Subversion Repositories openrisc

2002-10-06  Andrew Lunn  

        * include/devs_eth_ebsa285.inl: EBSA has the ethernet devices

        multiplexed onto one interrupt. This got losted when generalising

        the driver.

2002-05-14  Jesper Skov  

        * include/devs_eth_ebsa285.inl: Fix warnings.

2001-11-19  Hugo Tyson  

2001-11-19  Anssi Pulkkinen 

        * src/if_ebsa285.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]

2001-08-31  Hugo Tyson  

2001-08-31  Sanjay Bisen 

        * src/if_ebsa285.c (TxDone): Check from prev change is moved to

        the logical driver.  Also, we zero the key in global state before

        the callback - this should make no difference, but again it's good

        practice, and ASCOM's tests indicate a benefit for CASE 106059.

2001-08-22  Gary Thomas  

        * src/if_ebsa285.c:

        printf() is no longer a part of RedBoot.  Thus all programs

        must use diag_printf() and related functions instead.

2001-06-08  Gary Thomas  

        * src/if_ebsa285.c (TxDone): Defensive programming: only call

        tx_done with new/unique key values.

2001-03-12  Hugo Tyson  

        * src/if_ebsa285.c: Make device config more consistent with that

        used in other 82559 platforms, and fix a typo.

2001-02-20  Hugo Tyson  

2001-02-15  Martin Buck  

        * src/if_ebsa285.c: Configure i82559 not to drop oversized frames.

        This is required for VLAN support (because the maximum frame size

        including the VLAN tag is 1518 instead of 1514) and shouldn't

        hurt otherwise. To do this, we now always configure the chip when

        i82559_start is called, instead of relying on the default

        configuration if promiscuous mode is off. Another advantage is that

        we're now able to turn promiscuous mode off again after it was

        enabled once.

2001-01-02  Hugo Tyson  

        * src/if_ebsa285.c (find_82559s_match_func): New func to match

        complex PCI device requirements.

        (pci_init_find_82559s): Use it, and the new PCI API to correctly

        match a disparate bunch of devices on the one board.

2000-12-21  Hugo Tyson  

        * src/if_ebsa285.c (pci_init_find_82559s): Also check for the PCI

        device whose name is ... 0x1209; "82559ER".  Thanks to Martin Buck

         for the ID for that.

2000-11-19  Gary Thomas  

        * src/if_ebsa285.c (pci_init_find_82559s): Intel has at least

        two devices equivalent to the 82559.  Support both (0x1229, 0x01030).

2000-10-05  Hugo Tyson  

        * src/if_ebsa285.c: Deal with device interrupts in a nested

        fashion - disable/restore is the semantics now, rather than

        unconditionally unmasking.  Also go directly to the 21285 PIC's

        interrupt control registers to gain atomicity for these.  Poll for

        ready received packets when acknowledging an interrupt in the

        tranmitting world; a race here could lose an Rx interrupt.  Which

        doesn't matter on a busy system, but in quieter times...  there

        will always be such a race because of the vague way the i82559's

        status bits reflect how it's yanking the interrupt line; you have

        to poll until the interrupt is gone before returning else spurious

        interrupt failures occur.  The issue is to close the window as

        tightly as possible, which this change achieves at a minor cost in

        performance - because of the time spent polling when not required.

2000-09-11  Hugo Tyson  

        * src/if_ebsa285.c (i82559_poll): Only diddle the interface we

        were asked to.  This is more correct in terms of the intent of the

        API, though it shouldn't really be necessary.

2000-09-06  Hugo Tyson  

        * src/if_ebsa285.c (pci_init_find_82559s): Add asserts and an

        unconditional check that the PCI window as configured matches the

        address and size of the pci_window region from the MLT.  This is

        here because at present the MLT and CT cannot intercommunicate

        enough.  The separation of the PCI window is needed because

        otherwise the malloc heap will eat all memory.  [This is related

        to CR 902624-CR, "MLT needs to be configuration aware"]

2000-09-01  Hugo Tyson  

        * OVERVIEW: This is part of the change to the network stack to

        greatly reduce latencies both of (other) DSRs and of thread

        scheduling.  All the work that the network stack *and* individual

        ether drivers used to do in DSRs (including alarm callbacks and

        data copies to/from the device memory) is moved into a "fast

        network thread" instead.  It calls a device's "deliver" function

        to do the work that was previously in the DSR.  This is a separate

        thread so that it can be set higher priority than application

        threads in order to minimize packet loss (depending on the

        driver), if required (the application threads presumed to be

        higher priority in turn than the network thread).  A crucial

        consequence of this is that we are no longer locking against DSRs,

        so a plain mutex can be used rather than the global scheduler

        lock, thus simplifying all the splfoo/splx() style functions.

