Shao Miller [Tue, 25 Jan 2011 15:44:13 +0000 (10:44 -0500)]
memdisk: Enhance disk-probe debugging output
In an effort to trouble-shoot a problem report on some Dell
models (including an Optiplex GX260), we add further debugging
output to try to find out at which point things go wrong. The
problem units are apparently probing as "drive present" for all
drives, which obviously isn't right.
Signed-off-by: Shao Miller <shao.miller@yrdsb.edu.on.ca>
Shao Miller [Tue, 25 Jan 2011 02:47:43 +0000 (21:47 -0500)]
memdisk: Make debug-mode a tad prettier
By removing the use of #ifdef-#endif blocks.
Signed-off-by: Shao Miller <shao.miller@yrdsb.edu.on.ca>
H. Peter Anvin [Mon, 17 Jan 2011 19:54:45 +0000 (11:54 -0800)]
diskio: Support EDD 4 16-byte device paths
EDD 4 has 16-byte device path information. Make the EDD buffer big
enough to capture that. Note that the location of the device path
checksum actually depends on the device path length information field!
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
H. Peter Anvin [Mon, 17 Jan 2011 19:48:55 +0000 (11:48 -0800)]
diskio: add reminder that we need clear in the future
Currently, disk_init() can only be called once, and edd_params get
statically initialized to zero. This will change once this function
can be called more than once, and we depend on the pre-initialization
to zero to avoid known BIOS bugs.
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
H. Peter Anvin [Thu, 23 Dec 2010 20:04:48 +0000 (12:04 -0800)]
NEWS: Document MBR fix.
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
H. Peter Anvin [Thu, 23 Dec 2010 20:02:52 +0000 (12:02 -0800)]
mbr: Make sure the MBR code starts with the byte 0x33
Apparently some BIOSes (including some Acer Travelmate machines)
require an MBR to start with 0x33; apparently Micro$oft MBRs start
with 33 C0, an alternate coding of the "xorw %ax,%ax" instruction. As
such, follow suit to work on these braindead BIOSes.
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Gene Cumm [Tue, 21 Dec 2010 12:23:06 +0000 (07:23 -0500)]
core/diskboot.inc: fix bug; Sect1Ptr1_VAL
Gene Cumm [Tue, 21 Dec 2010 02:05:51 +0000 (21:05 -0500)]
core/diskstart.inc: Remove code copied to diskboot.inc
Gene Cumm [Tue, 21 Dec 2010 02:02:52 +0000 (21:02 -0500)]
core/diskboot.inc: The boot sector code from diskstart.inc
Split to allow it to be used by debugging/diagnostic images
Gene Cumm [Tue, 21 Dec 2010 01:53:39 +0000 (20:53 -0500)]
core/diskstart.inc: Sect1Ptr values moved to constants
This allows another program to re-use the boot sector and define
different values, once split.
Gene Cumm [Tue, 21 Dec 2010 01:39:11 +0000 (20:39 -0500)]
core/diskstart.inc: Use xint13 when it saves space or protects
Additional note on where it will cost to possibly not protect anything
Gene Cumm [Tue, 21 Dec 2010 01:24:29 +0000 (20:24 -0500)]
core/diskstart.inc: Wrap INT13h in xint13 to preserve ES
Sometimes INT13h AH08h kills ES to return info about a "floppy" (since
an unpartitioned HDD might look like a floppy). Also, change to tabs
on previous line.
H. Peter Anvin [Fri, 17 Dec 2010 03:45:57 +0000 (19:45 -0800)]
core: Add workaround in the Makefile for GNU Make 3.82 bug
Apparently GNU Make 3.82 mishandles pattern rules with fixed
dependencies, sigh.
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Sebastian Herbszt [Tue, 27 Jul 2010 20:32:54 +0000 (22:32 +0200)]
core: remove HAS_LOCALBOOT
HAS_LOCALBOOT is set unconditionally in config.inc.
