summaryrefslogtreecommitdiffstats
path: root/bsd_eth_drivers (unfollow)
Commit message (Collapse)AuthorFilesLines
2011-07-202011-07-20 Till Straumann <Till.Straumann@TU-Berlin.de>Till Straumann5-39/+241
* if_em/Makefile.am, if_em/e1000_osdep.h, if_em/e1000_osdep.c, if_em/if_em.c: added e1000_osdep.c which implements all os-dependent parts in a bsdnet-agnostic fashion. Hence, the low-level driver can be used w/o bsdnet. FIXES: dependency on bsdnet initialization; PCI bus header was not initialized when not used via bsdnet (lanIpBasic driver crashed).
2010-07-232010-07-23 Till Straumann <Till.Straumann@TU-Berlin.de>Till Straumann2-0/+16
* libbsdport/devicet.c: Initialize libbspExt.
2010-06-30*** empty log message ***Till Straumann1-0/+12
2010-06-302010-06-30 Till Straumann <Till.Straumann@TU-Berlin.de>Till Straumann1-8/+8
* libbsdport/bus.h: PPC's inlines from libcpu/io.h still have un-intuitive argument types (e.g., unsigned instead of uint32_t). Added additional 'iotype' arg to BUS_SPACE_DECL macro (ppc only) so that we can cast to the correct type as expected by libcpu/io.h. Hopefully that doesn't bring us into C99 alias hell.
2010-06-302010-06-30 Till Straumann <Till.Straumann@TU-Berlin.de>Till Straumann1-1/+1
* libbsdport/devicet.c: Silence compiler warnings by casting (char) argument passed to 'ctype.h' functions [toupper + friends] -> (unsigned char) as these functions expect 'int'.
2010-03-072010-03-07 Till Straumann <Till.Straumann@TU-Berlin.de>Till Straumann2-5/+7
* links.am: POSIXified sed commands for sake of portability.
2010-02-122010-02-12 Till Straumann <Till.Straumann@TU-Berlin.de>Till Straumann2-22/+5
* libbsdport/Makefile.am: Removed obsolete (and commented) stuff.
2009-10-202009-10-20 Till Straumann <Till.Straumann@TU-Berlin.de>Till Straumann3-1/+19
* libbsdport/libbsdport.h, if_em/e1000_osdep.h: We need __BSD_VISIBLE defined. Some things in the rtems headers changed with 4.10 (can't pinpoint the details). We now #define __INSIDE_RTEMS_BSD_TCPIP_STACK__ before explicitly including <rtems/rtems_bsdnet.h> and this seemed to fix things for 4.10.
2009-10-202009-10-20 Till Straumann <Till.Straumann@TU-Berlin.de>Till Straumann1-0/+15
* libbsdport/libbsdport_post.h: added missing declaration of m_defrag() (Chris replaced the dummy #define a while ago and provided an implementation instead [thanks!] but we also need a declaration).
2009-10-202009-10-20 Till Straumann <Till.Straumann@TU-Berlin.de>Till Straumann1-1/+1
* libbsdport/ifmedia.c: fixed 'unused variable' warning by removing unused var.
2009-10-202009-10-20 Till Straumann <Till.Straumann@TU-Berlin.de>Till Straumann1-0/+2
* if_re/if_re.c: fixed 'unused variable' warning by commenting via #ifndef __rtems__ (since code which uses the 'msi_disable' var is commented the same way).
2009-10-202009-10-20 Till Straumann <Till.Straumann@TU-Berlin.de>Till Straumann1-1/+16
* if_pcn/if_pcn.c: fixed 'type-punned pointer' warning. Use 'memcpy' to copy 2 bytes to a short. This requires another #ifdef __rtems__, however.
2009-10-202009-10-20 Till Straumann <Till.Straumann@TU-Berlin.de>Till Straumann2-1/+32
* if_bge/if_bge.c: bge_intr must be passed as a 'handler' to bus_setup_intr(). However, the irq_check_dis and irq_enable methods seem to still be skeletons so this driver won't work... (I have no hardware or emulation available to test).
2009-10-202009-10-20 Joel Sherrill <joel.sherrill@OARcorp.com>Joel Sherrill2-2/+4
* libbsdport/alldrv.c: Revert.
2009-10-202009-10-20 Joel Sherrill <joel.sherrill@OARcorp.com>Joel Sherrill3-1/+17
* if_bge/if_bge.c, libbsdport/alldrv.c: Add local changes (untested).