        * src/if_ebsa285.c: Minor: fix the big assert in i82559_send()

        which suffered a race condition when called from the fast thread

        rather than from a DSR.  Major: Add a "deliver" entry to the

        interface record for the "fast thread" implementation of the

        network internal comms system.  Provide a pass-up DSR to the

        logical ether driver's DSR and appropriate delivery routine(s).

        i82559_poll() now calls i82559_deliver() rather than the DSR.  Add

        valid data for mux'd DSR to pass on up.

2000-09-01  Hugo Tyson  

        * tests/test_net_realtime.h: Tighten up the latency requirements

        by a factor of 5; it all seems happy, so committed.

2000-08-25  Hugo Tyson  

        * src/if_ebsa285.c (i82559_ioctl): A little further diddling; have

        a bool to say whether the dot3 info is filled in.

2000-08-24  Hugo Tyson  

        * cdl/ebsa285_eth_drivers.cdl: Do not export a symbol for the

        device info file (include/ebsa285_info.h) since nobody needs

        (portably) to include it now.

        * src/if_ebsa285.c (i82559_ioctl): Handle new ioctl calls

        ETH_DRV_GET_IF_STATS_UD and ETH_DRV_GET_IF_STATS to get loads of

        statistical information out.  _UD means update.  The nonUD one can

        be used instead mostly, if we find the performance hit too large.

        This should allow SNMP (a) to not explode, (b) to get useful info

        out of other device implementations than this one.

        * include/ebsa285_info.h: Remove all the macro cruft for feature

        detecting of lots of individual statistics; we now just have a

        catch-all struct that SNMP uses, defined in the common ether

        driver environment.

2000-08-15  Hugo Tyson  

        * src/if_ebsa285.c (PacketRxReady): Put back the check for very

        small packets into the driver; the layer above checks for that

        (defensive programming) but only *after* asserting that the size

        is large enough, to help detect that scenario from other drivers.

        I believe we only have struct ether_header available if CYGPKG_NET

        but I could be wrong.

        [CASE 104353]

2000-08-08  Hugo Tyson  

        * src/if_ebsa285.c (eth_set_promiscuous_mode):

        - ccs->config_bytes[18]=0x70;

        + ccs->config_bytes[18]=0x72; // Keep the Padding Enable bit

        ...otherwise short frame sends don't work in promisc mode.

        [CASE 104289]

2000-08-07  Gary Thomas  

        * src/if_ebsa285.c (pciwindow_mem_alloc): Take out very noisy debug.

2000-08-03  Gary Thomas  

        * src/if_ebsa285.c: Changes for stand-alone mode.

        * cdl/ebsa285_eth_drivers.cdl: Ethernet driver package hierarchy changed.

        Add option to control number of interfaces.

2000-07-28  Hugo Tyson  

        * src/if_ebsa285.c (PacketRxReady): Do not attempt to forward

        short packets; eth_drv.c assumes there is at least a header there.

        (i82559_recv): Also be more careful and ASSERTive about -ve buffer

        sizes; be more defensive about sglists.  [CASE 104206]

2000-07-26  Gary Thomas  

        * src/if_ebsa285.c: Update for new eth_drv interfaces.

2000-07-18  Hugo Tyson  

        * src/if_ebsa285.c (i82559_poll): Fill in the flesh of this, it

        just calls ISR and DSR repeatedly.

        (i82559_start): Look in the device record for promiscuous mode

        flag; it should be passed though the common layer, but it's not

        [yet] - this change from Andrew Lunn/ASCOM.  Also a fix and delay

        to the promisc mode code per se.

2000-07-17  Hugo Tyson  

        * src/if_ebsa285.c (i82559_poll): New function, just to fill in

        the interface record; not used.

2000-06-27  Hugo Tyson  

        * cdl/ebsa285_eth_drivers.cdl: Add sesquipedalian option

        CYGDBG_DEVS_ETH_ARM_EBSA285_KEEP_82559_STATISTICS in (now)

        component CYGDBG_DEVS_ETH_ARM_EBSA285_KEEP_STATISTICS to control

        keeping (well, harvesting really) the i82559's internal stats.

