- 29 Aug, 2009 21 commits
-
-
Stephen Hemminger authored
Signed-off-by:
Stephen Hemminger <shemminger@vyatta.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Stephen Hemminger authored
Simpler to have one place that spins and accounts for delays, this will also make the last packet be detected faster for more repeatable timing. Signed-off-by:
-
Stephen Hemminger authored
This changes how the pktgen thread spins/waits between packets if delay is configured. It uses a high res timer to wait for time to arrive. Signed-off-by:
-
Stephen Hemminger authored
The kernel ktime_t is a nice generic infrastructure for mananging high resolution times, as is done in pktgen. Signed-off-by:
-
Stephen Hemminger authored
If not using delay then no need to update next_tx after each packet sent. This allows pktgen to send faster especially on systems with slower clock sources. Signed-off-by:
-
Stephen Hemminger authored
Handle standard (and non-standard) return values in a switch. Signed-off-by:
-
Stephen Hemminger authored
netdev_alloc_skb is NUMA node aware. Also, don't exhaust atomic emergency pool. Don't want pktgen to cause OOM behaviour. Signed-off-by:
-
Stephen Hemminger authored
The if statement to test for "should a new packet be used" can be simplified. Signed-off-by:
-
Stephen Hemminger authored
Do some reorganization of transmit logic path: * move transmit queue full idle to separate routine * add a cpu_relax() * eliminate some of the uneeded goto's * if queue is still stopped, go back to main thread loop. * don't give up transmitting if quantum is exhausted (be greedy) Signed-off-by:
-
Stephen Hemminger authored
All the callers were freeing skb after stopping device. Remove unneeded forward decl. Signed-off-by:
-
Stephen Hemminger authored
Signed-off-by:
-
Stephen Hemminger authored
Don't force inlining where not needed. Gcc does better job of deciding to inline local functions. Signed-off-by:
-
Stephen Hemminger authored
A couple of minor functions can be written more compactly. Signed-off-by:
-
Ron Mercer authored
TX completions were running in a workqueue queued by the ISR. This patch moves the processing of TX completions to an existing RSS NAPI context. Now each irq vector runs NAPI for one RSS ring and one or more TX completion rings. Signed-off-by:
Ron Mercer <ron.mercer@qlogic.com> Signed-off-by:
-
Ron Mercer authored
Currently we downshift to MSI/Legacy if we don't get enough vectors for cpu_count RSS rings plus cpu_count TX completion rings. This patch allows running MSIX with the vector count that the platform provides. Signed-off-by:
-
Ron Mercer authored
Currently we have three types of RX rings. 1) Default ring - services rx_ring for broadcast/multicast, handles firmware events, and errors. 2) TX completion ring - handles only outbound completions. 3) RSS ring - handles only inbound completions. This patch gets rid of the default ring type and moves it's functionality into the first RSS ring. This makes better use of MSIX vectors since they are a limited resource on some platforms. Signed-off-by:
-
-
Petri Gynther authored
bonding: Have bond_check_dev_link examine netif_running Some network devices do not call netif_carrier_off when they are set administratively down. Have the bonding link check function also inspect the netif_running state. Ignore netif_running if the bond_check_dev_link function is called with "reporting" set, as in that case it's inspecting the capabilities of the non-netif_carrier device driver. Signed-off-by:
Petri Gynther <pgynther@google.com> Signed-off-by:
Jay Vosburgh <fubar@us.ibm.com> Signed-off-by:
-
Nicolas de Pesloüan authored
max_bonds is of type int and cannot be greater than INT_MAX. Signed-off-by:
Nicolas de Pesloüan <nicolas.2p.debian@free.fr> Signed-off-by:
-
Stephen Hemminger authored
Bonding can use compare_ether_addr() in bond_release. Signed-off-by:
-
Jay Vosburgh authored
Propogate the vlan_features of the slave devices to the bonding master device, using the same logic as for regular features. Tested by Or Gerlitz <ogerlitz@voltaire.com>, who also removed the debug logic from the original test patch. Signed-off-by:
Or Gerlitz <ogerlitz@voltaire.com> Signed-off-by:
-
- 28 Aug, 2009 19 commits
-
-
Kiran Divekar authored
Most of the places in debugfs.c are missing a NULL check on the return value of get_zeroed_page API call. Added required NULL check at appropriate places. Signed-off-by:
Kiran Divekar <kirandivekar@gmail.com> Signed-off-by:
John W. Linville <linville@tuxdriver.com>
-
Jussi Kivilinna authored
Now that cfg80211 functions are added and wext converted to use wext-compat functions, remove wext structures and disabled code. Signed-off-by:
Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by:
-
Jussi Kivilinna authored
Signed-off-by:
-
Jussi Kivilinna authored
Signed-off-by:
-
Jussi Kivilinna authored
Signed-off-by:
-
Jussi Kivilinna authored
Stats worker no longer poll stats from device anymore. It's still needed to poll device control channel for connect/disconnect events, so rename stats worker as device poller. Signed-off-by:
-
Jussi Kivilinna authored
Signed-off-by:
-
Jussi Kivilinna authored
Add cfg80211 get_station and convert SIOCGIWRATE and get_wireless_stats to cfg80211. Signed-off-by:
-
Jussi Kivilinna authored
Add cfg80211 add_key/del_key/set_default_key and convert wext to use theim. Signed-off-by:
-
Jussi Kivilinna authored
Add cfg80211 set_channel and convert wext to use it. Signed-off-by:
-
Jussi Kivilinna authored
Add cfg80211 connect functions for station and ad-hoc modes and convert wext to use theim. Signed-off-by:
-
Jussi Kivilinna authored
Random essid must be set to turn on radio when not connected. If device is in ad-hoc mode, this results 'media connect' indications with the random essid which should be ignored. Signed-off-by:
-
Jussi Kivilinna authored
Signed-off-by:
-
Jussi Kivilinna authored
Use is_zero_ether_addr() and is_broadcast_ether_addr() instead of memcmp against ffff_bssid/zero_bssid. Signed-off-by:
-
Jussi Kivilinna authored
Move link up/down work to separate functions and use local array for allocating memory for info structure instead of kzmalloc. Signed-off-by:
-
Jussi Kivilinna authored
Increase scan delay from 1 sec to 6 sec. Spec says that scan by OID_802_11_BSSID_LIST_SCAN completes in 6 seconds. Before rfkill patch too short delay was not problem as device was always active (radio on) and performing background scanning. Signed-off-by:
-
Jussi Kivilinna authored
Buffer used for bssid list might be too small. Change rndis_query_oid() to return required buffer length to caller and make rndis_check_bssid_list() resize buffer when needed. Signed-off-by:
-
Jussi Kivilinna authored
OID_802_11_BSSID_LIST_SCAN clears device's bssid list, so retrieve current bssid list from device before issuing new scan. Signed-off-by:
-
Jussi Kivilinna authored
Setting WPA keys with OID_802_11_ADD_KEY sometimes trigger instant media connect indication. These indications are extranous and should be ignored, as otherwise driver would send reassociation event to userspace which in this case is not needed. Signed-off-by:
-
