people/indolent/gpxe.git/.git
13 years agomake md5_steps static
Holger Lubitz [Fri, 27 Jul 2007 19:36:16 +0000 (21:36 +0200)]
make md5_steps static

13 years agomake imgfill_cmdline static
Holger Lubitz [Fri, 27 Jul 2007 19:35:17 +0000 (21:35 +0200)]
make imgfill_cmdline static

13 years agodisable AES_convert_key by #if 0'ing it out (suggested by mcb30)
Holger Lubitz [Fri, 27 Jul 2007 19:33:53 +0000 (21:33 +0200)]
disable AES_convert_key by #if 0'ing it out (suggested by mcb30)

13 years agomake netboot() static
Holger Lubitz [Fri, 27 Jul 2007 19:31:45 +0000 (21:31 +0200)]
make netboot() static

13 years agodisable AES_convert_key by #if 0'ing it out (suggested by mcb30)
Holger Lubitz [Fri, 27 Jul 2007 19:31:11 +0000 (21:31 +0200)]
disable AES_convert_key by #if 0'ing it out (suggested by mcb30)

13 years agoAdded two lines of code that were completely missing! origin/origin
Michael Brown [Wed, 25 Jul 2007 12:32:02 +0000 (13:32 +0100)]
Added two lines of code that were completely missing!

13 years agoDon't overwrite %dl with a (potentially) modified drive number if we
Michael Brown [Wed, 25 Jul 2007 01:03:29 +0000 (02:03 +0100)]
Don't overwrite %dl with a (potentially) modified drive number if we
already handled the INT13 call.

13 years agoWe probably shouldn't leave interrupts disabled when starting an OS.
Michael Brown [Wed, 25 Jul 2007 01:02:24 +0000 (02:02 +0100)]
We probably shouldn't leave interrupts disabled when starting an OS.

13 years agoAdd per-file error identifiers
Michael Brown [Tue, 24 Jul 2007 16:11:31 +0000 (17:11 +0100)]
Add per-file error identifiers

13 years agoDefine -DOBJECT when generating Makefile rules, for consistency.
Michael Brown [Tue, 24 Jul 2007 14:19:09 +0000 (15:19 +0100)]
Define -DOBJECT when generating Makefile rules, for consistency.

Inhibit -Werror when generating Makefile rules, to prevent warnings
from killing the dependency list.

13 years agoInclude errno.h in the few drivers which use it, rather than in nic.h
Michael Brown [Tue, 24 Jul 2007 13:05:32 +0000 (14:05 +0100)]
Include errno.h in the few drivers which use it, rather than in nic.h

13 years agoAllow -Werror to be temporarily disabled using "make NO_WERROR=1"
Michael Brown [Tue, 24 Jul 2007 13:01:14 +0000 (14:01 +0100)]
Allow -Werror to be temporarily disabled using "make NO_WERROR=1"

13 years agoRemove some obsolete sections of librm.h and libkir.h.
Michael Brown [Tue, 24 Jul 2007 11:46:07 +0000 (12:46 +0100)]
Remove some obsolete sections of librm.h and libkir.h.

Add some of the missing parts to libkir.h.

13 years agoWe can finally enable -Werror :)
Michael Brown [Tue, 24 Jul 2007 11:43:11 +0000 (12:43 +0100)]
We can finally enable -Werror :)

13 years agoMerge branch 'master' of rom.etherboot.org:/pub/scm/gpxe
Michael Brown [Tue, 24 Jul 2007 11:37:09 +0000 (12:37 +0100)]
Merge branch 'master' of rom.etherboot.org:/gpxe

13 years agoChange #warning to FIXME for ns8390.c
Marty Connor [Mon, 23 Jul 2007 01:00:55 +0000 (21:00 -0400)]
Change #warning to FIXME for ns8390.c

This driver really needs to be rewritten.
It tries to build both ISA and PCI images,
and makes life diffifult for the build system
and rom-o-matic.net.  The warning was just a reminder
that it needs to be cleaned up and re-factored
to split the PCI and ISA drivers.