Signed-off-by: Sebastian Herbszt <herbszt@gmx.de>
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
H. Peter Anvin [Tue, 14 Dec 2010 17:55:12 +0000 (09:55 -0800)]
NEWS: document fix for non-partitioned devices.
H. Peter Anvin [Tue, 14 Dec 2010 17:53:42 +0000 (09:53 -0800)]
core, diskstart: add more sanity checks for handover info
In particular, somehow we failed to actually test for DS:SI = 0:0!
This tests both for SI < 16 and DS:SI < 1024. The former of those
tests is a bit iffy and probably should be removed.
This should also be promoted into isohdpfx.
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
H. Peter Anvin [Tue, 14 Dec 2010 17:51:33 +0000 (09:51 -0800)]
syslinux.ld: ld 2.20.51 seems to want .bss16 explicitly (NOLOAD)
ld 2.20.51 seems to want the .bss16 explicitly marked (NOLOAD). We do
that for most of the other bss sections, so we might as well.
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
H. Peter Anvin [Wed, 8 Dec 2010 00:35:01 +0000 (16:35 -0800)]
memdiskfind: abstract get page size so we can compile against klibc
klibc doesn't have sysconf(), because it involves a hideously
inefficient multiplex. Thus, if _SC_PAGESIZE is not defined, invoke
getpagesize(). Call the routine get_page_size() so it doesn't
conflict on a platform which has both sysconf(_SC_PAGESIZE) and
getpagesize().
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
H. Peter Anvin [Tue, 7 Dec 2010 00:02:12 +0000 (16:02 -0800)]
NEWS: further updates
H. Peter Anvin [Tue, 7 Dec 2010 00:00:58 +0000 (16:00 -0800)]
Merge remote branch 'genec/news-for-hpa'
H. Peter Anvin [Mon, 6 Dec 2010 23:54:24 +0000 (15:54 -0800)]
com32: fix a ffile descriptor leak on open() of a nonexistent file
If we try to open a nonexistent file, free the resulting file
descriptor.
Reported-by: Antonio Carlini <arcarlini@iee.org>
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
Sebastian Herbszt [Sun, 5 Dec 2010 20:43:45 +0000 (21:43 +0100)]
gfxboot: require LABEL keywords
Require LABEL keywords in config file.
Signed-off-by: Sebastian Herbszt <herbszt@gmx.de>
Sebastian Herbszt [Sun, 5 Dec 2010 20:34:42 +0000 (21:34 +0100)]
gfxboot: support TEXT and ENDTEXT keywords
Support TEXT and ENDTEXT keywords; ignore the help text.
Signed-off-by: Sebastian Herbszt <herbszt@gmx.de>
Gene Cumm [Sat, 4 Dec 2010 02:18:08 +0000 (21:18 -0500)]
NEWS: Update for 4.04 as of 4.04-pre1
H. Peter Anvin [Fri, 3 Dec 2010 00:04:35 +0000 (16:04 -0800)]
Bump version to 4.04
H. Peter Anvin [Thu, 18 Nov 2010 20:08:47 +0000 (12:08 -0800)]
man: remove long-since-obsolete note about name mangling
We haven't mangled label names for a very long time...
H. Peter Anvin [Sun, 14 Nov 2010 21:14:49 +0000 (13:14 -0800)]
core, pxe: Fix handling of unqualified DNS names
Actually append the domain name to the end of an unqualified DNS name
like we should have done all along.
Reported-by: Gene Cumm <gene.cumm@gmail.com>
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
H. Peter Anvin [Sun, 14 Nov 2010 21:13:25 +0000 (13:13 -0800)]
memdisk: use char array for external symbols
Use "extern const char foo[];" for an external symbol rather than
"extern void" (gcc 4.5 fix.)
