time in ms, or the number of beacons for the
beacon miss alarm. If we haven't had a beacon
in this period, the alarm is sounded and
action (i.e. roaming) is taken.
This patch changes the storage of the beacon
miss threshold to integral beacon count but
corrects the previous intent.
The default beacon miss alarm is going to be
850ms which is half way between the two hard
coded limits that were there before.
The old hard coded limits assumed that the
beacon interval was 100ms and set the limit to
10 (1000ms) for software beacon miss timer
and 7 for hardware beacon miss timer.
The new default is 850ms (the midpoint between
the two previous defaults). This value is rounded
up to the next nearest beacon interval.
There is no upper bound on the beacon miss
threshold specified, since it may need to be
wildly inflated with 25ms beacon interval, or
wildly reduced with 1000ms beacon interval.
The minimum is still 2 beacons, regardless of
the beacon interval.
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@3314 0192ed92-7a03-0410-a25b-9323aeb14dbd
Section mismatch checker in the forthcoming Linux 2.6.25 complains about
ath_pci_drv_id and recommends to rename it. Rename ath_pci_drv_id to
ath_pci_driver, which is consistent with other PCI drivers.
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@3312 0192ed92-7a03-0410-a25b-9323aeb14dbd
* Based on patches from nbd
* av_beacon_alloc code has been reimplemented using atomic bit operations; I think this is safe for SMP
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@3310 0192ed92-7a03-0410-a25b-9323aeb14dbd
dropped frames a little more consistent and
fixes an issue in hardstart where I was being
too aggressive about requeing.
I thought the slab allocator would stop handing
out skbuff instances before the system was out
of free memory based upon some kind of limit on
the number of buffers, but this was nonsense.
Therefore, it is NOT a good idea to requeue
when skb allocation fails because this pretty
much basically means we are out of free memory.
Keeping the packets we do not have the RAM to
send and then stopping the transmit queue has
very bad side effects under low/zero memory
conditions.
Symptoms may include OOM killer and panics,
among other things.
It's better to drop tx packets when there's no
more RAM than to lock all the tx packets up
in the transmit queue. Even if more skb
become freed up in the slab cache due to
transmissions already in progress, evil things
still happen.
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@3308 0192ed92-7a03-0410-a25b-9323aeb14dbd
This changes some iwpriv names, because I believe they will be more readily usable and understandable
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@3279 0192ed92-7a03-0410-a25b-9323aeb14dbd
HAL_INT_RX, HAL_INT_RXPHY, or HAL_INT_SWBA. In
such cases, we will collect the earliest hw tsf
sample we can get and this is passed into send_beacon
and the trigger processing helper functions.
This cuts the number of hardware TSF reads by a
factor of three per interrupt. We should be reading
the hardware TSF now once on interrupt entry, and
once in the second part of trigger processing when
we correct for hardware TSF rollover.
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@3271 0192ed92-7a03-0410-a25b-9323aeb14dbd
Move one hw tsf read into a conditional block as it is only used in a debug printout.
Replace one hw tsf read with a reference to the existing hw tsf in the buffer (which is more accurate anyway).
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@3270 0192ed92-7a03-0410-a25b-9323aeb14dbd
symbol names to their corresponding HAL function names (ah.h)
- Add a script to generate a SED script from the current HAL (ah.h)
that will replace instances of obfuscated names with the correct names.
- Add the SED script for cleaning up text containing obfuscated function names for the current HAL
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@3244 0192ed92-7a03-0410-a25b-9323aeb14dbd
in "11.1.2.2 Beacon generation in an IBSS" the IEEE802.11 standard says, each
STA should...
"b) Calculate a random delay uniformly distributed in the range between zero
and twice aCWmin × aSlotTime,"
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@3232 0192ed92-7a03-0410-a25b-9323aeb14dbd
Since r3076 ath_tx_capture no longer exists, and ath_capture no longer frees the SKB because it was a confusing place to do that. The reverted changeset has been tested to cause leaks, and once reverted the code operates correctly.i
*fwap*
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@3222 0192ed92-7a03-0410-a25b-9323aeb14dbd
ath_tx_capture, for each frame in the list we have already freed the skb but the original pointer to
the list of skbs is still in bf->bf_skb. So when bf is returned with ath_return_txbuf it tries to
double-free, resuilting a spew of backtraces/warnings when reference debugging is enabled.
Not to mention the fact that the skb is double-freed.
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@3220 0192ed92-7a03-0410-a25b-9323aeb14dbd
whether or not beacons were going to be reconfigured. Add logic so that
any time a STA on an AP mode radio comes back to RUN state we will
reconfigure beacons. Also reconfigure beacons on AP return to RUN state
in a AP+STA configuration. In this AP+STA mode, we need to be careful to update the beacons any time a STA or AP comes online to make sure the APs start beaconing. This fixes issues where beacons are not restarted correctly after the station returns from SCAN/ASSOC/AUTH/INIT to RUN state while APs are pending completation of the STAs activity.
