changelog

来自「开放源码实时操作系统源码.」· 代码 · 共 845 行 · 第 1/3 页

2007-06-12  Andrew Lunn  <[e-mail]>

	* src/if_i82559.c (eth_dsr): Fixed the previous change which broken
	linkage under some conditions.

2006-11-23  David Fernandez  <[e-mail]>

	* src/if_i82559.c Modifications to remove compile warnings.

2005-01-22  Andrew Lunn  <[e-mail]>

	* src/if_i82559.c Added string.h to remove compiler warnings.

2004-08-12  Jani Monoses <[e-mail]>

	* src/if_i82559.c: Fix builing with lwip.

2004-01-15  Jeff Daly  <[e-mail]>

	* src/if_i82559.c: support for large EEPROMs

2004-01-05  Gary Thomas  <[e-mail]>

	* src/if_i82559.c: Names of FLASH config types changed.

2003-02-27  Mark Salter  <[e-mail]>

	* src/if_i82559.c (eth_set_mac_address): Correct writing
	of wrong address to EEPROM.

2003-02-22  Mark Salter  <[e-mail]>

	* src/if_i82559.c: Fix endian issues with EEPROM writing.
	Use read-modify-write to set EEPROM mac address. Support
	writing to EEPROM the mac address obtained from
	CYGHWR_DEVS_ETH_INTEL_I82559_GET_ESA.

2003-02-12  Mark Salter  <[e-mail]>

	* src/if_i82559.c: Remove restriction preventing combination of
	*_GET_ESA and *_HAS_EEPROM.

2003-01-03  Gary Thomas  <[e-mail]>

	* src/if_i82559.c: 
	* cdl/intel_i82559_eth_drivers.cdl: Allow finer control over
	debug (chatter) by making control booldata.

2002-12-17  Mark Salter  <[e-mail]>

	* src/if_i82559.c (CYGHWR_DEVS_ETH_INTEL_I82559_ENDIAN_NEUTRAL_IO): New
	flag to support systems where PCI IO operations are not affected by
	CPU endianess.

2002-10-06  Andrew Lunn  <[e-mail]>

	* src/if_i82559.c (pci_init_find_82559s): Changed scope of 
	max_interrupt_handle since its needed in other places.
	* src/if_i82559.c (i82559_stop): Corrected wrong name of variable
	in a diag_printf function.

2002-07-24  Gary Thomas  <[e-mail]>

	* src/if_i82559.c (pci_init_find_82559s): 
	Allow platform to define CYGHWR_DEVS_ETH_INTEL_I82559_USE_MEMORY
	which forces use of memory instead of I/O space for device access.

2002-06-14  Gary Thomas  <[e-mail]>

	* src/if_i82559.c: 
	Need to include <pkgconf/io_eth_drivers.h> for proper configuration
	of stand-alone (polled) vs. system (interrupt driven) mode.

2002-04-09  Hugo Tyson  <[e-mail]>

	* src/if_i82559.c (i82559_stop): Free up any pending transmissions
	to prevent MBUF store leaks if the calling thread - the one that's
	shutting down the interface - priority ends up higher than (or
	same as) the network service threads.

2002-03-26  Jonathan Larmour  <[e-mail]>

	* src/if_i82559.c (wait_for_cmd_done): Don't assert if RUC_ADDR_LOAD
	command doesn't clear - on some chips it never does.

2002-03-18  Hugo Tyson  <[e-mail]>

	* src/if_i82559.c (pci_init_find_82559s): Move allocation of
	statistics pointer for each device to here, out of
	i82559_start().  It used to leak PCI window store every time the
	interface cycled down/up.  [Case 107110]

2002-02-25  Mark Salter  <[e-mail]>

	* src/if_i82559.c: Don't call PacketRxReady in standalone configuration
	unless ready to accept a packet. Limit packet delivery to one packet at
	a time in standalone configuration.

2002-02-20  Hugo Tyson  <[e-mail]>

	* include/i82559_info.h: New fields to record/control multicast
	all reception, oversized packet reception and promiscuous mode
	neatly.

	* src/if_i82559.c: Initial support for multicast address reception
	filtering.
	(i82559_restart): New routine to restart the device after fiddling
	with its setup - a tail recursion from i82559_start().
	(i82559_configure): More generic flags passed in and handled.
	(i82559_set_multicast): Routine to pass a set of addresses into
	the device; it works out the bitmap et al for us!
	(i82559_ioctl): Handle new case arms ETH_DRV_SET_MC_LIST and
	ETH_DRV_SET_MC_ALL if defined.  Reset and restart the device,
	diddling its status between.
	
2002-02-19  Mark Salter  <[e-mail]>

	* src/if_i82559.c (pci_init_find_82559s): Don't use IRQ support if
	CYGPKG_IO_ETH_DRIVERS_STAND_ALONE.

2002-01-28  Mark Salter  <[e-mail]>

	* src/if_i82559.c (pci_init_find_82559s): Removed CYG_ASSERT which
	assumed system PCI window exists.

2002-01-10  Jonathan Larmour  <[e-mail]>

	* src/if_i82559.c (find_82559s_match_func): Add some more compatible PCI
	ids.

2002-01-03  Mark Salter  <[e-mail]>

	* src/if_i82559.c (pci_init_find_82559s): Remove check that
	assumes driver owns entire PCI window.

	* cdl/intel_i82559_eth_drivers.cdl: Add
	CYGNUM_DEVS_ETH_INTEL_I82559_MAX_TX_DESCRIPTORS and
	CYGNUM_DEVS_ETH_INTEL_I82559_MAX_RX_DESCRIPTORS.

2001-11-19  Hugo Tyson  <[e-mail]>
2001-11-19  Anssi Pulkkinen <[e-mail]>

	* src/if_i82559.c (TxMachine): The test to see if there is a new
	tx to start must also say "go" if the tx queue is full.  Normally,
	if the txqueue is full, a tx will be occurring at all times - so
	one will complete soon, so the tx queue full flag will soon be
	cleared, and this condition will recover.  I suspect a subtle race
	which effectively means a new tx is queued and fills the queue in
	between two tx's, so no TxDone() follows, causes the hang which we
	get without this extra test under high load.  [CASE 106686]

2001-09-20  Gary Thomas  <[e-mail]>

	* src/if_i82559.c: Add declaration of 'i82559_stop' to quiet warning.

2001-09-13  Hugo Tyson  <[e-mail]>

	* cdl/intel_i82559_eth_drivers.cdl: 
	* src/if_i82559.c: 
	Remove the tickle thread from this driver.  The functionality is
	now in the network timeout functionality which handles the
	"deliver" calls.  So it's only called if the net is quiet.

2001-09-12  Hugo Tyson  <[e-mail]>

	* src/if_i82559.c (TxDone): Defensive programming; zero the Tx key
	in the global data before making the callback.  This change
	appeared useful in our forebear, the EBSA-specific driver.

2001-08-29  Hugo Tyson  <[e-mail]>

	* cdl/intel_i82559_eth_drivers.cdl: New CDL options called 
	CYGPKG_DEVS_ETH_INTEL_I82559_TICKLE_THREAD, .._PRIORITY and
	.._DELAY to control the thread.  By default this is all disabled
	because it is not needed for a real network application which
	itself proactively tries to use the net - it's only needed when
	the app is totally passive, so the driver otherwise does not get
	CPU time at all.
	
	* src/if_i82559.c (starti82559ticklethread): New code to start a
	thread (!) to catch lost interrupts and thus restart jammed
	hardware.  It's only possible if KERNEL and NET - RedBoot is
	unaffected, and it's not needed for RedBoot 'cos RedBoot polls.
	(TxMachine): An assert about the CU idle status sometimes fired;
	made the code more conservative wrt timing here.
	(comments): more documentation on the possible external configury
	by .inl file, described the two TIMEOUT macros that can be set,
	and indeed must be to use the tickling thread meaningfully.

2001-08-22  Gary Thomas  <[e-mail]>

	* src/if_i82559.c: 
	printf() is no longer a part of RedBoot.  Thus all programs
	must use diag_printf() and related functions instead.

2001-06-22  Jesper Skov  <[e-mail]>

	* src/if_i82559.c (i82559_init): Handle EEPROMS without CRC.

2001-05-16  Jonathan Larmour  <[e-mail]>

	* src/if_i82559.c (udelay): Use virtual vector rather than assuming
	hal_delay_us() exists.

2001-04-09  Hugo Tyson  <[e-mail]>

	* src/if_i82559.c (pciwindow_mem_alloc): If asked, by
	CYGHWR_DEVS_ETH_INTEL_I82559_PCIMEM_DISCONTIGUOUS, check for
	breaks in the physical address of PCI window memory, and discard
	any block with a break in it, then redo the alloc.  This is for
	targets where smaller SDRAM chips are fitted than the address
	width, so they end up with gaps in (physical) memory.

2001-04-09  Hugo Tyson  <[e-mail]>

	* src/if_i82559.c (i82559_init): Support hardware with only one
	serial EEPROM attached to one device, but multiple devices.  This
	generally involves moving some decisions for eg. programming
	EEPROM or not, optionally from compiletime to runtime.  Improved
	the diagnostics a bit also, and fixed some bugs.
	(eth_set_mac_address): now takes an arg to say whether we want to
	program the EEPROM, or are just setting up the chip.
	(i82559_ioctl): This is the only place we set that arg to write
	the EEPROM, startup never tries to write EEPROM from any external
	source such as configured ESA or RedBoot params.

2001-04-09  Hugo Tyson  <[e-mail]>

	* src/if_i82559.c (i82559_start): Warnings fixed; unused vars
	following previous change.

2001-04-09  Hugo Tyson  <[e-mail]>

	* src/if_i82559.c (i82559_start): Do not wait for command
	completion after the initial DUMPSTATS operation.  The code to
	wait was bogus; it was looking at the structure in the wrong way.
	In any case, there is no need to wait, the wait_for_cmd_done() in
	any following activity is good enough.

2001-03-28  Hugo Tyson  <[e-mail]>

	* src/if_i82559.c (update_statistics): Fix warning - only declare
	op_i82559 if we use it - from previous change.

2001-03-23  Hugo Tyson  <[e-mail]>

	* src/if_i82559.c (i82559_can_send, i82559_send): Count up and
	down entries into these routines, and always say "no" to
	can_send() if we are recursing, ie. within_send != 0.  This is a
	good idea because it limits stack usage.  The recursion happens
	because of the very neccessary polling of rx state within the tx
	routines - which can deliver packets up into the stack, which can
	trigger a response tx and so on...  Also fixed a harmless bug
	where, after polling, the p_i82559 current device state pointer
	pointer to the wrong device, if we polled both devs regardless; it
	only caused problems with the within_send recursion count.

	* include/i82559_info.h (I82559): Add within_send field.

2001-03-15  Hugo Tyson  <[e-mail]>

	* src/if_i82559.c (i82559_init): Support external macro
	CYGHWR_DEVS_ETH_INTEL_I82559_GET_ESA(...), which if defined, is
	used to acquire a MAC address, on the assumption that there is no
	EEPROM present.  CYGHWR_DEVS_ETH_INTEL_I82559_HAS_NO_EEPROM
	confirms that there is no EEPROM attached to the 82559(s) so we
	can omit all that code.  It's in the negative sense because it's
	the unusual case.

2001-03-13  Hugo Tyson  <[e-mail]>

	* src/if_i82559.c (CYGHWR_DEVS_ETH_INTEL_I82559_DEMUX_ALL): Fully
	enable "DEMUX_ALL" option; all devices are scanned for
	activity/wedging/interrupts at every opportunity.  Multiple
	devices on the same interrupts are now supported.

2001-03-12  Hugo Tyson  <[e-mail]>

	* src/if_i82559.c (i82559_configure): New routine combines actions
	of eth_set_promiscuous_mode() and eth_set_config() both of which
	disappear.  Thus the device is always configured correctly from
	the start.  [from a patch from Martin Buck  <[e-mail]>