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
H. Peter Anvin [Sun, 14 Nov 2010 21:12:28 +0000 (13:12 -0800)]
gfxboot: use a character array for an absolute symbol
The best way to access an absolute symbol is "extern const char
foo[];" -- older gccs let you get away with "extern void" but gcc 4.5
chokes on it.
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
H. Peter Anvin [Sun, 14 Nov 2010 21:11:30 +0000 (13:11 -0800)]
disklib: remove improper "inline"
Remove improper inlines from static functions which cause gcc 4.5 to
abort due to inlining failure.
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
H. Peter Anvin [Sun, 14 Nov 2010 21:09:48 +0000 (13:09 -0800)]
com32: add a centralized bitops header
Add a centralized bitops header <sys/bitops.h> which uses x86 bitops
instructions. This is necessary to keep gcc 4.5 from aborting
compilation due to the inlined code being larger than the non-inlined
version, and well, we should really use the bitops.
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
H. Peter Anvin [Sun, 14 Nov 2010 20:51:07 +0000 (12:51 -0800)]
Merge branch 'genec-rosh-for-hpa', remote branches 'genec/sample-make-for-hpa', 'genec/drmk-ver-for-hpa' and 'genec/pxechain-for-hpa'
Gene Cumm [Sat, 13 Nov 2010 16:29:08 +0000 (11:29 -0500)]
PXELINUX: shorten code in pxenv call
Now that both scenarios are clear (do/do not call timer_cleanup/
timer_init), shorten it to eliminate a needless jmp.
Gene Cumm [Sat, 13 Nov 2010 16:22:54 +0000 (11:22 -0500)]
PXELINUX: add gPXE PXENV_FILE_EXEC in addition to PXENV_RESTART_TFTP; comments
Both of these calls will not return if successful
Gene Cumm [Sat, 6 Nov 2010 12:09:15 +0000 (08:09 -0400)]
ver.com: Update some of how DRMK is handled
Gene Cumm [Fri, 5 Nov 2010 17:32:52 +0000 (13:32 -0400)]
PXELINUX: Fix timer bug
Certain calls to the PXE stack could potentially remove us from
memory. Revert the INT 1Ch timer to its previous state.
Gene Cumm [Wed, 3 Nov 2010 02:52:09 +0000 (22:52 -0400)]
ver.com: Fix month calculation for DRMK Build Date
Gene Cumm [Mon, 25 Oct 2010 17:01:24 +0000 (13:01 -0400)]
sample/Makefile: re-add the include
Without the include, .ppm.gz to .lss messes up as $(GZIPPROG) is
not expanded properly
Gene Cumm [Mon, 25 Oct 2010 16:59:47 +0000 (12:59 -0400)]
modules/Makefile: add ver.com
Gene Cumm [Mon, 25 Oct 2010 16:56:36 +0000 (12:56 -0400)]
ver.com: Fixed width number output; Conditional skip DOS serial
Create two functions, writedecb[23], to output a byte AL as fixed with of
2 or 3 characters wide, 0 prepend
If all of the DOS Serialnumber is 0, don't bother with printing anything.
Gene Cumm [Mon, 25 Oct 2010 00:03:23 +0000 (20:03 -0400)]
ver.com: Compact file by reuse of string
Gene Cumm [Sun, 24 Oct 2010 21:30:30 +0000 (17:30 -0400)]
ver.com: Update to pull info from DRMK
Gene Cumm [Sun, 24 Oct 2010 21:15:41 +0000 (17:15 -0400)]
core/writedec: Remove writechr; fix overflow in .cloop
writechr is found in other files and this prevents a compile error.
Gene Cumm [Sun, 24 Oct 2010 14:14:53 +0000 (10:14 -0400)]
ver.com: Import from previously private branch as new file
Colin Watson [Wed, 20 Oct 2010 19:23:02 +0000 (21:23 +0200)]
gfxboot: fix buffer overrun when loading kernel/initramfs
If the file size wasn't a multiple of 64KB, we could overwrite the next
entry in the malloc arena so reading the initramfs would fail.
