Commit Graph

38606 Commits

Author SHA1 Message Date
Poul-Henning Kamp
7194d335cf Run a revision of the devstat interface:
Kernel:

Change statistics to use the *uptime() timescale (ie: relative to
boottime) rather than the UTC aligned timescale.  This makes the
device statistics code oblivious to clock steps.

Change timestamps to bintime format, they are cheaper.

Remove the "busy_count", and replace it with two counter fields:
"start_count" and "end_count", which are updated in the down and
up paths respectively.  This removes the locking constraint on
devstat.

Add a timestamp argument to devstat_start_transaction(), this will
normally be a timestamp set by the *_bio() function in bp->bio_t0.
Use this field to calculate duration of I/O operations.

Add two timestamp arguments to devstat_end_transaction(), one is
the current time, a NULL pointer means "take timestamp yourself",
the other is the timestamp of when this transaction started (see
above).

Change calculation of busy_time to operate on "the salami principle":
Only when we are idle, which we can determine by the start+end
counts being identical, do we update the "busy_from" field in the
down path.  In the up path we accumulate the timeslice in busy_time
and update busy_from.

Change the byte_* and num_* fields into two arrays: bytes[] and
operations[].

Userland:

Change the misleading "busy_time" name to be called "snap_time" and
make the time long double since that is what most users need anyway,
fill it using clock_gettime(CLOCK_MONOTONIC) to put it on the same
timescale as the kernel fields.

Change devstat_compute_etime() to operate on struct bintime.

Remove the version 2 legacy interface: the change to bintime makes
compatibility far too expensive.

Fix a bug in systat's "vm" page where boot relative busy times would
be bogus.

Bump __FreeBSD_version to 500107

Review & Collaboration by:	ken
2003-03-15 21:59:06 +00:00
Matthew N. Dodd
0c61d694a8 Update interface statistics after MAC and IFF_UP|IFF_RUNNING checks. 2003-03-15 21:51:39 +00:00
Matthew N. Dodd
f156dd7b68 - Adopt tests for (IFF_UP|IFF_RUNNING) and non local unicast packets
in promiscuous mode from if_fddisubr.c.
- Add comment to reduce diffs.
2003-03-15 21:42:19 +00:00
Matthew N. Dodd
9a3c9f3971 Add MAC support.
This is the same code that was added in 1.70 of if_fddisubr.c
2003-03-15 21:30:00 +00:00
Matthew N. Dodd
15eccea260 Use llc_control rather than llc_snap.control. 2003-03-15 20:35:19 +00:00
Matthew N. Dodd
ad20d0f577 - Add comment.
- Whitespace fixes.
2003-03-15 20:33:30 +00:00
Matthew N. Dodd
06f684b004 Reduce code differences. 2003-03-15 19:37:44 +00:00
Matthew N. Dodd
a93a116153 Use ISO88025_ADDR_LEN where appropriate. 2003-03-15 19:25:00 +00:00
Matthew N. Dodd
dbd05f2e56 Don't use etherbroadcastaddr; use iso88025_broadcastaddr. 2003-03-15 19:16:39 +00:00
Matthew N. Dodd
fd87fbf6b8 - Remove definition of senderr() from iso88025.h.
- Use definition of senderr() from if_ethersubr.c.
2003-03-15 19:10:19 +00:00
Maxim Sobolev
71fbba351a Remove duplicate line introduced in the previous commit.
Submitted by:	bde
MFC after:	2 weeks
2003-03-15 18:49:41 +00:00
Matthew N. Dodd
386dca02cb Some whitespace/style/readability changes. 2003-03-15 17:54:49 +00:00
Matthew N. Dodd
7c16bd3a8b Add iso88025_resolvemulti().
Cribbed from net/if_fddisubr.c
2003-03-15 16:49:08 +00:00
Matthew N. Dodd
bf54cb290a Fix formatting of iso88025_ifattach(). 2003-03-15 16:41:35 +00:00
Matthew N. Dodd
fe7e109c6f Re-order and prune includes. 2003-03-15 16:37:28 +00:00
Maxim Sobolev
084254f807 - Add minimal support for TI16754 4xUART chip into sio(4) driver and remove
now unnecessary hack from the previous commit;
- Add support for Interrupt Latch Register (ILR) into puc(4). So far only
  ILRs compatible with specifications from Digi International are supported.
  Support for other types of ILRs could be easily added later;
- Correct clock frequency for IC Book Labs Dreadnought x16 Lite board;
- Enable ILR detection/usage for IC Book Labs Dreadnought x16 boards.

