Work around an integer division resulting in 0 and thus the
congestion window not being incremented, if cwnd > maxseg^2.
As suggested in RFC2581 increment the cwnd by 1 in this case.
See http://caia.swin.edu.au/reports/080829A/CAIA-TR-080829A.pdf
for more details.
Submitted by: Alana Huebner, Lawrence Stewart,
Grenville Armitage (caia.swin.edu.au)
Reviewed by: dwmalone, gnn, rpaulo
MFC After: 3 days
IGP based radeon cards such as the rs485 and rs690 require that the GART table
be un-cached. Our previous memory barrier was not sufficient. This patch
allocates the IGP GART tables using the BUS_DMA_NOCACHE flag to get these
cards working.
We should never call drm_pci_alloc() while holding locks, due the the
calls to bus_dma. There were multiple paths that held different locks or
no locks at all. This patch ensures that all of the calling paths drop
their lock(s) before calling drm_pci_alloc().
Add new TAPGIFNAME tap(4) character device ioctl. This is a
convenient shortcut to obtain network interface name using
file descriptor for character device.
- Reduce scope of #ifdef's in uma_zcreate() call in init_turnstile0().
- Set UMA_ZONE_NOFREE so that the per-turnstile spin locks are type stable
to avoid a race where one thread might dereference a lock in a free'd
turnstile that was previously used by another thread.
marius [Mon, 8 Sep 2008 21:24:25 +0000 (21:24 +0000)]
For cheetah-class CPUs ensure that the dt512_0 is set to hold 8k pages
for all three contexts and configure the dt512_1 to hold 4MB pages for
them (e.g. for direct mappings).
This might allow for additional optimization by using the faulting
page sizes provided by AA_DMMU_TAG_ACCESS_EXT for bypassing the page
size walker for the dt512 in the superpage support code.
marius [Mon, 8 Sep 2008 20:38:48 +0000 (20:38 +0000)]
USIII and beyond CPUs have stricter requirements when it comes
to synchronization needed after stores to internal ASIs in order
to make side-effects visible. This mainly requires the MEMBAR #Sync
after such stores to be replaced with a FLUSH. We use KERNBASE as
the address to FLUSH as it is guaranteed to not trap. Actually,
the USII synchronization rules also already require a FLUSH in
pretty much all of the cases changed.
We're also hitting an additional USIII synchronization rule which
requires stores to AA_IMMU_SFSR to be immediately followed by a DONE,
FLUSH or RETRY. Doing so triggers a RED state exception though so
leave the MEMBAR #Sync. Linux apparently also has gotten away with
doing the same for quite some time now, apart from the fact that
it's not clear to me why we need to clear the valid bit from the
SFSR in the first place.
marius [Mon, 8 Sep 2008 20:20:44 +0000 (20:20 +0000)]
o Move the MODULE_DEPEND() for cam(4) from the esp_sbus.c front-end to
the ncr53c9x.c core where it actually belongs so future front-ends
don't need to add it.
o Use the correct OFW property when looking for the initiator ID of the
SBus device.
o Don't specify an alignment when creating the parent DMA tag for
SUNW,fas; their DMA engine doesn't require an alignment constraint
and it's no inherited by the child DMA tags anyway (which probably
is a bug though).
o Drop the superfluous sc_maxsync and use sc_minsync instead. The
former apparently was added due to a confusion with the maximum
frequency used in cam(4), which basically corresponds to the
inverse of minimum sync period.
o Merge ncr53c9x.c from NetBSD:
1.116: NCRDMA_SETUP() should be called before NCR_SET_COUNT() and
NCRCMD_DMA command in ncr53c9x_select().
1.125: free allocated resources on detach.
o Static'ize ncr53c9x_action(), ncr53c9x_init() and ncr53c9x_reset()
as these are not required outside of ncr53c9x.c.
o In ncr53c9x_attach() don't leak the device mutex in case attaching
fails.
o Register an asynchronous notification handler so in case cam(4)
reports a lost device we can cancel outstanding commands and
restore the default parameters for the target in question.
o For FAS366 correctly support 16-bit target IDs and let it know
that we use 32-bit transfers.
o Overhaul the negotiation of transfer settings. This includes
distinguishing between current and goal transfer settings of the
target so we can renegotiate their goal settings when necessary
and correcting the order in which tagged, wide and synchronous
transfers are negotiated.
o If we are requesting sense, force a renegotiation if we are
currently using anything different from asynchronous at 8 bit
as the target might have lost our transfer negotiations.
o In case of an XPT_RESET_BUS just directly call ncr53c9x_init()
instead of issuing a NCRCMD_RSTSCSI, which in turn will issue an
interrupt that is treated as an unexpected SCSI bus reset by
ncr53c9x_intr() and thus calls ncr53c9x_init(). Remove the now
no longer used ncr53c9x_scsi_reset().
o Correct an off-by-one error when setting cpi->max_lun.