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@3213 0192ed92-7a03-0410-a25b-9323aeb14dbd
On requeue, stop the kernel queue and flag the fact that we stopped it in sc_dev_stopped so that it can be restarted automatically when resources become available. This is already done in the case of taking a txbuf, but can/should be done for every code path leading to requeuing.
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@3208 0192ed92-7a03-0410-a25b-9323aeb14dbd
avoid hitting the size limitations of wireless extensions. wext is a real pain.
Now the values are stored in ic_chan_non_occupy[] instead.
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@3207 0192ed92-7a03-0410-a25b-9323aeb14dbd
* Lots of coding style and formatting, tedious and annoying
* Quite a few bad merges as far as I can tell; at least one of which was serious - reintroduced a leak
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@3202 0192ed92-7a03-0410-a25b-9323aeb14dbd
hard_start_xmit() functions must either return NETDEV_TX_OK or
NETDEV_TX_BUSY (they might also return negative errno values as well,
like -ENETDOWN)
Correct a small missing static reported by sparse
This revert part of r3075
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@3123 0192ed92-7a03-0410-a25b-9323aeb14dbd
When hardware queue is full, stop the linux device queue.
This fix a bug where CPU was 100% in softirq context
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@3122 0192ed92-7a03-0410-a25b-9323aeb14dbd
* Use roundup() to calculate the next intval boundary for nexttbtt. This method is more simple to understand and computationally simpler.
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@3099 0192ed92-7a03-0410-a25b-9323aeb14dbd
* Use roundup() to calculate the next intval boundary for nexttbtt. This method is more simple to understand and computationally simpler.
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@3095 0192ed92-7a03-0410-a25b-9323aeb14dbd
from the madwifi-dfs branch (benoit) r3086
Added MAC_FMT & MAC_ADDR macros that can be used as a drop-in
replacement for ether_sprintf(). Fixed some use of ether_sprintf() that
uses the same static buffer twice.
- Without the functional changes.
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@3091 0192ed92-7a03-0410-a25b-9323aeb14dbd
It's called with NULL vap in several places, and the old
ath_beacon_config() had such protection. Without this check, the kernel
panics after association in managed mode.
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@3064 0192ed92-7a03-0410-a25b-9323aeb14dbd
IRQF_SHARED appeared in Linux 2.6.18. Define it in compat.h for older
kernels. Stop using obsolete SA_SHIRQ in if_ath_ahb.c and if_ath_pci.c
This fixes compilation of if_ath_ahb.c for Linux 2.6.22 and newer.
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@3041 0192ed92-7a03-0410-a25b-9323aeb14dbd
They are of no use, because they are not updated as the developmnet
continues. It's better to have a single version for MadWifi.
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@3039 0192ed92-7a03-0410-a25b-9323aeb14dbd
branch to trunk. especially notable is, that due to improved nexttbtt (next
"target beacon transmit time") calculation we will now get the correct backoff
behaviour for beacons, resulting in only one beacon per beacon interval (the
current head version will send N beacons for N stations in the beacon interval
because they are poorly synchronized). also because of the better timer
synchronization there is no more time lag of up to 1 minute until we see
beacons after a merge. thanks to benoit for figuring that out!
the difference between this patch and the version in the dfs-branch is that it
still uses self linked descriptors and uses the SWBA interrupt only for
updating the nexttbtt. this is necessary to recognize HW merges, for which we
don't get any notification by the hardware (see the thread "IBSS testing" on
this list for more details).
also i tried to clean up a bit, use a more descriptive function name for timer
updates (ath_beacon_update_timers instead of ath_beacon_config) and generally
better distingush between a HW merge and a SW merge.
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@3027 0192ed92-7a03-0410-a25b-9323aeb14dbd
i verified that 0 rates also create the same problem in G mode so i removed the reference to A bands.
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@3015 0192ed92-7a03-0410-a25b-9323aeb14dbd
* Remove comments that merely duplicate the code (i.e., they don't provide information on Why or explain the opaque)
* Rename some variable for consistency
* Line length -> 80 chars
* Reorder non-dependent code blocks in a function to make it more readable (at least to me)
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@3003 0192ed92-7a03-0410-a25b-9323aeb14dbd
ARRAY_SIZE is present in all kernel versions we support. Use it instead
of other definitions. Define ARRAY_SIZE in the userspace tools as well.
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@2988 0192ed92-7a03-0410-a25b-9323aeb14dbd
Split skb and node debug flags - one ending in _ref tracks minute changes to reference counts, while the non _ref flag trackes more coarse grained debug events.
Fix a node reference leak before skb_orphan detected by the debug destructor.
Fix debug flag display order for athdebug and 80211debug so that the shorter
name comes first and matches, otherwise only the _ref flags match.
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@2964 0192ed92-7a03-0410-a25b-9323aeb14dbd
ath_key_alloc() was using pointer subtraction to determine the group key
index. This is not optimal, as it's done by dividing a pointer-sized
integer by a number that is generally not a power of two.
Since there are only 4 (IEEE80211_WEP_NKID) keys to try, it's easier to
try them all in a loop. It also makes the code more reliable, as it now
detects the cases when the key pointer is within the valid range, but
doesn't point to the beginning of an array element.
