summaryrefslogtreecommitdiff
path: root/sys/dev
AgeCommit message (Collapse)Author
2012-10-31Some fixes for the receive filter handling.Brad Smith
ok sf@
2012-10-30introduce specialized register read and write functionsMike Belopuhov
oce_{read,write}_{cfg,csr,db} for different pci bars
2012-10-30backout previous over the m_freem contention concernsMike Belopuhov
2012-10-30Bump the default number of descriptors.Brad Smith
ok mikeb@
2012-10-29prefer to return oce_cmd and save on the error variableMike Belopuhov
2012-10-29shorten fwcmd to cmdMike Belopuhov
2012-10-29merge oce.c into if_oce.c and rename oce{reg,var}.h to if_oce{reg,var}.hMike Belopuhov
2012-10-29don't split mq sge paddr into low and high partsMike Belopuhov
2012-10-29rename oce_fw to oce_cmd and shorten MBX_SUBSYSTEM_* defines to SUBSYS_Mike Belopuhov
2012-10-29rearrange the function layout a bitMike Belopuhov
2012-10-29oce_get_fw_config is useless; ditch itMike Belopuhov
2012-10-29Add support for Elantech touchpads to pms(4). This allows the X synaptics(4)Stefan Sperling
input driver to configure these touchpads properly, e.g. toggle tap-to-click, two-finger scrolling, edge-scrolling, etc. So far these pads were working only in PS/2 mouse compatibility mode. Support for up to hardware version 3 has been implemented. But only version 3 has been tested so versions 1 and 2 remain disabled via #ifdef notyet until someone tests (bugs in this code could seriously mess up the mouse and I don't want to end up shipping such bugs in 5.3). help and ok mpi@ shadchin@
2012-10-28Add braces to some code in ixgbe_setup_vlan_hw_support().Brad Smith
ok mikeb@
2012-10-27Implememnt AP side power saving for acx(4). Tested on acx11.Claudio Jeker
With some input and OK from kettenis@ and OK sthen@
2012-10-27kill some unused definitionsMike Belopuhov
2012-10-26oce_first_mcc_cmd doesn't return anything of value so make it void;Mike Belopuhov
shorten it to oce_first_mcc while here
2012-10-26simplify the way we collect statisticsMike Belopuhov
2012-10-26remove stuff that either is no longer or was never usedMike Belopuhov
2012-10-26cleanup oce_init_fw; use less bitfieldsMike Belopuhov
2012-10-26rename oce_config_nic_rss to oce_config_rss and merge oce_rss_itbl_init in;Mike Belopuhov
cleanup oce_set_promisc and oce_config_vlan function arguments, a bunch of defines and prototypes while at it.
2012-10-25make oce_mbox_{dispatch,wait} more readable and use less bitfieldsMike Belopuhov
2012-10-25simplify/unify writes to the rx and tx doorbell registersMike Belopuhov
2012-10-25don't use bitfields in oce_pci_alloc; kill some unused structuresMike Belopuhov
2012-10-25cleanup interrupt register definesMike Belopuhov
2012-10-25Fix a hilarious bug where oce_set_promisc wouldn't zero out theMike Belopuhov
command buffer it sent to the card but an incorrect assignment of the DMA address pointing to this buffer would offset the problem. Once I've fixed an incorrect assignment oce_init started to fail. Turns out that oce_set_promisc is about the only function whose command size is larger than the payload of the standard command buffer and needs an external chunk of DMA memory.
2012-10-25simplify oce_arm_eq and oce_arm_cqMike Belopuhov
2012-10-22- Provide flow control statusBrad Smith
- Use ANAR_PAUSE_TOWARDS define - Only advertise flow control if the MAC sets MIIF_DOPAUSE ok jsg@
2012-10-22Add flow control support.Brad Smith
ok jsg@
2012-10-22Provide flow control status.Brad Smith
ok jsg@
2012-10-22Add flow control support.Brad Smith
Tested with Apple/Sun gem(4) with and without flow control (Sun ERI/Apple GEM with FastE PHY). ok kettenis@
2012-10-22Rewrite the receive filter handling code and cleanup the ioctl bits.Brad Smith
Tested by and ok stsp@
2012-10-22Fix the multicast filter full size check.Brad Smith
ok mikeb@
2012-10-21repair pci powerstate handling. The extra timeouts were not needed.Theo de Raadt
Furthermore, all the extra powering up and down caused resumes to not work. ok ratchov
2012-10-20driver xxactivate() functins are run even if a driver xxattach() hasTheo de Raadt
"failed"; no indication is given up to the autoconf framework since xxattach() returns void. Until this situation is improved, there are a handful of drivers which must use driver-specific checks in xxactivate() to see if the xxattach() function succeeded or failed. Add such a check here. ok kettenis
2012-10-20Implement transmit DMA segments. The descriptor usage is from FreeBSD,Chris Cappuccio
which contradicts the datasheet (but actually works). Also, use a segment (instead of a new mbuf and data copy) to zero-pad small packets, modeled after if_myx. ok sthen@, gerhard@, henning@
2012-10-20Init values for the AR9485 were for version 1.0 of this chip, which accordingStefan Sperling
to Atheros Linux developers was never sold. So update initvals to what Linux is using for the 1.1 generation. Because the serdes values are written to different registers on the AR9485 this involves tweaking the serdes init code for all athn(4) chip families. This commit doesn't make AR9485 devices work yet but is a step in the right direction. Tested by krw, kettenis, and Andrew Ngo. ok kettenis@
2012-10-20Fix hardware kill switch detection for the ar9300 chip family. The driver wasStefan Sperling
checking the wrong bits of eeprom to determine rfkill switch pin and polarity, and was reading the wrong register to determine rfkill gpio pin state. ok kettenis@
2012-10-19andre.s@list.ru has a laptop which does not bring the azalia out ofTheo de Raadt
D3 state, so the attach funciton must do this. ok kettenis
2012-10-19small cleanups and output prettyingTheo de Raadt
ok yuo matthieu
2012-10-19Simplify xl_iff_90x() a bit and only check ac->ac_multicnt since ac_multicntBrad Smith
is also bumped for multicast ranges. ok mikeb@
2012-10-18Convert a number of old private copies of code which predatesTheo de Raadt
pci_set_powerstate() to using it instead. Many of these chunks of code had bugs in them, especially missing delay() calls. Some of them were doing things our PCI subsystem is now responsible for handling. If you have any of the affected devices, please keep an eye out for regressions. ok kettenis
2012-10-18Use pci_set_powerstate() unconditionally for all chip variants, ratherTheo de Raadt
than a kludged together chunk of code for just one chip, which happens to be entirely wrong. We need to watch for regressions on cardbus, though. ok kettenis
2012-10-18make oce_arm_{eq,cq} functions look like the other queueMike Belopuhov
manupulation functions in this file; do some minor style cleanup while here.
2012-10-17Swap arguments to wdog_register() since it is nicer, and prepareTheo de Raadt
wdog_shutdown() for external usage.
2012-10-17Narrow the spltty/splx pair in com_activate to only enclose the minimumTheo de Raadt
operation which requires it.
2012-10-17cleanup more of the driver soft-state when suspending; in certainTheo de Raadt
async operational modes a timeout and some requests on a queue were still active helps a machine mlarkin has ok miod shadchin mlarkin
2012-10-17AH_BYTE_ORDER and AH_BIG_ENDIAN are holdovers from the old FreeBSDJonathan Gray
code, we test but don't define them. Replace with _BYTE_ORDER and _BIG_ENDIAN to make the iv show up properly for big endian bpf consumers. ok stsp@ reyk@
2012-10-17#if NWSDISPLAY > 0 not #if WSDISPLAY > 0Jonathan Gray
ok miod@
2012-10-16Remove uneeded <net/bpf.h> include which was never used as itJonathan Gray
was protected by #if NBPFILTER > 0 and "bpfilter.h" wasn't included.
2012-10-15eliminate a couple of gotosMike Belopuhov
generated by cgit v1.2.3 (git 2.25.1) at 2024-12-21 05:57:50 +0000