o In replace printf(9) with device_printf(9) calls where appropriate
and in ncr53c9x_action() remove some unnecessarily verbose messages.
o In ncr53c9x_sched() use TAILQ_FOREACH() instead of reimplementing
it and consolidate two tagging-related target info checks into one.
o In ncr53c9x_done() set the CAM status to CAM_SCSI_STATUS_ERROR when
appropriate, respect CAM_DIS_AUTOSENSE and teach it to return SCSI
status information.
o In ncr53c9x_dequeue() ensure the tags are cleared.
o Use ulmin() instead of min() where appropriate.
o In ncr53c9x_msgout() consistently use the reset label.
o When we're interrupted during a data phase and the DMA engine is
still active, don't panic but reset the core and the DMA engine as
this should be sufficient. Also, the typical problem for triggering
this was the lack of renegotiation when requesting sense.
o Correctly handle DEVICE RESETs.
o Adapt the locking of esp(4) to MPSAFE cam(4). This includes moving
the calls of lsi64854_attach() to the bus front-ends so it can pass
the esp(4) mutex to bus_dma_tag_create(9).
o Change the LSI64854 driver to not create a DMA tag and map for the
Ethernet channel as le(4) will handle these on its own as well as
sync and unload the DMA maps for the SCSI and parallel port channel
after a DMA transfer.
o Cam(4)'ify some NetBSD-centric comments.
o Use bus_{read,write}_*(9) instead of bus_space_{read,write}_*(9)
and take advantage of rman_get_rid(9) in order to save some softc
members.
Close a race in sleepq_broadcast() where the sleepq could be reused after
it had been assigned to the last sleeping thread. That thread might have
started running on another CPU and have reused that sleep queue. Fix it
by just walking the thread queue using TAILQ_FOREACH_SAFE() rather than
a while loop.
peter [Mon, 8 Sep 2008 11:41:20 +0000 (11:41 +0000)]
Pass 1 of mpsafetty-ifying si(4). It compiles and has basic functionality,
but needs a lot more work. In particular, it has no flow control and has
a tendency to race when giving commands. It still uses Giant for the
tty and driver lock, but this is a keep-it-simple feature for now.
Some of the [temporary] proliferation of messages lines are way too long.
The pcb_gs32p should be per-cpu, not per-thread pointer. This is
location in GDT where the segment descriptor from pcb_gs32sd is
copied, and the location is in GDT local to CPU.
Commit the no-brainer parts of my space saving kernel experiments. We
don't inline the locking primitives, and only grab those parts of mii
that we really need. Other space optimizations are too agressive for
the generic file (removing all of usb, and loading it as kernel
modules).
To my reading there are no real consumers of ip6_plen (IPv6
Payload Length) as set in tcpip_fillheaders().
ip6_output() will calculate it based of the length from the
mbuf packet header itself.
So initialize the value in tcpip_fillheaders() in correct
(network) byte order.
With the above change, to my reading, all places calling tcp_trace()
pass in the ip6 header via ipgen as serialized in the mbuf and with
ip6_plen in network byte order.
Thus convert the IPv6 payload length to host byte order before printing.
Minor cleanup of this config file:
o It has been tested only on KB9202, KB9202A and KB9202B boards
o Better comments about hints
o option<space><tab>
o Add newer SX_NOINLINE option.
o Fix a few comments
Split tcp_mss() in tcp_mss() and tcp_mss_update() where the former
calls the latter.
Merge tcp_mss_update() with code from tcp_mtudisc() basically
doing the same thing.
This gives us one central place where we calcuate and check mss values
to update t_maxopd (maximum mss + options length) instead of two slightly
different but almost equal implementations to maintain.
Segment registers are stored in the uc_mcontext member of the struct
l_ucontext. To restore the registers content, trampoline needs to
dereference uc_mcontext instead of taking some undefined values from
l_ucontext.
V_irtualize SVN r182846 tcp_mssdflt/tcp_v6mssdflt procedure based
sysctl implementations for VIMAGE the same way we did elsewhere:
update the implementation but leave the globals and the SYSCTL
statement untouched.
Convert SYSCTL_INTs for tcp_mssdflt and tcp_v6mssdflt to
SYSCTL_PROCs and check that the default mss for neither v4 nor
v6 goes below the minimum MSS constant (216).
This prevents people from shooting themselves in the foot.
PR: kern/118455 (remotely related)
Reviewed by: silby (as part of a larger patch in March)
MFC after: 2 months
- Make use of the new and g_device_path utility function in libgeom to avoid
duplication of code in fdisk and boot0cfg. Also make use of g_providername to
fix an issue with fdisk and boot0cfg not using the correct provider when
writing the MBR.
- Add a new ioctl for getting the provider name of a geom provider.
- Add a routine for looking up a device and checking if it is a valid geom
provider given a partial or full path to its device node.
Catch a possible NULL pointer deref in case the offsets got mangled
somehow.
