Without it, on ports where splhigh() is inline, the compiler will optimise
the second SOFTINT_PENDING test in softint_schedule(). A dissasembly
of softint_schedule() with and without the volatile sh_flags confirm this
on sparc.
Because of this there is a race that could lead to the softhand_t
being enqueued twice on si_q, leading to a corrupted queue and
some handler being SOFTINT_PENDING but never called.
Should fix PR kern/38637
big endian mode:
- add IEE_REV_A flag to indicate if chip support the 32 bit BE mode or not
- add IEE_SWAPA32() macro and use it on necessary 32 bit DMA pointers
- rename IEE_SWAP() macro for the SCP address pointer and statistics
counters which require word swap even on Rev B/C chips to IEE_SWAP32()
for clarification
- add comments about these BE mode quirks
Tested on HP9000 735/125 by me and also tested on 715/50 by skrll@
with netbsd-5 branch, and fixes MI part of PR port-hp700/35531.
network to be functionnal, and therefore must be called after
config_finalize(). Unfortunately, config_finalize() attaches md0
through mdattach(), and when tftproot_getfile() attempts to attach it
later by calling mdattach() again, it gets a failure.
This change checks for the RAMdisk size in mdattach(). If it is zero, then
the RAMdisk has not been loaded yet, and nothing is done. It will be
attached later when tftproot_getfile() will call mdattach()
being reclaimed by another thread. after recent changes in cache_lookup_raw,
there's a race between cache_lookup_raw/vtryget and getcleanvnode/vclean.
PR/41028.
way they can be included without having to include DDB.
(arguably all print routines should be behind #ifdef DEBUGPRINT
and options DDB should define that macro, but I'll tackle that later)
acquired bus space from bus_space_alloc(9) or from rbus_space_alloc(),
so that we use the appropriate routine to release the bus space in
pccbbdetach(). This stops panics during cbb(4) detachment (esp. at
shutdown) on some systems.
Thanks to Kurt Schreiner for testing the fix, Kurt, Patrick Welche,
Pouya D. Tafti, and Matt Green for reporting.
our uhci_poll_hub callout might strike. The detaching of the uhub has left
the device member in the pipe NULL, so make sure we do not dereference it.
We could halt the callout before detaching child devices, but it this
would likely interfere with proper servicing the child devices while
they are still around.
write a "feature-rx-notify" to the xenstore, which is used by recent
linux dom0 kernels. This reduce packet loss when using a NetBSD domU
on such linux dom0.
This entry is ignored by NetBSD and older linux domUs.
the name to rump_module_init(). Also, adjust the signature to take
a direct pointer to modinfo and allow passing of props. Finally,
provide rump_module_fini().