13 years agoChange #warning to FIXME for depca
Marty Connor [Mon, 23 Jul 2007 00:56:26 +0000 (20:56 -0400)]
Change #warning to FIXME for depca

This ancient ISA driver should probably be removed.
It is hard to get a card to test it with, and there
are comments to the effect that it cannot work with
relocation.  I would be quite interested to get a
bug report by someone who actually has this card.

13 years agoChanged #warning to FIXME for nvo hack warning.
Marty Connor [Mon, 23 Jul 2007 00:52:32 +0000 (20:52 -0400)]
Changed #warning to FIXME for nvo hack warning.

13 years agoAdd NIC to targets that are always made.
Marty Connor [Mon, 23 Jul 2007 00:41:53 +0000 (20:41 -0400)]
Add NIC to targets that are always made.

Even though it is only currently used for rom-o-matic.net code,
it is good to make sure it gets built, and that the parsing code
is working.

13 years agoMinor debug improvement
Michael Brown [Thu, 19 Jul 2007 16:01:21 +0000 (17:01 +0100)]
Minor debug improvement

13 years agoDon't trash the %ecx value returned by relocate(). This was causing
Michael Brown [Thu, 19 Jul 2007 14:54:33 +0000 (15:54 +0100)]
Don't trash the %ecx value returned by relocate().  This was causing
us to round down the size for the relocation copy to the nearest 64kB
(+0x10 bytes); this just happened to work on most machines because the
last 64kB of the image is all-zeroes anyway (it's the .bss).

13 years agoRemove INT 13,4b handler; it represents an incomplete feature (CD-ROM
Michael Brown [Thu, 19 Jul 2007 10:16:05 +0000 (11:16 +0100)]
Remove INT 13,4b handler; it represents an incomplete feature (CD-ROM
emulation support) for an unknown purpose, and breaks grub.

13 years agoCompleted definition of struct int13_cdrom_specification, and moved to
Michael Brown [Thu, 19 Jul 2007 10:15:01 +0000 (11:15 +0100)]
Completed definition of struct int13_cdrom_specification, and moved to
int13.h.

13 years agoReal-mode code might set the direction flag, which would cause carnage.
Michael Brown [Tue, 17 Jul 2007 12:46:37 +0000 (13:46 +0100)]
Real-mode code might set the direction flag, which would cause carnage.

13 years agoMove .zinfo to libprefix.S; it doesn't belong with the decompression code.
Michael Brown [Tue, 17 Jul 2007 11:52:54 +0000 (12:52 +0100)]
Move .zinfo to libprefix.S; it doesn't belong with the decompression code.

13 years agoProduce no output unless -v is specified or an error occurs.
Michael Brown [Tue, 17 Jul 2007 00:59:48 +0000 (01:59 +0100)]
Produce no output unless -v is specified or an error occurs.

13 years agoRemoved dead code identified by symcheck
Michael Brown [Tue, 17 Jul 2007 00:43:19 +0000 (01:43 +0100)]
Removed dead code identified by symcheck

13 years agoAdded missing "static" declarations
Michael Brown [Tue, 17 Jul 2007 00:42:10 +0000 (01:42 +0100)]
Added missing "static" declarations

13 years agoRestore the "shared symbol found only in a single object should be
Michael Brown [Tue, 17 Jul 2007 00:39:06 +0000 (01:39 +0100)]
Restore the "shared symbol found only in a single object should be
treated as being provided by that object" code.

13 years agoAdded symbols that the DBG() macros will use.
Michael Brown [Tue, 17 Jul 2007 00:24:04 +0000 (01:24 +0100)]
Added symbols that the DBG() macros will use.

