[illumos-gate] 18309 format: remove -Wno-clobbered -- Toomas Soome <tsoome⊙mc>

[illumos-gate] 18302 libsecdb crashes on memory exhaustion due to unchecked allocations -- Andy Fiddaman <illumos⊙fn>

[illumos-gate] 18303 pfexecd door thread pool grows without bound -- Andy Fiddaman <illumos⊙fn>

jbk, jclulow: turns out the linux pivot does much more than I thought. what I'm in favour of is just mounting a real root filesystem normally, after we're done with the archive

maybe that comes off a USB and ends up in a ramdisk

always fun with the panic handler itself locks up...

my best luck there is to break before the crosscalls

(but then you haven't really "panicked" and the debugger can be a bit odd)

but if I remember, anywhere before panicsys raises the spl is for the best. even just `panic` or `vpanic`

(obviously for the _best_ is to not screw up page tables/interrupts/etc, so this isn't a problem)

heh.. that's the thing, wasn't touching any of that (well possibly using interrupts, but not the interrupt infra itself)...i'm sure it's something silly..

it's a debug kernel, so panic prints all of that extra stuff out

i'm wondering if there's a timing bug w/ writing to a framebuffer and that

also annoyingly... that tends to push out the actual useful contents of hte msgbuf on sytstems w/ large # of cores