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@2954 0192ed92-7a03-0410-a25b-9323aeb14dbd
Preprocessor flag now used to toggle printk output prior to blocking on
a spinlock.
There is of course a chance that the spinlock will be taken between the time we check it and print our little message, but this should still be of some assistance troubleshoot locking mistakes that come up in the future.
If you turn on the flag, you can find out what locks are being contended for and from which functions... i.e. ath_intr fights for a particular lock, etc.
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@2944 0192ed92-7a03-0410-a25b-9323aeb14dbd
Pointers don't fit u_int32_t on 64-bit systems. Besides, conversion
from pointers to integers and vice versa should be generally avoided.
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@2906 0192ed92-7a03-0410-a25b-9323aeb14dbd
* Counters for total outstanding instances for each resource type (skb, ath_node and ath_buf)
* One pair of acquisition/release functions per resource type in unlocked and one in locked
* Adds some more _debug versions of functions in the call chain that acquire/release resources so that the original func/line in the driver as well as the func/line that affected the resource use can be shown in the trace. Intermediate stack frames aren't necessary to trace the leaks.
* Changes naming convention for "lock-required" functions to suffix _locked for the versions that expect locking, to be consistent with some other places in the code.
* Consolidate debug messages to the helper functions that actually affect the reference count or acquire/release a resource
* Additional sanity checks and leak detection (esp for detecting node ref leaks through skb)
* skb references are nulled out by the new sbk unref/free function.
I've tested these changes extensively and found lots of cases where we didn't get enough node references when cloning skbuff, and where the kernel drops packets due to performance issues and leaks our node references.
With these changes and the tracing enabled I have verified that:
* TX BUF: tx buffers always go down to zero when the tx queue is done, and you can watch tx queue usage ratio go up and down again as the driver is working. There are no leaks here at the moment, although there *are* some in the madwifi-dfs branch during CAC at the moment.
* skbuff leaks in all the common flows are fixed. We were leaking node references in a lot of places where kernel was dropping skb's due to congestion and we were failing to increment node references when cloning skbuffs. These are now detected, as are skbuffs that are reaped by the kernel while still holding a node reference.
* the ath_node count works correctly and on an idle system we get about 5 references per station table node, with 2 node instances per VAP. One for the bss and one for the node in the station table, I believe. The ath_node count goes up and down but always lands back at the stable number based on the vaps you have configured and the number of actual stations in the station table. The point here is that it's pretty constant what you will see over time, despite excessive node creation/release in our code during input (esp input_all). Thank god for the slab allocator.
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@2902 0192ed92-7a03-0410-a25b-9323aeb14dbd
catch messages that occur during VAP startup (before you can possibly invoke
80211debug).
Support for shared debug flags where they take effect across all devices
(i.e. shared flags in athdebug), and when they take effect across all VAPs
(i.e. shared flags in 80211debug).
Some additional debugging flags, including some in preparation for more leak
detection code to be merged shortly as well as some from madwifi-dfs.
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@2888 0192ed92-7a03-0410-a25b-9323aeb14dbd
* Enhance readability of some functions WRT node referencing by putting blocks around areas where node references are held
* Various bits of consistency of the usage of node references; always unreference them at the point at which they are no longer used. Don't reuse variables implicitly.
* Significant overhaul of the usage of bf_node and ieee80211_cb.ni. Always hold a reference with aforementioned variable and always unreference it properly.
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@2839 0192ed92-7a03-0410-a25b-9323aeb14dbd
Linux 2.6.24 introduces a sysctl checker that disallows using arbitrary
ctl_name values for sysctl entries. The entries with non-standard
values should set ctl_name to CTL_UNNUMBERED.
On the other hand, using consistent non-zero ctl_name values is required
for older kernels (tested on Linux 2.4.33.3 from Slackware 11).
Thus the solution is to use CTL_UNNUMBERED on Linux 2.6.24+ and the
original values in the older kernels. Move CTL_AUTO and DEV_ATH to
include/compat.h and define them correspondingly.
When copying sysctl entries, check procname for being non-zero, since
ctl_name is zero (CTL_UNNUMBERED) on Linux 2.6.24 and newer.
For the same reason, don't use ctl_name to distinguish entries in
ath_sysctl_template, use extra2 instead.
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@2814 0192ed92-7a03-0410-a25b-9323aeb14dbd
Linux 2.6.24 has a sysctl table checker that rejects the tables
presented by the driver. Make the code survive this condition. Free
devname (or dev_name) before freeing the table. Don't do it at the
unregister time, it's too late.
The sysctl tables still need to be fixed.
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@2795 0192ed92-7a03-0410-a25b-9323aeb14dbd
SET_MODULE_OWNER() is not defined in 2.6.24, but it's already a no-op in
2.6.23. Make sure that it will be a no-op in 2.6.23 and newer kernels.
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@2754 0192ed92-7a03-0410-a25b-9323aeb14dbd
This value is roughly around the value that other devices are setting. The previous value appeared to be stupidly high.
This may solve some interactivity problems and machine performance issues.
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@2732 0192ed92-7a03-0410-a25b-9323aeb14dbd