13 years agonameserver should be static
Michael Brown [Tue, 17 Jul 2007 00:21:20 +0000 (01:21 +0100)]
nameserver should be static

13 years agoObsolete code removal
Michael Brown [Tue, 17 Jul 2007 00:19:35 +0000 (01:19 +0100)]
Obsolete code removal

13 years agoAdded missing extern
Michael Brown [Tue, 17 Jul 2007 00:16:24 +0000 (01:16 +0100)]
Added missing extern

13 years agoUse objdump -t for symcheck; it allows us to take the symbol section
Michael Brown [Tue, 17 Jul 2007 00:15:39 +0000 (01:15 +0100)]
Use objdump -t for symcheck; it allows us to take the symbol section
into account (and so remove the false positives due to .tbl.* symbols).

13 years agoKill off redundant references to the .zprefixes.
Michael Brown [Mon, 16 Jul 2007 16:26:07 +0000 (17:26 +0100)]
Kill off redundant references to the .zprefixes.

13 years agoUpdate all prefixes currently referring to _load_size to include a
Michael Brown [Mon, 16 Jul 2007 16:17:26 +0000 (17:17 +0100)]
Update all prefixes currently referring to _load_size to include a
compressor fixup section.

13 years agoRename _rom_size to _load_size_sect, since hdprefix.S also uses it.
Michael Brown [Mon, 16 Jul 2007 16:03:19 +0000 (17:03 +0100)]
Rename _rom_size to _load_size_sect, since hdprefix.S also uses it.

13 years agoCompressed ROM images now work.
Michael Brown [Mon, 16 Jul 2007 15:58:38 +0000 (16:58 +0100)]
Compressed ROM images now work.

13 years agoSplit the objcopy stage out of final image preparation to a separate
Michael Brown [Mon, 16 Jul 2007 12:41:27 +0000 (13:41 +0100)]
Split the objcopy stage out of final image preparation to a separate
bin/%.bin target.

13 years agoMove uninitialised .data16 variables to .bss16; saves around 2000
Michael Brown [Mon, 16 Jul 2007 12:22:12 +0000 (13:22 +0100)]
Move uninitialised .data16 variables to .bss16; saves around 2000
bytes of useless zeroes in the final image.

13 years agoKill off some long-redundant CFLAGS
Michael Brown [Mon, 16 Jul 2007 12:15:12 +0000 (13:15 +0100)]
Kill off some long-redundant CFLAGS

13 years agoAdd __bss16() macro, and allow use of .bss16 section by removing
Michael Brown [Mon, 16 Jul 2007 12:10:49 +0000 (13:10 +0100)]
Add __bss16() macro, and allow use of .bss16 section by removing
link-time check for section overlaps.  (In order to avoid wasting
space in the executable image, .bss16 will overlap with the following
section, which is .text).

13 years agoRevert "no need for depending on build library ("blib")"
Michael Brown [Mon, 16 Jul 2007 12:06:58 +0000 (13:06 +0100)]
Revert "no need for depending on build library ("blib")"

This reverts commit 9dc2515d7ef16e1149c048493f0c7887dfd3d054.

I want "make" to actually build something.

13 years agoAdded nrv2b binary to ignore list
Michael Brown [Sun, 15 Jul 2007 12:31:31 +0000 (13:31 +0100)]
Added nrv2b binary to ignore list

13 years agoUse a single _payload_offset linker-defined variable to locate the
Michael Brown [Sun, 15 Jul 2007 01:52:02 +0000 (02:52 +0100)]
Use a single _payload_offset linker-defined variable to locate the
start of the non-prefix blocks in the loaded image, and rely on the
image ordering.  This should make introducing compression much easier.

13 years agoReverting unintended change caused by merge of indolent/natsemi branch.
Michael Brown [Sun, 15 Jul 2007 01:22:49 +0000 (02:22 +0100)]
Reverting unintended change caused by merge of indolent/natsemi branch.