2009-10-202009-10-20 Chris Johns <chrisj@rtems.org>Chris Johns11-875/+3980
* bsd_eth_drivers/libbsdport/sysbus.cL Fixed the print message for no IRQ handler. * bsd_eth_drivers/libbsdport/misc.c: Add m_defrag and what else it needed. The RL driver needs it. * bsd_eth_drivers/libbsdport/libbsdport_post.h: Remove the define for m_defrag. * bsd_eth_drivers/libbsdport/libbsdport.h: Add if_maddr_rlock and if_maddr_runlock. * bsd_eth_drivers/libbsdport/bus.h: Add the bus_space_write_stream_4 call. * bsd_eth_drivers/libbsdport/libbsdport_post.h, bsd_eth_drivers/libbsdport/alldrv.c, bsd_eth_drivers/if_re/Makefile.am: Add RL driver support. * bsd_eth_drivers/if_re/if_re.c, bsd_eth_drivers/if_re/if_rlreg.h: Updated to the lastest version. * bsd_eth_drivers/if_re/if_rl.c: New. This is for the Realtek 8129/8139 PCI NIC while the RE driver does only the 8139C+ version.
2009-10-122009-09-12 Chris Johns <chrisj@rtems.org>Chris Johns2-1/+14
* links.am: Fixed the links so they work with absolute paths.
2009-08-152009-08-15 Till Straumann <Till.Straumann@TU-Berlin.de>Till Straumann2-1/+6
* if_re/Makefile.am: added 'if_rlreg.h' to SOURCES so that it is 'distributed'.
2009-08-152009-08-15 Till Straumann <Till.Straumann@TU-Berlin.de>Till Straumann2-2/+7
* if_em/31000_osdep.h: added missing 'void' return type of __out_le32() inline function.
2009-08-062009-08-06 Till Straumann <Till.Straumann@TU-Berlin.de>Till Straumann2-0/+15
* libbsdport/rtems_callout.c: fixed possible race condition. callout_stop() must check again from critical/protected section of code if callout is still on the list/active. Otherwise, the callout-task could have executed and removed the callout between callout_stop() checking the p_prev pointer and entering the critical section.
2009-08-052009-08-05 Till Straumann <Till.Straumann@TU-Berlin.de>Till Straumann4-99/+152
* if_em/Makefile.am, if_em/e1000_osdep.h, if_em/if_em.c: Changed 'e1000_osdep.h' to provide stdalone I/O methods for select architectures (x86 and PPC -- others fall back on libbsdport/bus.h). This lets the low-level driver (everything except for 'if_em.c/if_em.h') API be independent of BSD networking and libbsdport which is desirable since certain applications may wish to just use the low-level API for implementing dedicated 'raw-ethernet' drivers for BSD-independent, proprietary GigE communication.
2009-08-052009-08-05 Till Straumann <Till.Straumann@TU-Berlin.de>Till Straumann2-2/+7
* libbsdport/alldrv.c: reactivate weak aliases for all known/supported drivers. This magic allows the user to just link the desired (sub-)set of drivers.
2009-08-052009-08-05 Till Straumann <Till.Straumann@TU-Berlin.de>Till Straumann2-0/+6
* libbsdport/libbsdport.h: define _KERNEL only if not defined already (e.g., from Makefile).
2009-06-042009-06-04 Joel Sherrill <joel.sherrill@oarcorp.com>Joel Sherrill3-0/+8
* libbsdport/alldrv.c, libbsdport/libbsdport_api.h: Add include <stdio.h> for FILE *.
2009-05-13 2009/05/12 (TS)Till Straumann2-1/+3
- bugfix: bus_write_2() must map to bus_space_write_2 (not 4)
2009-04-23 - had lost a few bits during the merge...Till Straumann1-0/+2
2009-04-23 - had lost a few things during the merge :-(Till Straumann2-0/+10
2009-04-23 - resolved conflictTill Straumann1-4/+0
2009-04-23 - ported to RTEMS.Till Straumann4-0/+219
2009-04-23 - added missing ETHER_SIOCMULTIFRAG() macro to ioctl (#ifdef __rtems__)Till Straumann1-0/+4
2009-04-23 - removed 'bootverbose' -- now defined by libbsdport.Till Straumann1-2/+0
2009-04-23 - removed definition of 'bootverbose' (now in libbsdport)Till Straumann1-7/+16
- removed definition of 'bswap32' (now in libbsdport) - removed definition of 'callout_active' (now in libbsdport) - removed definition of 'callout_pending' (now in libbsdport) - removed bus_dmamap_load() -- I believe the version from libbsdport should work (but setting the bus addr to zero probably not). - added (__rtems__) specific code to set multicast address filter - added ETHER_SIOCMULTIFRAG() macro (__rtems__ specific) to ioctl.
