

I think to actually land this, it probably makes sense to hide this behavior behind a non-default tunable. I don't think that complicates things too much, but please discuss. It's certainly a "Relnotes: yes". Is that acceptable from a secteam perspective, @delphij ?

Sun, May 19, 4:20 PM
cem added a comment to D20313: EXPERIMENTAL Fortuna: allow increased concurrency.

Thanks Mark for taking a look at both of these differentials!

Sun, May 19, 4:09 PM
cem added inline comments to D20312: random(4): Generalize algorithm-independent APIs.
Sun, May 19, 3:46 PM
cem committed rS347967: nvd.4: Reference nda(4).
nvd.4: Reference nda(4)
Sun, May 19, 6:01 AM
cem planned changes to D20312: random(4): Generalize algorithm-independent APIs.

Note that the fast path doesn't work on real systems with SMAP. (It worked in a bhyve VM, of course.) I'm not sure vslock provides sufficient protection against concurrent modification of the UVA to copy directly into it, and the actual difference was not large (2-3%). So I will probably revert that portion. The rest is useful for the follow-up patch, see child revision.

Sun, May 19, 5:54 AM

Sat, May 18

cem added a child revision for D20312: random(4): Generalize algorithm-independent APIs: D20313: EXPERIMENTAL Fortuna: allow increased concurrency.
Sat, May 18, 11:19 PM
cem added a parent revision for D20313: EXPERIMENTAL Fortuna: allow increased concurrency: D20312: random(4): Generalize algorithm-independent APIs.
Sat, May 18, 11:19 PM
cem created D20313: EXPERIMENTAL Fortuna: allow increased concurrency.
Sat, May 18, 11:04 PM
cem created D20312: random(4): Generalize algorithm-independent APIs.
Sat, May 18, 9:46 PM
cem added a comment to D20305: Set a specific value in vm_guest for older detection methods.

Thanks for doing this! It mostly looks great to me, just some minor const and bool style quibbles below.

Sat, May 18, 8:44 PM
cem committed rS347948: Add DragonFly's partition number to fdisk(8) and diskmbr.h.
Add DragonFly's partition number to fdisk(8) and diskmbr.h
Sat, May 18, 12:22 AM

Thu, May 16

cem committed rS347888: mips: Implement basic pmap_kenter_device, pmap_kremove_device.
mips: Implement basic pmap_kenter_device, pmap_kremove_device
Thu, May 16, 7:11 PM
cem abandoned D20280: mips: Implement basic pmap_kenter_device, pmap_kremove_device.

Committed in r347888.

Thu, May 16, 7:11 PM
cem created D20280: mips: Implement basic pmap_kenter_device, pmap_kremove_device.
Thu, May 16, 6:23 PM
cem committed rS347828: xdma(4): Fix invalid pointer use (breaks arm.SOCFPGA build).
xdma(4): Fix invalid pointer use (breaks arm.SOCFPGA build)
Thu, May 16, 5:34 PM
cem accepted D20273: db show thread: avoid overlow in tick conversion.
Thu, May 16, 5:11 AM
cem committed rS347639: x86: Correctly identify bhyve hypervisor.
x86: Correctly identify bhyve hypervisor
Thu, May 16, 1:33 AM

Mon, May 13

cem committed rS347555: Revert r346292 (permit_nonrandom_stackcookies).
Revert r346292 (permit_nonrandom_stackcookies)
Mon, May 13, 11:37 PM
cem added a comment to D20205: sysctl.name false positive.

Hi Alfonso,

Mon, May 13, 9:52 PM
cem committed rS347546: Fortuna: Fix false negatives in is_random_seeded().
Fortuna: Fix false negatives in is_random_seeded()
Mon, May 13, 7:35 PM
cem closed D20239: Fortuna: Fix false negatives in is_random_seeded().
Mon, May 13, 7:35 PM
cem added a comment to D20239: Fortuna: Fix false negatives in is_random_seeded().

Thanks for reviewing, Xin!

Mon, May 13, 7:10 PM

Sun, May 12

cem accepted D20242: Depessimize arc4rand.
Sun, May 12, 2:24 AM

Sat, May 11

cem updated the test plan for D20239: Fortuna: Fix false negatives in is_random_seeded().
Sat, May 11, 4:40 PM
cem updated the diff for D20239: Fortuna: Fix false negatives in is_random_seeded().

Fix lock recursion; didn't recall that pre_read invoked seeded().