        Reputedly, it doesn't service the net whilst this is happening, so

        it could be viewed a bad thing.  Hence the option.

        * include/ebsa285_info.h: Only describe the I82559_COUNTERS

        i82559_counters[2]; structs if full stats are to be kept.

        * src/if_ebsa285.c (update_statistics): Only include this if full

        stats are to be kept.

2000-06-27  Hugo Tyson  

        * src/if_ebsa285.c (ResetRxRing): Re-do the management of the

        RxRing; have an end-of-list flag (EL) in the last entry, and as

        you unload filled slots, drag it round after you.

2000-06-14  Hugo Tyson  

        * cdl/ebsa285_eth_drivers.cdl: Add option to control statistics

        keeping.

        * include/ebsa285_info.h: Act on it.

2000-06-13  Hugo Tyson  

        * cdl/ebsa285_eth_drivers.cdl: Properly include the new header

        file and define symbols to allow client code to get at it without

        knowing the particular device driver name.

        * include/ebsa285_info.h: New file: export various statistics

        information about the driver for use by monitoring and

        network-management systems.  This requires exposing the

        (otherwise) internal structures of the driver.

        * src/if_ebsa285.c: remove a lot of structure definitions &c that

        are now necessarily in the new header; add a couple of new

        routines which provide status and update statistics from the

        device into memory; tidy up control of whether stats-keeping is

        enabled.

2000-06-06  Hugo Tyson  

        * cdl/ebsa285_eth_drivers.cdl (define_proc): Add #define of

        CYGDAT_DEVS_ETH_DESCRIPTION in the config file for information.

2000-05-12  Hugo Tyson  

        * tests/test_net_realtime.h (tnr_print_activity): New routine to

        check the system is working, tidied up the API.  It works!

2000-05-11  Hugo Tyson  

        * cdl/ebsa285_eth_drivers.cdl: Added export of the test header

        below, and config opts for controlling EEPROM writing and all the

        status chatter as the device starts up.

        * src/if_ebsa285.c: Reworked the code for reading and setting the

        EEPROM that holds the MAC address.  This is very ugly, but now

        more reliable.  Also tidied up printing cruft with neater

        configury, and made it an option (for safety) whether it's

        possible to write the EEPROM at all.

        * tests/test_net_realtime.h: New file - it is intended to be used

        by networking tests to verify that latency is not compromised by

        the stack and driver.  It's very platform specific, hence the

        location in here.  This is a preliminary version only.

2000-04-27  Hugo Tyson  

        * src/if_ebsa285.c: A serious re-write.  This cuts out a lot of

        code from the old version and improves the performance greatly.

        The cruft was mainly doing lots of explicit event communication

        between the ISR and DSR, when in fact all the state needed is

        present in the tx/rx rings.  So both ISRs and DSRs regard their

        call as an opportunity to progress everything they can, rather

        than only dealing with one interrupt cause at a time; the

        connection between them is now rather looser.

        Interrups can now be re-enabled after the ISR (in other words they

        are not masked in the ISR), no need to wait for the DSR, but in

        consequence some DSR code must mask/unmask intrs as it works.

        The 82559 appears to be a little slow in reacting to commands and

        state changes, so some interrupts were being lost - or persisting

        beyond their desired life - so there's some kinda polling code to

        deal with that also.  We also rely on the foreground to kind of

        poll in the same way, in the send/can_send calls - we know the

        stack will re-try if necessary, though this is rare.

        The driver now works (albeit at much reduced performance) with as

        few as 6 rx and tx buffers - in other words the "queue full/out of

        rx buffers" states have been tested and all is well.  It works

        generally fine with 8 buffers of each kind.

        The mux ISR and DSR are now rather more polled than the old

        versions; we just try to do things with both devices (if active)

        by simply calling each unitary ISR/DSR respectively.

        I also re-ordered some of the code, moving utilities to the end of

        the file and grouping together Tx and Rx machines a bit better.

2000-04-13  Hugo Tyson  

        * src/if_ebsa285.c: Attribution to Ron Spence, Pacific Softworks

        added as a contributor.

2000-04-07  Hugo Tyson  

        * ecos.db: Re-organize device packages.  This is a massive change

        involving deleting all the sources for serial and ethernet drivers

        from where they used to live in

            packages/io/serial/current/src/ARCH/PLATFORM.[ch]

            packages/net/drivers/eth/PLATFORM/current/src/...

        and reinstating them in

            packages/devs/serial/ARCH/PLATFORM/current/src/...

            packages/devs/eth/ARCH/PLATFORM/current/src/...