As a consequence we may now get an unexpected result(*).
Catch that error cases with a well defined panic giving appropriate
pointers to ease debugging.
(*) While the concensus was that the case should never happen unless
there was a bug, noone was definitively sure.
Discussed with: kmacy (about 8 months back)
Reviewed by: silby (as part of a larger patch in March)
MFC after: 2 months
Add a second KASSERT checking for len >= 0 in the tcp output path.
This is different to the first one (as len gets updated between those
two) and would have caught various edge cases (read bugs) at a well
defined place I had been debugging the last months instead of
triggering (random) panics further down the call graph.
When restoring a directory, allow symlinks to be followed. The full
logic here gets a little complex, but the net effect is that the
SECURE_SYMLINKS flag will prevent us from ever following a symlink.
Without it, we'll only follow symlinks to dirs. bsdtar specifies
SECURE_SYMLINKS by default, suppresses it for -P.
I've also beefed up the write_disk_secure test to verify this
behavior.
- Add the i386_memio_map_load() function to load I/O address table.
- Add the bus_space_compare macro for bus_space consistency.
- Switch using the bus_space_map_load() in isa_load_resourcev().
sam [Sat, 6 Sep 2008 17:42:51 +0000 (17:42 +0000)]
o kill IEEE80211_NODE_PSUPDATE; it was never used
o kill IEEE80211_NODE_CHWUPDATE; it's use was pointless and it was defined
wrong (overlap w/ existing flags)
sam [Sat, 6 Sep 2008 17:38:20 +0000 (17:38 +0000)]
o validate the ba policy in addba response
o leave a check for the max ba window disabled; we accept out of range
values and just truncate them but may want to act differently in the future
Export the IPFW_DEFAULT_RULE outside ip_fw2.c. This number in not only
the default rule number but also the maximum rule number. User space
software such as ipfw and natd should be aware of its value. The
software that already includes ip_fw.h should use the defined value. All
other a expected to use sysctl (as discussed on net@).
Make TIOCCONS use priv_check() instead of checking /dev/console permissions.
As discussed with Robert on IRC, checking the permissions on
/dev/console to see if we can call TIOCCONS could be unreliable. When we
run a chroot() without a devfs instance mounted inside, it won't
actually check the permissions on the device node inside the devfs
instance.
Using the already existing PRIV_TTY_CONSOLE for this seems like a better
idea.
Fix double free in setfacl(1). Description from the author:
Initially, 'acl' (an 'acl_t *') is allocated, and its ACCESS_ACL and
DEFAULT_ACL fields are passed to the 'libc' ACL routines for subsequent
allocation. If the '-m' option (merge existing ACL with a new one) is
specified, then 'set_acl_mask()' will be called and passed one of the
two ACLs. This function, in turn, replaces this given ACL structure by
another, freshly allocated. However, the pointer in the 'acl' variable
in the caller is not updated. The caller then proceeds to free the ACL,
incurring in a double free condition.
Submitted by: Pedro Martelletto <pedro at ambientworks.net>
Approved by: rwatson (mentor)
Kill bogus #if 0'd stuff for interrupts. They don't happen, and this
driver will need more serious help to work with an interrupt driven
path. There's many subtleties in driving the DMA engine with
interrupts in many configurations. Best to not "guess" what the right
way would be and mislead people.
Kill vestiges of the special case code we once had in place for usb
memory allocation. It was change to include the range in the normal
memory area, so these ifdef'd out special cases are no longer useful
to keep around.
Be consistent about whether these multi-lined macros are separated by
a blank line. Some were, some weren't. Decide in favour of the line
as it matches what an inline would do, and it's easier to read.
Keep track of the active buffer on output. For the moment, panic if
the device indicates that it wasn't able to write all the data in the
buffer out.
Ed Schouten doesn't like the idea of a panic here. I think for
production code, we need something better. For right now, while we're
trying to assess the impact of this issue, a panic is OK. So complain
to me, not him if this is hit.
Add two example regexps: (1) one for matching all the characters
that belong in a character class, and (2) one for matching all
the characters *not* in a character class.
Submitted by: Mark B, mkbucc at gmail.com
MFC after: 3 days
Catch up after last insmntque() changes:
- The vnode has to be locked exclusively before calling insmntque().
- Until I find a way to handle insmntque() failures use VV_FORCEINSMQ flag
to force insmntque() to always succeed.
MfP4: Handle entries with unset size properly: Regular files with
unspecified size are "unlimited" (required by Zip reader, which
sometimes does not know the uncompressed size of an entry until it
gets to the end). Also, hardlinks with unspecified (or zero) size do
not overwrite the data on disk nor do they set metadata. This is
compatible with GNU tar and NetBSD pax behavior.
MfP4: Use the set/unset tracking to determine when mtime or atime
is unavailable. Use start_time as a substitute when one is missing;
if both are missing, skip the call to {f,l,}utime{s,}() entirely.