13 years agoMerge commit 'indolent/natsemi'
Michael Brown [Sun, 15 Jul 2007 01:21:55 +0000 (02:21 +0100)]
Merge commit 'indolent/natsemi'

13 years agoremoved type casting from DBG statements. natsemi
Udayan Kumar [Sun, 15 Jul 2007 00:01:30 +0000 (20:01 -0400)]
removed type casting from DBG statements.

13 years agoreplaces printf with dbg
Udayan Kumar [Sat, 14 Jul 2007 19:36:15 +0000 (15:36 -0400)]
replaces printf with dbg

13 years agoadded natsemi.h
Udayan Kumar [Sat, 14 Jul 2007 04:44:56 +0000 (00:44 -0400)]
added natsemi.h

13 years agomdc's+nvs
Udayan Kumar [Sat, 14 Jul 2007 02:55:17 +0000 (22:55 -0400)]
mdc's+nvs

13 years agomore debugging
Udayan Kumar [Wed, 11 Jul 2007 16:30:44 +0000 (12:30 -0400)]
more debugging

13 years agomore debugging
Udayan Kumar [Tue, 10 Jul 2007 19:47:34 +0000 (15:47 -0400)]
more debugging

13 years agoadded silicon revision number
Udayan Kumar [Tue, 10 Jul 2007 19:29:55 +0000 (15:29 -0400)]
added silicon revision number

13 years agoinit_fix_up
Udayan Kumar [Tue, 10 Jul 2007 19:08:58 +0000 (15:08 -0400)]
init_fix_up

13 years agoadded cable magic for 100Mps in natsemi
Udayan Kumar [Tue, 10 Jul 2007 17:29:30 +0000 (13:29 -0400)]
added cable magic for 100Mps in natsemi

13 years agoduplex setting added to natsemi.c
Udayan Kumar [Tue, 10 Jul 2007 01:54:02 +0000 (21:54 -0400)]
duplex setting added to natsemi.c

13 years agomore indentation and styling done
Udayan Kumar [Sun, 8 Jul 2007 22:41:12 +0000 (18:41 -0400)]
more indentation and styling done

13 years ago:x
Udayan Kumar [Sun, 8 Jul 2007 19:12:38 +0000 (15:12 -0400)]
:x

13 years agono changes
Udayan Kumar [Sun, 8 Jul 2007 18:29:55 +0000 (14:29 -0400)]
no changes

13 years agoadded nat_irq to natsemi.c
Udayan Kumar [Sun, 8 Jul 2007 17:37:42 +0000 (13:37 -0400)]
added nat_irq to natsemi.c

13 years agostopped memory leak in natsemi::nat_open()
Udayan Kumar [Sun, 8 Jul 2007 15:03:30 +0000 (11:03 -0400)]
stopped memory leak in natsemi::nat_open()

13 years agoNatsemi commented and almost done
Udayan Kumar [Sun, 8 Jul 2007 00:42:53 +0000 (20:42 -0400)]
Natsemi commented and almost done

13 years agonatsemi.c is workin
Udayan Kumar [Sat, 7 Jul 2007 20:07:30 +0000 (16:07 -0400)]
natsemi.c is workin

13 years agoirq still not working will have to remove it
Udayan Kumar [Thu, 5 Jul 2007 12:38:37 +0000 (08:38 -0400)]
irq still not working will have to remove it

13 years agonatsemi
Udayan Kumar [Wed, 4 Jul 2007 02:23:17 +0000 (22:23 -0400)]
natsemi

13 years agointerrupt in natsemi
Udayan Kumar [Wed, 4 Jul 2007 02:20:47 +0000 (22:20 -0400)]
interrupt in natsemi

13 years agoenabled interrupt in natsemi.c
Udayan Kumar [Tue, 3 Jul 2007 23:29:31 +0000 (19:29 -0400)]
enabled interrupt in natsemi.c

13 years agoadded change log to natsemi.c
Udayan Kumar [Tue, 3 Jul 2007 02:45:50 +0000 (22:45 -0400)]
added change log to natsemi.c

13 years agoadded netdev_tx_complete to natsemi.c
Udayan Kumar [Sun, 1 Jul 2007 22:05:58 +0000 (18:05 -0400)]
added netdev_tx_complete to natsemi.c

13 years agoadded endianness to natsemi.
Udayan Kumar [Sun, 1 Jul 2007 21:11:22 +0000 (17:11 -0400)]
added endianness to natsemi.