        All these new packages are properly defined in ecos.db, and are

        all of type "hardware" so that a "target" can grab them.

        This directory layout is descriptive of the devices we have right

        now, arch and platform are separate levels just to make it easier

        to navigate in the filesystem and similar to the HAL structure in

        the filesystem.

        It is *not* prescriptive of future work; for example, the mythical

        common highly-portable 16550 serial driver which works on many

        targets would be called "devs/serial/s16550/current", or a serial

        device for a particular board (cogent springs to mind) that can

        work with different CPUs fitted is "devs/serial/cogent/current".

        Changelogs have been preserved and replicated over all the new

        packages, so that no history is lost.

        The contents of individual source files are unchanged; they build

        in just the same emvironment except for a very few cases where the

        config file name changed in this movement.

        Targets in ecos.db have been redefined to bring in all relevant

        hardware packages including net and serial drivers (but the newly

        included packages are only active if their desired parent is

        available.)

        The names of CDL options (and their #defines of course) stay the

        same for the serial drivers, for backward compatibility.

        * templates/*/current.ect: these have had CYGPKG_IO_SERIAL added

        rather than it being in (almost) all target definitions.

2000-03-29  Hugo Tyson  

        * src/if_ebsa285.c (i82559_recv): Be happy with NULLs in the

        SGlist; it means the caller is out of memory so drop the packet on

        the floor.  Also remove a completely redundant test.

2000-03-06  Hugo Tyson  

        * src/if_ebsa285.c (i82559_can_send): Update net driver to new

        interface style.  This is incomplete wrt promiscuous mode, but

        that's probably about all.

2000-02-14  Gary Thomas  

        * cdl/ebsa285_eth_drivers.cdl: Update CDL to indicate multiple

        interface support.

2000-02-14  Hugo Tyson  

        * src/if_ebsa285.c (pci_init_find_82559s): Tidy comments somewhat

        and set debug and stats collecting defines to most friendly

        settings.

2000-02-10  Hugo Tyson  

        * src/if_ebsa285.c (PacketRxReady): Fix bug; current descriptor

        was not being write back for the callback to use.  Hence asserts

        on state of rfd were firing in busy times - that leading rfd had

        already been drained.

        Also rationalized meaning of DEBUG printy symbols a bit - it's now

        chatty during startup/foreground manipulations but not in any

        performance related activities ie. rx or tx.

2000-02-09  John Dallaway  

        * cdl/ebsa285_eth_drivers.cdl:

        Reparent under CYGPKG_NET_ETH_DRIVERS and tidy display strings.

2000-02-08  Hugo Tyson  

        * src/if_ebsa285.c: New File.

        * cdl/ebsa285_eth_drivers.cdl: New File.

        Initial Checkin of EBSA285 Ethernet driver.

        It's one monolithic file at present, and should be split up into a

        more generic Intel 82559 driver plus platform-specific parts (PCI

        et al) plus eCos/Red-Hat-BSD-stack parts.

//===========================================================================

//####ECOSGPLCOPYRIGHTBEGIN####

// -------------------------------------------

// This file is part of eCos, the Embedded Configurable Operating System.

// Copyright (C) 1998, 1999, 2000, 2001, 2002 Red Hat, Inc.

//

// eCos is free software; you can redistribute it and/or modify it under

// the terms of the GNU General Public License as published by the Free

// Software Foundation; either version 2 or (at your option) any later version.

//

// eCos is distributed in the hope that it will be useful, but WITHOUT ANY

// WARRANTY; without even the implied warranty of MERCHANTABILITY or

// FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License

// for more details.

//

// You should have received a copy of the GNU General Public License along

// with eCos; if not, write to the Free Software Foundation, Inc.,

// 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA.

//

// As a special exception, if other files instantiate templates or use macros

// or inline functions from this file, or you compile this file and link it

// with other works to produce a work based on this file, this file does not

// by itself cause the resulting work to be covered by the GNU General Public

// License. However the source code for this file must still be made available

// in accordance with section (3) of the GNU General Public License.

//

// This exception does not invalidate any other reasons why a work based on

// this file might be covered by the GNU General Public License.

//

// Alternative licenses for eCos may be arranged by contacting Red Hat, Inc.

// at http://sources.redhat.com/ecos/ecos-license/

// -------------------------------------------

//####ECOSGPLCOPYRIGHTEND####

//===========================================================================