In that case I think it is reasonable to patch the leak, if only to avoid seeing the same reports again and again.
Wed, Jul 8
Sun, Jul 5
Sat, Jul 4
Fri, Jul 3
I think this is still useful, is there any concern of committing this to head?
Thu, Jul 2
Tue, Jun 30
approach seems reasonable to me
collate.c looks like an independent bugfix you can just commit first?
Mon, Jun 29
Looks good to my cursory review.
Fri, Jun 26
LGTM, maybe worth a comment?
Wed, Jun 24
I'm not sure off hand of other representative test cases - most likely we can just find them from user feedback as folks start using this.
Also add a WITH_CLANG_FORMAT text description file
Tue, Jun 23
Anyway, it’s easier to change base.
Seems reasonable to me, although I'd also argue that ports ought not set MK_KERNEL_SYMBOLS=no.
Mon, Jun 22
@glebius have you heard back from Cronyx?
Sun, Jun 21
Just FYI I'm planning to add module loading support to flua soon, although the current plan is not to have it enabled under BOOTSTRAPPING.
Sat, Jun 20
Thu, Jun 18
I think for metadata we probably have:
@gbe indeed you can commit to src with approval from any src committer, however I noticed after that this code comes from vendor/NetBSD/vis, so we will want to import the update there and merge to contrib/
Tue, Jun 16
Still need approval from:
- Assar Westerlund
- Peter Wemm
- Tim J. Robbins
Mon, Jun 15
Looking good. Two inline comments and I have not yet reviewed usr.sbin/bhyve/pci_virtio_9p.c
Sun, Jun 14
After setarch linux32 the architecture says armv8l
Fri, Jun 12
looks like it's actually aarch64 for 32-bit uname on 64-bit kernel
lazytwitter turned up https://unix.stackexchange.com/questions/136407/is-my-linux-arm-32-or-64-bit for me,
Thu, Jun 11
Wed, Jun 10
I very much like the cleanup/simplification that comes from this
Jun 9 2020
this lgtm, thanks for all the effort here!
Jun 8 2020
(And, thank you very much for your submission!)
I think this is fine, although I wish we had some standard comment (along the lines of /* XXX */) we could use to identify cases where we intentionally implement a low-fidelity version of Linux behaviour.
Jun 7 2020
Jun 6 2020
Jun 5 2020
The other option would have been to change the build system to pass -Wa,-defsym, SKEIN_USE_ASM=XXX (that's what I did for my godbolt testing.
But either solution seems fine.
We assemble skein_block_asm.S four times - once for sys/modules/crypto, and three times for lib/libmd (.o, .pico, .po). All four instances have the same change when assembled with Clang IAS vs. GNU as 2.17.50: