| Age | Commit message (Collapse) | Author | |
|---|---|---|---|
| 2002-05-13 | When collecting the completed tx descriptors, dmasync all of them | Artur Grabowski | |
| before reading, not just the first one. jason@ ok. | |||
| 2002-05-13 | Don't leak memory when bus_dmamap_load_mbuf fails. | Artur Grabowski | |
| 2002-05-10 | Zap an unsued define. | Artur Grabowski | |
| 2002-05-10 | What we store on all archs just before the rfa is a pointer | Artur Grabowski | |
| to a bus_dmamap_t, not a bus_dmamap_t (just a cosmetic change, bus_dmamap_t is a pointer on all archs, but this was confusing) | |||
| 2002-03-15 | Missing splx | Artur Grabowski | |
| 2002-03-14 | First round of __P removal in sys | Todd C. Miller | |
| 2002-02-15 | Don't cast nonexistent return value from splx to (void). ok art@ | Thomas Nordin | |
| 2001-11-06 | Replace inclusion of <vm/foo.h> with the correct <uvm/bar.h> when necessary. | Miod Vallat | |
| (Look ma, I might have broken the tree) | |||
| 2001-11-05 | Switch everything to the new bus_dmamap_sync API. | Artur Grabowski | |
| Most work by Wilbern Cobb <vedge@csoft.org> with some fixes from me, mickey@ and drahn@. | |||
| 2001-09-20 | occured->occurred | Mike Pechkin | |
| idea from deraadt@ via NetBSD millert@ ok | |||
| 2001-09-17 | Remove all of the last vestiges of vtophys (namely RFA's): | Jason Wright | |
| o stash bus_dmamap_t in the mbuf cluster (alignment fudge is now 2+sizeof(bus_dmamap_t) o borrow FXP_RXMAP_{GET|PUT} macros from NetBSD o move a few definitions around to make it all work | |||
| 2001-08-27 | get direction of read/write correct in calls to bus_dmamap_sync() | Jason Wright | |
| 2001-08-25 | deal with 5 arg form of bus_dmamap_sync() if available. | Jason Wright | |
| 2001-08-10 | - make transmit use linked lists not indexed arrays (more like it was when I ↵ | Jason Wright | |
| started) - also, only bus_dmamap_unload() loaded maps in fxp_stop() | |||
| 2001-08-09 | - scb_wait() is now time based (not cpu speed based) | Jason Wright | |
| - rearrange multicast setup to be part of init - convert cu operations to bus_dma(9) (still have to do RFA ops). | |||
| 2001-08-03 | Fix logic for detecting 82558 and newer chips in pci; | Chris Cappuccio | |
| This driver allows large frames for those chips so set IFCAP_VLAN_MTU for 82558+ | |||
| 2001-07-18 | fix vtophys usage, mainly ppc-required; art@ ok and tried to ↵ | Michael Shalayeff | |
| bribe^H^H^H^H^Hmotivate me to bussydma 'em too (; | |||
| 2001-06-27 | ALTQ'ify network drivers. | Kenjiro Cho | |
| - use the new queue macros. - use IFQ_POLL() to peek at the next packet. - use IFQ_IS_EMPTY() for empty check. - drivers should always check if (m == NULL) after IFQ_DEQUEUE(), since it could return NULL even when IFQ_IS_EMPTY() is FALSE under rate-limiting. - drivers are supposed to call if_start from tx complete interrupts (in order to trigger the next dequeue under rate-limiting). | |||
| 2001-06-23 | ether_input_mbuf(). | Federico G. Schwindt | |
| 2001-06-13 | From NetBSD, 82562 requires a nop before a resume (tested by mickey) | Jason Wright | |
| 2001-05-27 | Put back the fix for the possible leak and fix another bug | Artur Grabowski | |
| that the fix uncovered. The tx descriptors were not initialized when allocated. The initialization was done in fxp_init, but the first thing fxp_init does is to call fxp_stop and fxp_stop expects the tx descriptors to be already initialized. How did this ever work? | |||
| 2001-05-26 | remove code which was obviously not tested. begone, sloppy code monkeys | Theo de Raadt | |
| 2001-05-24 | When releasing the tx buffers, walk through all of them, don't abort when | Artur Grabowski | |
| we find a descriptor with mb_head == NULL. It happens sometimes that we have NOPs in the tx chains and normal buffers after the NOP. | |||
| 2001-05-24 | typo | Artur Grabowski | |
| 2001-04-06 | Move offsetof define into sys/param.h | Constantine Sapuntzakis | |
| 2001-03-22 | Reformat comment. | Angelos D. Keromytis | |
| 2001-02-20 | for ethernet ifaces attach bpf from ether_ifattach; jason@, aaron@, itojun@ ok | Michael Shalayeff | |
| 2000-10-16 | Use mii_attach() directly instead of mii_phy_probe(). | Aaron Campbell | |
| 2000-09-17 | Support detach of fxp devices to allow ejection of Intel CardBus adapters. | Aaron Campbell | |
| 2000-08-02 | No need to include pci header files here. | Aaron Campbell | |
| 2000-07-20 | Convert to new timeouts. (art@ ok) | Hakan Olsson | |
| 2000-06-03 | remove fxp_ether_ioctl(), and move it's processing to fxp_ioctl(). | Jason Wright | |
| 2000-04-27 | Whoops | Chris Cappuccio | |
| 2000-04-27 | Fix SIOCSIFMTU | Chris Cappuccio | |
| 2000-04-26 | Turn on a bit (for 82558/82559 only) which disables the chip's | Chris Cappuccio | |
| builin feature to discard frames >1500 bytes (useful for jumbo frames, also 802.1Q VLANs) ok aaron@ | |||
| 2000-04-26 | After the fxp split, sizeof(enaddr) is now 4, not 6, since enaddr is a pointer | Aaron Campbell | |
| that's passed in from the bus-dependent parts. Use ETHER_ADDR_LEN so we get the whole thing; sigh@kuzirabekon.econ.nagasaki-u.ac.jp | |||
| 2000-04-18 | split fxp into bus independent and dependent (pci) parts (cardbus to come) | Jason Wright | |
 |
index : src | |
| OpenBSD base system |
| summaryrefslogtreecommitdiff |