Subversion Repositories openrisc_me

2002-10-06  Andrew Lunn  <andrew.lunn@ascom.ch>

        * 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  <jskov@redhat.com>

        * include/devs_eth_ebsa285.inl: Fix warnings.

2001-11-19  Hugo Tyson  <hmt@redhat.com>
2001-11-19  Anssi Pulkkinen <Anssi.Pulkkinen@ascom.ch>

        * 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  <hmt@redhat.com>
2001-08-31  Sanjay Bisen <sanjay.bisen@ascom.ch>

        * 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  <gthomas@redhat.com>

        * 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  <gthomas@redhat.com>

        * src/if_ebsa285.c (TxDone): Defensive programming: only call
        tx_done with new/unique key values.

2001-03-12  Hugo Tyson  <hmt@redhat.com>

        * 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  <hmt@redhat.com>
2001-02-15  Martin Buck  <martin.buck@ascom.ch>

        * 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  <hmt@redhat.com>

        * 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  <hmt@redhat.com>

        * src/if_ebsa285.c (pci_init_find_82559s): Also check for the PCI
        device whose name is ... 0x1209; "82559ER".  Thanks to Martin Buck
        <martin.buck@ascom.ch> for the ID for that.

2000-11-19  Gary Thomas  <gthomas@redhat.com>

        * 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  <hmt@redhat.com>

        * 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  <hmt@cygnus.co.uk>

        * 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  <hmt@cygnus.co.uk>

        * 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  <hmt@cygnus.co.uk>

        * 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  <hmt@cygnus.co.uk>

        * 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  <hmt@cygnus.co.uk>

        * 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  <hmt@cygnus.co.uk>

        * 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  <hmt@cygnus.co.uk>

        * 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  <hmt@cygnus.co.uk>

        * 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  <gthomas@redhat.com>

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

2000-08-03  Gary Thomas  <gthomas@redhat.com>

        * 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  <hmt@cygnus.co.uk>

        * 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  <gthomas@redhat.com>

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

2000-07-18  Hugo Tyson  <hmt@cygnus.co.uk>

        * 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  <hmt@cygnus.co.uk>

        * src/if_ebsa285.c (i82559_poll): New function, just to fill in
        the interface record; not used.

2000-06-27  Hugo Tyson  <hmt@cygnus.co.uk>

        * 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  <hmt@cygnus.co.uk>

        * 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  <hmt@cygnus.co.uk>

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

        * include/ebsa285_info.h: Act on it.

2000-06-13  Hugo Tyson  <hmt@cygnus.co.uk>

        * 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  <hmt@cygnus.co.uk>

        * 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  <hmt@cygnus.co.uk>

        * 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  <hmt@cygnus.co.uk>

        * 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  <hmt@cygnus.co.uk>

        * 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  <hmt@cygnus.co.uk>

        * src/if_ebsa285.c: Attribution to Ron Spence, Pacific Softworks
        added as a contributor.

2000-04-07  Hugo Tyson  <hmt@cygnus.co.uk>

        * 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  <hmt@cygnus.co.uk>

        * 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  <hmt@cygnus.co.uk>

        * 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  <gthomas@cygnus.co.uk>

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

2000-02-14  Hugo Tyson  <hmt@cygnus.co.uk>

        * 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  <hmt@cygnus.co.uk>

        * 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  <jld@cygnus.co.uk>

        * cdl/ebsa285_eth_drivers.cdl:

        Reparent under CYGPKG_NET_ETH_DRIVERS and tidy display strings.

2000-02-08  Hugo Tyson  <hmt@cygnus.co.uk>

        * 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####
//===========================================================================