via SCM_RIGHTS messages are dealt with:
1. unp_gc: make this a kthread.
2. unp_detach: go not call unp_gc directly. instead, wake up unp_gc kthread.
3. unp_scan: do not close files here. instead, put them on a global list
for unp_gc to close, along with a per-file "deferred close count". if
file is already enqueued for close, just increment deferred close count.
this eliminates the recursive calls.
3. unp_gc: scan files on global deferred close list. close each file N
times, as specified by deferred close count in file. continue processing
list until it becomes empty (closing may cause additional files to be
queued for close).
4. unp_gc: add additional bit to mark files we are scanning. set during
initial scan of global file list that currently clears FMARK/FDEFER.
during later scans, never examine / garbage collect descriptors that
we have not marked during the earlier scan. do not proceed with this
initial scan until all deferred closes have been processed. be careful
with locking to ensure no races are introduced between deferred close
and file scan.
5. unp_gc: use dummy file_t to mark position in list when scanning. allow
us to drop filelist_lock. in turn allows us to eliminate kmem_alloc()
and safely close files, etc.
6. prohibit transfer of descriptors within SCM_RIGHTS messages if
(num_files_in_transit > maxfiles / unp_rights_ratio)
7. fd_allocfile: ensure recycled filse don't get scanned.
this is 97% work done by andrew doran, with a couple of minor bug fixes
and a lot of testing by yours truly.
mine) - not quite working, but improves the situation for non-MULTIPROCESSOR
kernels (makes LOCKDEBUG kernels work) and does not make SMP kernels worse:
Rearange cpu_info access and hide the actual implementation of the mapping
from all parts of the code that do not directly deal with it. Do the
mapping early in pmap_bootstrap, so that post-vmlocking2 kernels have
a chance to work.
The actual mapping of the cpus array for SMP kernels has to be fixed still,
but both mrg and me ran out of time and this lay around in our trees far
too long.
between a NULL string and an error.
This fixes sh from exiting with libedit now allowing EINTR to return.
We may need to expand this to an errno check in the future.
- x86_round_page, x86_trunc_page, x86_btop and x86_ptob macros are used with
physical addresses; cast to paddr_t instead of u_long. Issue pointed out
by jym@
- machine_to_phys_mapping[] is a long. This is fine as it holds page
frame numbers (and this fits in a 32bit int as physical addresses are
only 36bits), but cast to paddr_t before << PAGE_SHIFT
- xen_start_info.store_mfn is a long; cast it to paddr_t before << PAGE_SHIFT.
should fix issue pointed out by cegger@
number instead of allocating a new one. Force allocating a new interrupt number
for PCI devices, as the number stored in the PCI interrupt register
may be wrong.
This should help using a pciide controller in compat mode or ISA devices
in a non-0 domain.
- Set the flag to use multicast filter...
Without this flag and with a shared hub, the MAC gets ALL packets and the
packets will be forwarded if the forwarding is enabled. It results in DUP!
- Workaround for "Non-Intel XScale Technology Eratta" No. 29.
AA:BB:CC:DD:EE:xF's packet (the 2nd nibble from LSB is the same as the
MAC's one) had matched the filter (both unicast and multicast).
- Check if too {short,long} frame.
- Add the MD hook to get MAC address