Sponsored by:   IC Book Labs
MFC after:      2 weeks
2003-03-15 16:25:40 +00:00
Matthew N. Dodd
40811c1473 Add module data and version. 2003-03-15 15:38:02 +00:00
Matthew N. Dodd
4fd80d5b9d s/llc_un.type_snap/llc_snap/g 2003-03-15 15:35:25 +00:00
Matthew N. Dodd
67df65de97 Formatting and whitespace changes. 2003-03-15 15:09:11 +00:00
Takanori Watanabe
869c12cbb4 Add module dependency. This makes udbp.ko to be loaded. 2003-03-15 13:03:30 +00:00
Poul-Henning Kamp
fe002e7546 Fix a bug I introduced by overlooking the fact that "unit number" can
be one of several things in this driver.

Spotted & Tested by:	alfred
2003-03-15 12:16:33 +00:00
Poul-Henning Kamp
0be3115328 Don't use the devstat->busy_count for state decisions in the device
drivers.  Doing so imposes atomicity and locking constraints on the
devstat API.

By:	ken
2003-03-15 11:00:56 +00:00
Poul-Henning Kamp
09a6b2aad6 Call devstat_start_transaction_bio() instead of devstat_start_transaction(). 2003-03-15 10:56:19 +00:00
Poul-Henning Kamp
1ecc485c1b Call devstat_start_transaction_bio() instead of devstat_start_transaction() 2003-03-15 10:50:44 +00:00
Poul-Henning Kamp
81876757ec Use devstat_{start,end}_transaction_bio().
Remember to set bio_resid correctly first.
2003-03-15 10:49:26 +00:00
Poul-Henning Kamp
9fa85de269 Add a devstat_start_transaction_bio() to match the
devstat_end_transaction_bio() we already have.

For now it just calls devstat_start_transaction(), but that will change
shortly.
2003-03-15 10:33:32 +00:00
Crist J. Clark
010dabb047 Add a 'verrevpath' option that verifies the interface that a packet
comes in on is the same interface that we would route out of to get to
the packet's source address. Essentially automates an anti-spoofing
check using the information in the routing table.

Experimental. The usage and rule format for the feature may still be
subject to change.
2003-03-15 01:13:00 +00:00
Mike Barcroft
13141dbc38 Add missing types, update standard namespaces, and use variable types
that don't require <sys/types.h>.
2003-03-14 16:09:48 +00:00
Maxime Henrion
40b63da2a9 Bah, get it right this time and add sys/lock.h before sys/mutex.h. 2003-03-14 13:30:31 +00:00
Jake Burkholder
7b66da6e32 lock.h must be included before mutex.h. 2003-03-14 07:19:29 +00:00
Jeff Roberson
8501ead911 - Add a forgotten BUF_LOCK()
Most sincere apologies to:	jake
2003-03-14 05:13:19 +00:00
David Xu
9a4b78c9da Export current time when returning from never blocked syscall. 2003-03-14 03:52:16 +00:00
Maxime Henrion
7541142438 Oops, add missing includes. Pass me the pointy hat.
Reported by:	jake
2003-03-14 00:04:37 +00:00
John Baldwin
2d055ab20f Trim some trailing whitespace. 2003-03-13 23:07:09 +00:00
Scott Long
6f420b1756 Now that bus_dmamem_alloc() handles its Giant mutex requirements itself,
don't bother doing the same in the code that calls it.
2003-03-13 22:56:11 +00:00
John Baldwin
b62f75cf44 - Change the linux_[gs]et_os{name, release, s_version}() functions to
take a thread instead of a proc for their first argument.
- Add a mutex to protect the system-wide Linux osname, osrelease, and
  oss_version variables.
- Change linux_get_prison() to take a thread instead of a proc for its
  first argument and to use td_ucred rather than p_ucred.  This is ok
  because a thread's prison does not change even though it's ucred might.
- Also, change linux_get_prison() to return a struct prison * instead of
  a struct linux_prison * since it returns with the struct prison locked
  and this makes it easier to safely unlock the prison when we are done
  messing with it.
2003-03-13 22:45:43 +00:00
John Baldwin
75768576cc Add a new userland-visible ktrace flag KTR_DROP and an internal ktrace flag
KTRFAC_DROP to track instances when ktrace events are dropped due to the
request pool being exhausted.  When a thread tries to post a ktrace event
and is unable to due to no available ktrace request objects, it sets
KTRFAC_DROP in its process' p_traceflag field.  The next trace event to
successfully post from that process will set the KTR_DROP flag in the
header of the request going out and clear KTRFAC_DROP.

