introduce a modified allocation mechanism for mbufs and mbuf clusters; one
which can scale under SMP and which offers the possibility of resource
reclamation to be implemented in the future. Notable advantages:
o Reduce contention for SMP by offering per-CPU pools and locks.
o Better use of data cache due to per-CPU pools.
o Much less code cache pollution due to excessively large allocation macros.
o Framework for `grouping' objects from same page together so as to be able
to possibly free wired-down pages back to the system if they are no longer
needed by the network stacks.
Additional things changed with this addition:
- Moved some mbuf specific declarations and initializations from
sys/conf/param.c into mbuf-specific code where they belong.
- m_getclr() has been renamed to m_get_clrd() because the old name is really
confusing. m_getclr() HAS been preserved though and is defined to the new
name. No tree sweep has been done "to change the interface," as the old
name will continue to be supported and is not depracated. The change was
merely done because m_getclr() sounds too much like "m_get a cluster."
- TEMPORARILY disabled mbtypes statistics displaying in netstat(1) and
systat(1) (see TODO below).
- Fixed systat(1) to display number of "free mbufs" based on new per-CPU
stat structures.
- Fixed netstat(1) to display new per-CPU stats based on sysctl-exported
per-CPU stat structures. All infos are fetched via sysctl.
TODO (in order of priority):
- Re-enable mbtypes statistics in both netstat(1) and systat(1) after
introducing an SMP friendly way to collect the mbtypes stats under the
already introduced per-CPU locks (i.e. hopefully don't use atomic() - it
seems too costly for a mere stat update, especially when other locks are
already present).
- Optionally have systat(1) display not only "total free mbufs" but also
"total free mbufs per CPU pool."
- Fix minor length-fetching issues in netstat(1) related to recently
re-enabled option to read mbuf stats from a core file.
- Move reference counters at least for mbuf clusters into an unused portion
of the cluster itself, to save space and need to allocate a counter.
- Look into introducing resource freeing possibly from a kproc.
Reviewed by (in parts): jlemon, jake, silby, terry
Tested by: jlemon (Intel & Alpha), mjacob (Intel & Alpha)
Preliminary performance measurements: jlemon (and me, obviously)
URL: http://people.freebsd.org/~bmilekic/mb_alloc/
are duplicated by newly defined types/options in RFC3121
- We have no backward compatibility issue. There is no apps in our
distribution which use the above types/options.
Obtained from: KAME
MFC after: 2 weeks
lock. We now use temporary variables to save the process argument pointer
and just update the pointer while holding the lock. We then perform the
free on the cached pointer after releasing the lock.
. staticize out_fdc(), there's no longer an ft(4) driver sharing its use
. remove in_fdc(), has been used by ft(4) last time, long since obsoleted
by fd_in()
. move the declaration of fd_clone() to where most of the other function
declarations are
. de-__P()ify fd_clone(), it's been the only _P()ed function in the
entire file
something: offset into the first mbuf of the target chain before copying
the source data over.
Make drivers using m_devget() with a first argument "data - ETHER_ALIGN"
to use the offset argument to pass ETHER_ALIGN in. The way it was previously
done is potentially dangerous if the source data was at the top of a page
and the offset caused the previous page to be copied (if the
previous page has not yet been appropriately mapped).
The old `offset' argument in m_devget() is not used anywhere (it's always
0) and dates back to ~1995 (and earlier?) when support for ethernet trailers
existed. With that support gone, it was merely collecting dust.
Tested on alpha by: jlemon
Partially submitted by: jlemon
Reviewed by: jlemon
MFC after: 3 weeks
the console device was open. At other times, the interrupts that
are used to detect the break signal or ~^B sequence were disabled,
so these events would not be noticed until the next open (e.g. the
next kernel printf). This was mainly a problem while there was no
getty running on the console, such as during bootup or shutdown.
For serial consoles with break-to-debugger support, we now enable
the generation of interrupts at attach time, and we leave them
enabled while the device is closed.
Reviewed by: bde (I've since made chages as per his suggestions)
via the new DIGIIO_SETALTPIN ioctl, and allow the port's ALTPIN setting
to be queried via DIGIIO_GETALTPIN.
The initial state and lock devices are normally used to set and/or
lock ALTPIN settings although the device itself may also be used.
ALTPIN settings are applied per-device and apply to both the callin
and callout device at the same time.
sizeof(ro_dst) is not necessarily the correct one.
this change would also fix the recent path MTU discovery problem for the
destination of an incoming TCP connection.
Submitted by: JINMEI Tatuya <jinmei@kame.net>
Obtained from: KAME
MFC after: 2 weeks
of tunclose() rather than the end, and tunopen() grabbed that unit
before tunclose() finished (one process is allocating it while another
is freeing it!).
It may be worth hanging some sort of rw mutex around all specinfo
calls where d_close and the detach handler get a write lock and all
other functions get a read lock. This would guarantee certain levels
of ``atomicity'' (is that a word?) that people may expect (I believe
Solaris does something like this).
requirements(RFC1573, interface MIB). This change for 4.4BSD was
first introduced in if_ethersubr.c:1.17->1.18.
BTW, iflastchange on all of IFs are inconsistent. e.g.
ether, tun: update
fddi, tokenring, ppp: not update
I'll make patch later.
Obtained from: KAME
MFC after: 2 weeks
converting from the old external mbuf buffer code to the new (with the
MEXTADD() macro). Also free free list memory correctly in
foo_free_jumbo_mem() routines: grab the head of the list, then
remove it, _then_ free() it.
This fixes the memory corruption problem I've been chasing in the level 1
driver.
and its associated constants. Implement _SC_IOV_MAX in the usual way.
Be a bit sloppy about the namespace question; this should get cleared up
in time for 5.0.
MFC after: 1 month
startup routine more closely matches that of alpha and ia64. At some
point the common mutexes shared across all platforms probably should move
into sys/kern_mutex.c.
trace code that was brought over from NetBSD.)
- Check for "syscall_with_err_pushed" as the label prior to a syscall trap
frame rather than "Xlcall_syscall" and "Xint0x80_syscall". We don't
have a valid trapframe during the short range of code that those two
symbols now cover.
- Simplify db_next_frame() to avoid duplicating the code for the different
trap frame types.
- Don't try to trace a swapped-out process. (Brought over from NetBSD via
the new alpha trace code.)