13 years agofree_netdev -> netdev_put
Udayan Kumar [Sun, 1 Jul 2007 02:23:19 +0000 (22:23 -0400)]
free_netdev -> netdev_put

13 years agodebugging natsemi.c
Udayan Kumar [Sat, 30 Jun 2007 23:30:41 +0000 (19:30 -0400)]
debugging natsemi.c

13 years agoadded eeprom from rtl8139 but not working
Udayan Kumar [Tue, 26 Jun 2007 21:20:34 +0000 (17:20 -0400)]
added eeprom from rtl8139 but not working

13 years agonatsemi now needs eeprom access
Udayan Kumar [Tue, 26 Jun 2007 01:24:34 +0000 (21:24 -0400)]
natsemi now needs eeprom access

13 years agosame as before, but now compiling natsemi
Udayan Kumar [Wed, 20 Jun 2007 23:46:22 +0000 (19:46 -0400)]
same as before, but now compiling natsemi

13 years agoadded polling and transmit. eeprom access still remaining
Udayan Kumar [Wed, 20 Jun 2007 23:07:04 +0000 (19:07 -0400)]
added polling and transmit. eeprom access still remaining

13 years agoskel of new natsemi driver (still in developments)
Udayan Kumar [Mon, 18 Jun 2007 22:30:33 +0000 (18:30 -0400)]
skel of new natsemi driver (still in developments)

13 years agotesting if it works
Udayan Kumar [Sat, 16 Jun 2007 01:42:09 +0000 (21:42 -0400)]
testing if it works

13 years agoRevert mdc's warnings purge on natsemi.c, to allow for a clean rebase.
Michael Brown [Sat, 7 Jul 2007 21:33:45 +0000 (22:33 +0100)]
Revert mdc's warnings purge on natsemi.c, to allow for a clean rebase.

13 years agoSkip the length field prepended by util/nrv2b.c
Michael Brown [Sun, 15 Jul 2007 01:01:17 +0000 (02:01 +0100)]
Skip the length field prepended by util/nrv2b.c

13 years agoModified calling convention: we now update %esi and %edi just like a
Michael Brown [Sun, 15 Jul 2007 00:51:32 +0000 (01:51 +0100)]
Modified calling convention: we now update %esi and %edi just like a
"rep movsb".

13 years ago"not" doesn't affect flags, "xor $-1" does.
Michael Brown [Sun, 15 Jul 2007 00:25:13 +0000 (01:25 +0100)]
"not" doesn't affect flags, "xor $-1" does.

13 years agoRemove more obsolete files
Michael Brown [Sat, 14 Jul 2007 22:16:01 +0000 (23:16 +0100)]
Remove more obsolete files

13 years agoOn advice of hpa: be more patient with the KBC and SCPA methods; retry
Michael Brown [Sat, 14 Jul 2007 19:56:12 +0000 (20:56 +0100)]
On advice of hpa: be more patient with the KBC and SCPA methods; retry
in a long loop before giving up on them.  Record method which worked
and default to that method on next attempt.

13 years agoShrink cpu.c and render it useful for our purposes.
Michael Brown [Sat, 14 Jul 2007 18:12:13 +0000 (19:12 +0100)]
Shrink cpu.c and render it useful for our purposes.