Signed-off-by: Colin Watson <cjwatson@ubuntu.com>
Signed-off-by: Sebastian Herbszt <herbszt@gmx.de>
Gene Cumm [Sat, 16 Oct 2010 12:07:02 +0000 (08:07 -0400)]
rosh: rosh_argcat now has length. if DO_DEBUG, show startup command
Gene Cumm [Sat, 16 Oct 2010 11:58:31 +0000 (07:58 -0400)]
rosh: display command entered and a command unknown message when not known
H. Peter Anvin [Fri, 15 Oct 2010 23:41:02 +0000 (16:41 -0700)]
Merge remote branch 'liu/master'
Gene Cumm [Fri, 15 Oct 2010 03:50:58 +0000 (23:50 -0400)]
chain.c32: DRMK: Fix comment at top of source
Gene Cumm [Wed, 13 Oct 2010 23:51:06 +0000 (19:51 -0400)]
chain.c32: DRMK: Add comment about only using a DWORD for partition offset
Gene Cumm [Wed, 13 Oct 2010 23:39:19 +0000 (19:39 -0400)]
chain.c32: DRMK: clean up output; clean up and correct comments
Gene Cumm [Mon, 9 Aug 2010 16:34:32 +0000 (12:34 -0400)]
chain.c: Progress on DRMK and notes that it's not perfect yet.
Gene Cumm [Wed, 13 Oct 2010 20:32:13 +0000 (16:32 -0400)]
Merge branch 'master' into rosh_for_hpa
Gene Cumm [Wed, 13 Oct 2010 20:28:53 +0000 (16:28 -0400)]
rosh: add another item in the ToDos comment
Colin Watson [Wed, 14 Jul 2010 13:11:14 +0000 (14:11 +0100)]
gfxboot: add include and menu include support
Ubuntu CDs use multiple configuration files and the include directive as
well as gfxboot, so it's useful for gfxboot to support these.
Signed-off-by: Colin Watson <cjwatson@ubuntu.com>
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Ferenc Wagner [Mon, 11 Oct 2010 15:55:52 +0000 (17:55 +0200)]
doc: extlinux falls back to syslinux.cfg
Signed-off-by: Ferenc Wagner <wferi@niif.hu>
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Ferenc Wagner [Mon, 11 Oct 2010 15:55:51 +0000 (17:55 +0200)]
doc: reorganize config file intro
Signed-off-by: Ferenc Wagner <wferi@niif.hu>
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Ferenc Wagner [Mon, 11 Oct 2010 15:55:50 +0000 (17:55 +0200)]
doc: comments don't need a leading space
Signed-off-by: Ferenc Wagner <wferi@niif.hu>
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
H. Peter Anvin [Wed, 13 Oct 2010 05:46:15 +0000 (22:46 -0700)]
chain.c32: make CHS calculation match core/fs/diskio.c
Use the same format for the CHS calculation as in core/fs/diskio.c;
also, apply the correct limits, and propagate the fix to the write
routine.
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
H. Peter Anvin [Wed, 13 Oct 2010 05:42:27 +0000 (22:42 -0700)]
strspn.c: drop "inline"
Drop "inline" from bit operations. These should be recoded to use x86
bts/bt instructions, but for now just drop the "inline" which causes
failures on gcc 4.5 due to inlining failures.
Reported-by: Sebastian Herbszt <herbszt@gmx.de>
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
H. Peter Anvin [Wed, 13 Oct 2010 05:40:46 +0000 (22:40 -0700)]
NEWS: Update for 4.03
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
P J P [Tue, 14 Sep 2010 06:38:37 +0000 (12:08 +0530)]
isohybrid: Use ftruncate instead of seek for final padding
Pad the image via ftruncate instead of seeking to the end (which was
done incorrectly).