2009-04-23 - echo links that are to be made (for diagnostic purposes).Till Straumann1-2/+3
2009-04-23 - added FXP driverTill Straumann1-1/+1
2009-04-23*** empty log message ***Till Straumann1-0/+2
2009-04-23 - make entries in default driver table all weak aliases toTill Straumann1-27/+48
a sentinel. If the user decides only to link a subset of driver the probe algorithm skips entries matching the sentinel. - added more diagnostic messages which can be enabled by setting 'libbsdportAttachVerbose' to nonzero. This is quite helpful. - added the ugliest of all hack -- i386 pcibios API returns wrong number for bus count :-( [the max bus number instead of the count]. We simply add 1 to this number...
2009-04-23 - make entries in default driver table all weak aliases toTill Straumann1-2/+25
a sentinel. If the user decides only to link a subset of driver the probe algorithm skips entries matching the sentinel.
2009-04-23 - added ifmedia_removeall().Till Straumann1-0/+10
2009-04-23 - added le32enc()Till Straumann1-2/+29
- added definition of PCIR_CACHELNSZ - added definition of PCIM_CMD_MWRICEN - added pci_get_revid() - moved mii_mediachg() macro-stub to libbsdport_post.h - added definition for 'bootverbose'
2009-04-23 - added bus_alloc_resources()Till Straumann2-4/+115
- added bus_release_resources() - added dummy handle for giant mutex. - added resource_int_value() stub (returns ENOENT). - added resource_long_value() stub (returns ENOENT). - added bus_read_1(), bus_read_2(), bus_read_4() bus_write_1(), bus_write_2(), bus_write_4(). - honour BUS_DMA_ZERO flag. - changed 'vaddr' arg from caddr_t to void*
2009-04-23 - callout_stop() and callout_reset() actually must return an 'int' valueTill Straumann2-7/+63
indicating whether a callout was pending. - added callout_active(), callout_pending(), callout_deactivate().
2009-04-23 - added definition for MA_NOTOWNEDTill Straumann1-1/+2
2009-04-23 - fixed ETHER_SIOCMULTIFRAG() macro. Need to reset 'e' (error) to zero ifTill Straumann1-2/+15
ether_addmulti()/ether_delmulti() return ENETRESET (which is acually an OK status). - moved mii_mediachg() dummy from libbsdport.h to libbsdport_post.h
2009-04-23 - added FXP driverTill Straumann1-0/+2
2009-04-23 - added 'bootverbose' (in new 'misc.c' file)Till Straumann3-1/+60
- added rtems_mii_phy_probe(). This is a helper albeit RTEMS specific. Porting the whole MII/media stuff is still too much ATM.
2009-04-23 - reverted AC_PREREQ to 2.62Till Straumann5-2800/+0
- removed files that are generated by autreconf after checkout and added them to .cvsignore
2009-04-23 - importing original version I had sent Joel on 20090214baseTill Straumann5-0/+2800
(SLAC tag 'joel-bsd-drivers-20090214')
2009-04-22 - importing updated version from SLAC as of 20090422Till Straumann27-64/+4033
2009-02-202009-02-20 Joel Sherrill <joel.sherrill@oarcorp.com>Joel Sherrill7-2/+7563
* configure.ac, bsd_eth_drivers/Makefile.am, bsd_eth_drivers/libbsdport/alldrv.c, bsd_eth_drivers/libbsdport/libbsdport_api.h: Add first cut at Broadcomm BCM57xx driver. * bsd_eth_drivers/if_bge/.cvsignore, bsd_eth_drivers/if_bge/Makefile.am, bsd_eth_drivers/if_bge/if_bge.c, bsd_eth_drivers/if_bge/if_bgereg.h: New files.
2009-02-20 - use separate #ifdef __rtems__ branch when using bus_setup_intr() becauseTill Straumann1-1/+6
APIs are really different. - must not use INTR_FAST anymore since this will (deliberately) produce a compile-time error (see <sys/bus.h>)
generated by cgit v1.2.3 (git 2.25.1) at 2024-04-28 10:10:20 +0000