Sat, May 11, 4:38 PM
cem added a child revision for D20239: Fortuna: Fix false negatives in is_random_seeded(): D19928: random(4): More thoroughly attempt to ensure seeding during priming.
Sat, May 11, 4:33 PM
cem added a parent revision for D19928: random(4): More thoroughly attempt to ensure seeding during priming: D20239: Fortuna: Fix false negatives in is_random_seeded().
Sat, May 11, 4:33 PM
cem created D20239: Fortuna: Fix false negatives in is_random_seeded().
Sat, May 11, 4:31 PM
cem created P258 unblock random D19928-lite.
Sat, May 11, 4:06 PM

Fri, May 10

cem committed rS347473: netdump: Ref the interface we're attached to.
netdump: Ref the interface we're attached to
Fri, May 10, 11:13 PM
cem closed D20206: netdump: Ref the interface we're attached to.
Fri, May 10, 11:13 PM
cem committed rS347471: netdump: Fix boot-time configuration typo.
netdump: Fix boot-time configuration typo
Fri, May 10, 11:10 PM
cem added inline comments to D20206: netdump: Ref the interface we're attached to.
Fri, May 10, 11:09 PM
cem added a comment to D20206: netdump: Ref the interface we're attached to.

Where are we calling netdump_enabled() after a panic?

Fri, May 10, 10:21 PM
cem committed rS347467: netdump: Don't store sensitive key data we don't need.
netdump: Don't store sensitive key data we don't need
Fri, May 10, 9:55 PM
cem closed D20233: netdump: Don't store sensitive key data we don't need.
Fri, May 10, 9:55 PM
cem updated the diff for D20206: netdump: Ref the interface we're attached to.
  • 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, 9:54 PM
cem added inline comments to D20206: netdump: Ref the interface we're attached to.
Fri, May 10, 9:44 PM
cem added a child revision for D20233: netdump: Don't store sensitive key data we don't need: D20206: netdump: Ref the interface we're attached to.
Fri, May 10, 8:59 PM
cem added a parent revision for D20206: netdump: Ref the interface we're attached to: D20233: netdump: Don't store sensitive key data we don't need.
Fri, May 10, 8:59 PM
cem updated the diff for D20206: netdump: Ref the interface we're attached to.
  • Refactor de-configuration into netdump_unconfigure() routine
  • Rebase on D20233
Fri, May 10, 8:59 PM
cem updated the summary of D20233: netdump: Don't store sensitive key data we don't need.
Fri, May 10, 8:55 PM
cem updated the summary of D20233: netdump: Don't store sensitive key data we don't need.
Fri, May 10, 8:54 PM
cem created D20233: netdump: Don't store sensitive key data we don't need.
Fri, May 10, 8:35 PM
cem planned changes to D20206: netdump: Ref the interface we're attached to.
Fri, May 10, 7:38 PM
cem added a reviewer for D20206: netdump: Ref the interface we're attached to: vangyzen.

Tag :)

Fri, May 10, 4:02 PM
cem added a comment to D20219: ddb(4): Add 'sysctl' command.

Regarding the live version panic - it is a little odd... __mtx_assert() does check for SCHEDULER_STOPPED... I haven't poked around the live breakpoint in quite a while... wonder if that isn't lit? That would do it. Anyway, I trust you to dig a little deeper on that. Thanks. Otherwise, I'm fine with things.

Fri, May 10, 12:35 AM

Thu, May 9

cem accepted D20218: Do not return NULL from the Intel hw PRNG ifunc resolver..
Thu, May 9, 10:07 PM
cem added a comment to D20218: Do not return NULL from the Intel hw PRNG ifunc resolver..

Kernel ifuncs are resolved quite early during boot, well before MOD_LOAD.

Thu, May 9, 9:23 PM
cem created D20219: ddb(4): Add 'sysctl' command.
Thu, May 9, 9:15 PM
cem accepted D20218: Do not return NULL from the Intel hw PRNG ifunc resolver..

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, 8:50 PM
cem added a comment to D20205: sysctl.name false positive.

So, I think this was an intentional feature and there are automatically generated numerical nodes which are not actually present in the tree structure.