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
H. Peter Anvin [Wed, 13 Oct 2010 05:31:09 +0000 (22:31 -0700)]
chain.c32: fix conversion from LBA to CHS
Bits [9:8] of the cylinder goes into the top bits of the sector
register, not the bottom (and the math for that was wrong, anyway.)
Reported-by: Michal Soltys <soltys@ziu.info>
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Gene Cumm [Sat, 9 Oct 2010 15:28:48 +0000 (11:28 -0400)]
rosh: Fix rosh_ls(); fix FTBFS on rosh_more(); add ROSH_DEBUG2_ARGV_V macro
Liu Aleaxander [Sat, 9 Oct 2010 15:10:14 +0000 (23:10 +0800)]
lib: Update memmem function
The current memem implementation can not handle the case that the length
of the third parameter(needle) is equal *ONE*. So that the current
strstr implemention doesn't work when the lenght of substring is ONE.
For example, strstr("linux", "l") it will return NULL instead of
"linux", which is not right.
This code is stolen from klibc ;)
Signed-off-by: Liu Aleaxander <Aleaxander@gmail.com>
Gene Cumm [Sat, 9 Oct 2010 12:32:46 +0000 (08:32 -0400)]
unistd.h: Add missing getopt declarations
Gene Cumm [Sat, 9 Oct 2010 11:41:12 +0000 (07:41 -0400)]
Merge branch 'master' into rosh_for_hpa
Gene Cumm [Sun, 3 Oct 2010 22:28:02 +0000 (18:28 -0400)]
rosh: Remove unneeded cases in rosh_command() switch statements
Gene Cumm [Sun, 3 Oct 2010 16:05:45 +0000 (12:05 -0400)]
rosh: Clean up rosh_command to remove old stuff
Gene Cumm [Sun, 3 Oct 2010 14:02:11 +0000 (10:02 -0400)]
rosh: code cleanup
Gene Cumm [Sun, 3 Oct 2010 13:24:38 +0000 (09:24 -0400)]
rosh: Change rosh_ls()/rosh_dir() to use argc/argv; use getopt()
Gene Cumm [Sun, 3 Oct 2010 13:15:52 +0000 (09:15 -0400)]
rosh: Change rosh_more()/rosh_less() to use argc/argv
Gene Cumm [Sun, 3 Oct 2010 11:22:59 +0000 (07:22 -0400)]
rosh: remove cmdstr from rosh_pwd()
Gene Cumm [Sun, 3 Oct 2010 11:17:54 +0000 (07:17 -0400)]
rosh: clear errno before possibly using it. might be set by a previous call
Gene Cumm [Sun, 3 Oct 2010 11:12:06 +0000 (07:12 -0400)]
rosh: Add argc/argv rosh_reboot(); Allow warm reboots
Gene Cumm [Sun, 3 Oct 2010 11:00:25 +0000 (07:00 -0400)]
rosh: In main, use string macro directly
Gene Cumm [Sun, 3 Oct 2010 10:59:44 +0000 (06:59 -0400)]
rosh: Change rosh_run() to use argc/argv
Gene Cumm [Sat, 2 Oct 2010 17:20:07 +0000 (13:20 -0400)]
rosh: Change rosh_help() to use the string passed to it directly
Gene Cumm [Sat, 2 Oct 2010 05:11:23 +0000 (01:11 -0400)]
rosh: ROSH_DEBUG_ARGV_V in header
Gene Cumm [Sat, 2 Oct 2010 05:10:49 +0000 (01:10 -0400)]
rosh: convert rosh_cat() to argc/argv
Gene Cumm [Sat, 2 Oct 2010 05:05:03 +0000 (01:05 -0400)]
rosh: Add function header comments that were missing
Gene Cumm [Sat, 2 Oct 2010 04:55:02 +0000 (00:55 -0400)]
rosh: Fix warning on rosh_ls_arg(); remove rosh_echo() for rosh_pr_argv()
Gene Cumm [Sat, 2 Oct 2010 04:43:24 +0000 (00:43 -0400)]
rosh: Rename function as it's verbose
rosh_pr_argv and ROSH_DEBUG_ARGV are verbose, printing the
argument number and single-quoted string out.