13 years agoReduced etherboot.h dependencies to avoid unnecessary rebuilds.
Michael Brown [Sat, 14 Jul 2007 14:52:23 +0000 (15:52 +0100)]
Reduced etherboot.h dependencies to avoid unnecessary rebuilds.

etherboot.h is only now used by drivers that have not been updated to
the new API.

13 years agoDead code removal.
Michael Brown [Sat, 14 Jul 2007 14:42:26 +0000 (15:42 +0100)]
Dead code removal.

Kill off use of etherboot.h outside drivers/net.

13 years agoBe more aggressive in attempts to enable A20, now that we have the
Michael Brown [Sat, 14 Jul 2007 13:21:18 +0000 (14:21 +0100)]
Be more aggressive in attempts to enable A20, now that we have the
optimisation of only trying to do so when necessary.

13 years agoKill off the enforced RX quota; it only seems to hurt on real hardware.
Michael Brown [Fri, 13 Jul 2007 23:30:11 +0000 (00:30 +0100)]
Kill off the enforced RX quota; it only seems to hurt on real hardware.

13 years agoUse fast in-situ test for gate A20 being set, to cut down on the
Michael Brown [Fri, 13 Jul 2007 12:32:49 +0000 (13:32 +0100)]
Use fast in-situ test for gate A20 being set, to cut down on the
number of (potentially very slow) gateA20_set operations.

Die with a fatal error if we are unable to set gate A20; if this fails
then we are bound to experience memory corruption at a later stage,
and I'd prefer to pick it up early.

13 years agoAdd support for TCP timestamps
Michael Brown [Fri, 13 Jul 2007 10:32:53 +0000 (11:32 +0100)]
Add support for TCP timestamps

13 years agoAdjust received length to take into account any already-received data
Michael Brown [Fri, 13 Jul 2007 10:31:58 +0000 (11:31 +0100)]
Adjust received length to take into account any already-received data
in tcp_rx_data().

Clarify comments on discarding duplicate or out-of-order data.

13 years agoAvoid reusing auto-allocated ports after connection close.
Michael Brown [Fri, 13 Jul 2007 10:25:00 +0000 (11:25 +0100)]
Avoid reusing auto-allocated ports after connection close.

13 years agoImprove error reporting for strange length combinations reported by
Michael Brown [Tue, 10 Jul 2007 19:59:21 +0000 (20:59 +0100)]
Improve error reporting for strange length combinations reported by
the UNDI stack.

Ignore obviously invalid length combinations (as returned by
e.g. VMWare's PXE stack).

Limit to one packet per poll to avoid memory exhaustion.

13 years agoSet up %ds *before* testing a value in our data segment (d'oh!).
Michael Brown [Tue, 10 Jul 2007 16:08:32 +0000 (17:08 +0100)]
Set up %ds *before* testing a value in our data segment (d'oh!).

Always send EOI; do not chain to BIOS's default interrupt handler.
They are just too unpredictable; at least VMware's seems to kill the
machine if you go anywhere near it.

Disable interrupts after return from PXENV_UNDI_ISR, just in case some
dumb PXE stack enables them.

13 years agoReport RX errors via netdev_rx_err()
Michael Brown [Tue, 10 Jul 2007 16:01:18 +0000 (17:01 +0100)]
Report RX errors via netdev_rx_err()

13 years agoDump first 64 bytes of chained ISR when debugging is enabled.
Michael Brown [Tue, 10 Jul 2007 15:57:55 +0000 (16:57 +0100)]
Dump first 64 bytes of chained ISR when debugging is enabled.

13 years agoMay be required in some disconnect scenarios
Michael Brown [Tue, 10 Jul 2007 05:31:11 +0000 (06:31 +0100)]
May be required in some disconnect scenarios

13 years agoUnplug before sending close() message, to avoid screwing up interfaces
Michael Brown [Tue, 10 Jul 2007 05:25:30 +0000 (06:25 +0100)]
Unplug before sending close() message, to avoid screwing up interfaces
which respond to close with a reopen() (e.g. iSCSI).