The KTR_DROP flag is the high bit in the type field of the ktr_header
structure.  Older kdump binaries will simply complain about an unknown type
when seeing an entry with KTR_DROP set.  Note that KTR_DROP being set on a
record in a ktrace file does not tell you anything except that at least one
event from this process was dropped prior to this event.  The user has no
way of knowing what types of events were dropped nor how many were dropped.

Requested by:	phk
2003-03-13 18:31:15 +00:00
John Baldwin
a5881ea55a - Cache a reference to the credential of the thread that starts a ktrace in
struct proc as p_tracecred alongside the current cache of the vnode in
  p_tracep.  This credential is then used for all later ktrace operations on
  this file rather than using the credential of the current thread at the
  time of each ktrace event.
- Now that we have multiple ktrace-related items in struct proc that are
  pointers, rename p_tracep to p_tracevp to make it less ambiguous.

Requested by:	rwatson (1)
2003-03-13 18:24:22 +00:00
Maxime Henrion
c0796d1cb4 Grab Giant around calls to contigmalloc() and contigfree() so
that drivers converted to be MP safe don't have to deal with it.
2003-03-13 17:18:48 +00:00
Maxime Henrion
ea458bbcdb Memory allocated with contigmalloc() should be freed with
contigfree(), not with free().
2003-03-13 17:10:54 +00:00
Hartmut Brandt
57648f8aae This corrects a longstanding endian bug in processing LLC/SNAP encoded
frames. A comment in if_atm.h suggests that both macros, that for extracting
the ethertype and that for inserting it, handle their argument in host
byte order. In fact, the inserting macro treated its argument as an opposite
host order short and the calling code feeds it the result of htons(). This
happens to work on i386, but fails on sparc. Make the macro use real host
endianess.

Reviewed by:	kjc, atm@
2003-03-13 12:44:06 +00:00
Jeffrey Hsu
7792ea2700 Greatly simplify the unlocking logic by holding the TCP protocol lock until
after FIN_WAIT_2 processing.

Helped with debugging:	Doug Barton
2003-03-13 11:46:57 +00:00
Søren Schmidt
f711eff1d2 Remove the check for more than one open if one is a write op.
This allows using DVD+RW and DVD-RW as random storage, provided
the 32K blocksize is honoured for DVD-RW (DVD+RW has built in
read-modify-write).
2003-03-13 09:12:17 +00:00
Søren Schmidt
6918962fad Dont free the stats struct, this has been changed by the new stat code.
Spotted by: Ian Dowse <iedowse@maths.tcd.ie>
2003-03-13 09:09:10 +00:00
Søren Schmidt
394046e66f Dont call dma->free unless dma is there.
Spotted by: Ian Dowse <iedowse@maths.tcd.ie>
2003-03-13 09:04:55 +00:00
Ian Dowse
a80cc4e104 In m_dup_pkthdr(), convert the supplied `how' argument into malloc
flags when passing it into m_tag_copy_chain(), as m_tag* functions
use malloc, not mbuf flags.
2003-03-13 09:02:19 +00:00
Søren Schmidt
64eca1e1f2 Add pci id# for the sis648 2003-03-13 07:55:53 +00:00
Jeff Roberson
749ffa4ecd - Add a lock for protecting against msleep(bp, ...) wakeup(bp) races.
- Create a new function bdone() which sets B_DONE and calls wakup(bp). This
   is suitable for use as b_iodone for buf consumers who are not going
   through the buf cache.
 - Create a new function bwait() which waits for the buf to be done at a set
   priority and with a specific wmesg.
 - Replace several cases where the above functionality was implemented
   without locking with the new functions.
2003-03-13 07:31:45 +00:00
Jeff Roberson
e99215a614 - Remove a dead check for bp->b_vp == vp in vtruncbuf(). This has not been
possible for some time.
 - Lock the buf before accessing fields.  This should very rarely be locked.
 - Assert that B_DELWRI is set after we acquire the buf.  This should always
   be the case now.
2003-03-13 07:22:53 +00:00
Jeff Roberson
09f11da5a3 - Remove a race between fsync like functions and flushbufqueues() by
requiring locked bufs in vfs_bio_awrite().  Previously the buf could
   have been written out by fsync before we acquired the buf lock if it
   weren't for giant.  The cluster_wbuild() handles this race properly but
   the single write at the end of vfs_bio_awrite() would not.
 - Modify flushbufqueues() so there is only one copy of the loop.  Pass a
   parameter in that says whether or not we should sync bufs with deps.
 - Call flushbufqueues() a second time and then break if we couldn't find
   any bufs without deps.
2003-03-13 07:19:23 +00:00