Gene Cumm [Sat, 2 Oct 2010 04:18:27 +0000 (00:18 -0400)]
rosh: closedir() only if actually opened
Gene Cumm [Sat, 2 Oct 2010 04:16:54 +0000 (00:16 -0400)]
rosh: compact previous commit
Gene Cumm [Sat, 2 Oct 2010 04:13:38 +0000 (00:13 -0400)]
rosh: Fix rosh_ls_arg() to prevent segfault on access denied
Gene Cumm [Sat, 2 Oct 2010 04:03:29 +0000 (00:03 -0400)]
rosh: main, prompt and command try to use argc/argv mostly
prompt still uses a string but parses it and passes the parsed
argc/argv pair to command. main doesn't need a string anymore
Gene Cumm [Fri, 1 Oct 2010 03:11:10 +0000 (23:11 -0400)]
rosh: rosh_cd fix DEBUG
Gene Cumm [Fri, 1 Oct 2010 03:08:24 +0000 (23:08 -0400)]
rosh: fix errors in previous commit
Gene Cumm [Fri, 1 Oct 2010 03:03:35 +0000 (23:03 -0400)]
rosh: change rosh_cd() to use argc/argv
Gene Cumm [Fri, 1 Oct 2010 02:41:21 +0000 (22:41 -0400)]
rosh: Use internal parser for argc/argv; use argc/argv in rosh_command
Gene Cumm [Fri, 24 Sep 2010 02:06:10 +0000 (22:06 -0400)]
rosh: start parse_args(); prep for use; make print_argv()
parse_args() for parsing a string into an argv/argc pair for easier
parsing and the potential use of getopt().
print_argv() for printing an argv/argc pair to show functionality
Gene Cumm [Tue, 20 Jul 2010 15:56:28 +0000 (11:56 -0400)]
rosh: fix commented out WS usage in linux to grab screen size
Gene Cumm [Fri, 9 Jul 2010 00:05:39 +0000 (20:05 -0400)]
rosh: add echo
H. Peter Anvin [Fri, 8 Oct 2010 20:46:26 +0000 (13:46 -0700)]
Merge remote branch 'genec/rosh_for_hpa'
Gene Cumm [Fri, 24 Sep 2010 01:51:14 +0000 (21:51 -0400)]
core: fix handling new directory "/" for CONFIG in ISOLINUX and EXTLINUX
Arwin Vosselman [Thu, 5 Aug 2010 20:17:30 +0000 (22:17 +0200)]
Fix missing option -s in libinstaller
In 'libinstaller/syslxopt.c' on line 71 the "s" (undercast) is missing in:
const char short_options[] = "t:fid:UuzS:H:rvho:OM:ma";
Adding the 's' there seems to solve our problem.
H. Peter Anvin [Thu, 26 Aug 2010 21:45:53 +0000 (14:45 -0700)]
pxe: add a "pxeretry" option to deal with 404's in web apps
For web downloads, sometimes a mirror site will not be fully synced.
Add an option to retry the open a specific number of times before
giving up.
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
H. Peter Anvin [Sat, 14 Aug 2010 00:27:44 +0000 (17:27 -0700)]
extlinux: try to get the disk offset from sysfs
It is possible(?) that HDIO_GETGEO can't return the full offset, and
in either case it is too small -- only 32 bits on 32-bit platforms.
Thus query sysfs for the real value, if available.
sysfs also contains information for slave devices -- this is going to
really matter for the md/lvm issues.
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
Don Hiatt [Fri, 13 Aug 2010 16:51:12 +0000 (09:51 -0700)]
diskio: remove redundant register write
The disk number is set twice in diskio.c/edd_rdwr_sectors() this patch
removes the redundant setting.