Thu, May 9, 6:20 PM
cem created D20206: netdump: Ref the interface we're attached to.
Thu, May 9, 12:12 AM

Wed, May 8

cem committed rS347329: random(4): Don't complain noisily when an entropy source is slow.
random(4): Don't complain noisily when an entropy source is slow
Wed, May 8, 2:54 PM
cem closed D20195: random: Don't complain noisily when an entropy source is slow.
Wed, May 8, 2:54 PM
cem added a comment to D20195: random: Don't complain noisily when an entropy source is slow.

Thanks Mark and Xin!

Wed, May 8, 2:48 PM
cem created D20195: random: Don't complain noisily when an entropy source is slow.
Wed, May 8, 3:31 AM
cem committed rS347239: random: x86 driver: Prefer RDSEED over RDRAND when available.
random: x86 driver: Prefer RDSEED over RDRAND when available
Wed, May 8, 12:45 AM
cem closed D20192: random: x86 driver: Prefer RDSEED over RDRAND when available.
Wed, May 8, 12:45 AM
cem closed D20194: vmm(4): Pass through RDSEED feature bit to guests.
Wed, May 8, 12:40 AM
cem committed rS347238: vmm(4): Pass through RDSEED feature bit to guests.
vmm(4): Pass through RDSEED feature bit to guests
Wed, May 8, 12:40 AM
cem added a child revision for D20194: vmm(4): Pass through RDSEED feature bit to guests: D20192: random: x86 driver: Prefer RDSEED over RDRAND when available.
Wed, May 8, 12:38 AM
cem added a parent revision for D20192: random: x86 driver: Prefer RDSEED over RDRAND when available: D20194: vmm(4): Pass through RDSEED feature bit to guests.
Wed, May 8, 12:38 AM
cem updated the diff for D20192: random: x86 driver: Prefer RDSEED over RDRAND when available.
  • Drop ifndef wat
  • Drop bangbang "operator" -- bool implicitly coerces on assignment already
Wed, May 8, 12:35 AM
cem added inline comments to D20192: random: x86 driver: Prefer RDSEED over RDRAND when available.
Wed, May 8, 12:33 AM
cem created D20194: vmm(4): Pass through RDSEED feature bit to guests.
Wed, May 8, 12:31 AM
cem created D20192: random: x86 driver: Prefer RDSEED over RDRAND when available.
Wed, May 8, 12:15 AM

Tue, May 7

cem added a comment to D19928: random(4): More thoroughly attempt to ensure seeding during priming.

Now that some of the dust has settled around random, I'd like to work on getting this in.

Tue, May 7, 10:04 PM
cem closed D20187: Fix r347229 libsbuf symbol version.
Tue, May 7, 9:15 PM
cem committed rS347235: Fix libsbuf sbuf_printf_drain symbol version.
Fix libsbuf sbuf_printf_drain symbol version
Tue, May 7, 9:15 PM
cem created D20187: Fix r347229 libsbuf symbol version.
Tue, May 7, 8:10 PM
cem committed rS347229: device_printf: Use sbuf for more coherent prints on SMP.
device_printf: Use sbuf for more coherent prints on SMP
Tue, May 7, 5:47 PM
cem closed D16690: make device_printf use sbuf.
Tue, May 7, 5:47 PM
cem commandeered D16690: make device_printf use sbuf.
Tue, May 7, 5:47 PM
cem accepted D20183: makesyscalls: use @generated tag in generated files.
Tue, May 7, 3:56 PM
cem accepted D20182: Avoid literal @generated tag in file-generating scripts.
Tue, May 7, 3:55 PM

Mon, May 6

cem added a comment to D19948: Add witness(4) warnings around potentially blocking requests for random.
In D19948#434568, @jhb wrote:

Correct, in this case the crypto driver would know to not create the session unless it was seeded, so its associated code that invokes arc4random will only be run if the PRNG is seeded, but the check and the arc4random use are in different threads, so TDP_RNG_SEEDED doesn't help.

Mon, May 6, 11:33 PM
cem committed rS347192: List-ify kernel dump device configuration.
List-ify kernel dump device configuration
Mon, May 6, 6:24 PM
cem closed D19996: List-ify kernel dump device configuration.
Mon, May 6, 6:24 PM
cem edited P255 WIP arc4random stacks.
Mon, May 6, 6:16 PM
cem updated the test plan for D20172: tcp_syncache: Mute a false positive random-may-block warning.
Mon, May 6, 6:15 PM
cem added a child revision for D19948: Add witness(4) warnings around potentially blocking requests for random: D20172: tcp_syncache: Mute a false positive random-may-block warning.
Mon, May 6, 6:14 PM