Bug#462218: Bug#461442: detection of other OSes in update-grub
Hi there,
Robert Millan schrieb:
If you want the former, feel free to send a patch to grub-devel. For the
latter, look at my questions and send some feedback.
please find attached a patch that adds a new parameter '--device, -d' to
grub-probe. If this parameter is set, grub-probe expects the given
argument to be a block device. All of the '--target' parameters work
with this option as well. If the '--device' parameter is not set,
grub-probe will work as before.
Furthermore I had to add a new public function to getroot.c that returns
the given argument if it is a block device and returns NULL else. This
was necessary, because else you could force grub-probe to print 'foobar'
if run as 'grub-probe --target=device --device foobar'.
I consider this (i.e. adding a new option to grub-probe) a better
solution than writing a separate tool only for this purpose.
The second file that I attached is the 30_os-prober script making use of
the new grub-probe feature. Please note that the 'case ... in' part is
only run if 'grub-probe --device' actually returns something usefull.
I am looking forward to your feedback.
Cheers,
Fabian
--
Dipl.-Phys. Fabian Greffrath
Ruhr-Universität Bochum
Lehrstuhl für Energieanlagen und Energieprozesstechnik (LEAT)
Universitätsstr. 150, IB 3/134
D-44780 Bochum
Telefon: +49 (0)234 / 32-26334
Fax: +49 (0)234 / 32-14227
E-Mail: [EMAIL PROTECTED]
diff -urNad grub2-1.95+20080201~/include/grub/util/getroot.h grub2-1.95+20080201/include/grub/util/getroot.h
--- grub2-1.95+20080201~/include/grub/util/getroot.h 2008-01-12 16:11:56.0 +0100
+++ grub2-1.95+20080201/include/grub/util/getroot.h 2008-02-03 18:14:11.0 +0100
@@ -29,5 +29,6 @@
char *grub_get_prefix (const char *dir);
int grub_util_get_dev_abstraction (const char *os_dev);
char *grub_util_get_grub_dev (const char *os_dev);
+char *grub_util_check_block_device (const char *blk_dev);
#endif /* ! GRUB_UTIL_GETROOT_HEADER */
diff -urNad grub2-1.95+20080201~/util/getroot.c grub2-1.95+20080201/util/getroot.c
--- grub2-1.95+20080201~/util/getroot.c 2008-01-12 16:11:56.0 +0100
+++ grub2-1.95+20080201/util/getroot.c 2008-02-03 21:56:29.0 +0100
@@ -327,3 +327,17 @@
return grub_dev;
}
+
+char *
+grub_util_check_block_device (const char *blk_dev)
+{
+ struct stat st;
+
+ if (stat (blk_dev, st) 0)
+grub_util_error (Cannot stat `%s', blk_dev);
+
+ if (S_ISBLK (st.st_mode))
+return strdup(blk_dev);
+ else
+return 0;
+}
diff -urNad grub2-1.95+20080201~/util/grub-probe.c grub2-1.95+20080201/util/grub-probe.c
--- grub2-1.95+20080201~/util/grub-probe.c 2008-01-25 23:33:57.0 +0100
+++ grub2-1.95+20080201/util/grub-probe.c 2008-02-03 19:30:50.0 +0100
@@ -50,6 +50,7 @@
};
int print = PRINT_FS;
+unsigned int argument_is_device = 0;
void
grub_putchar (int c)
@@ -84,9 +85,18 @@
int abstraction_type;
grub_device_t dev = NULL;
- device_name = grub_guess_root_device (path);
+ if (argument_is_device)
+device_name = grub_util_check_block_device (path);
+ else
+device_name = grub_guess_root_device (path);
+
if (! device_name)
-grub_util_error (cannot find a device for %s.\n, path);
+{
+ if (argument_is_device)
+grub_util_error (%s is not a block device.\n, path);
+ else
+grub_util_error (cannot find a device for %s.\n, path);
+}
if (print == PRINT_DEVICE)
{
@@ -201,6 +211,7 @@
static struct option options[] =
{
+{device, no_argument, 0, 'd'},
{device-map, required_argument, 0, 'm'},
{target, required_argument, 0, 't'},
{help, no_argument, 0, 'h'},
@@ -217,10 +228,11 @@
Try ``grub-probe --help'' for more information.\n);
else
printf (\
-Usage: grub-probe [OPTION]... PATH\n\
+Usage: grub-probe [OPTION]... [PATH|DEVICE]\n\
\n\
-Probe device information for a given path.\n\
+Probe device information for a given path or device.\n\
\n\
+ -d, --device given argument is a system device, not a path\n\
-m, --device-map=FILE use FILE as the device map [default=%s]\n\
-t, --target=(fs|drive|device|partmap|abstraction)\n\
print filesystem module, GRUB drive, system device, partition map module or abstraction module [default=fs]\n\
@@ -246,13 +258,17 @@
/* Check for options. */
while (1)
{
- int c = getopt_long (argc, argv, m:t:hVv, options, 0);
+ int c = getopt_long (argc, argv, dm:t:hVv, options, 0);
if (c == -1)
break;
else
switch (c)
{
+ case 'd':
+ argument_is_device = 1;
+ break;
+
case 'm':
if (dev_map)
free (dev_map);
#! /bin/sh -e
# update-grub helper script.
# insert copyright and license blurb here
if [ -x `which os-prober 2/dev/null` ] ; then
OSPROBED=`os-prober | tr ' ' '|' | paste -s -d ' '`
fi
if [ -n ${OSPROBED} ] ; then
for OS in ${OSPROBED} ; do
DEVICE=`echo ${OS} | cut -d ':' -f 1`
Bug#462218: Bug#461442: detection of other OSes in update-grub
On Mon, Feb 04, 2008 at 03:47:26PM +0100, Fabian Greffrath wrote: Robert Millan schrieb: Why is that a problem? Because foobar is not a block device, but grub-probe claims that it will print a device if '--target=device' is given. Uhm I'm not sure if that's a good thing or may be overkill. What does everyone else think about this? I know that the call to probe() is not supposed to be reentrant, but I'd prefer not to break reentrancy if it can be easily avoided; it is possible that probe() needs to recurse onto itself in the future (because of RAID/LVM). OK, but should I keep it uninitialized? Depends on what you want to do ;-) Is it possible to share code with 10_linux.in here? Only if os-prober is installed. But then, os-prober does not check for kernels on / and /target. What I mean is that 10_linux.in already has code that probes for stuff like Linux images and initrd files, and it might be a good idea to use the same logic. Or does os-prober provide any additional feature/knowledge about this? -- Robert Millan GPLv2 I know my rights; I want my phone call! DRM What use is a phone call… if you are unable to speak? (as seen on /.) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#462218: Bug#461442: detection of other OSes in update-grub
Robert Millan schrieb: Why is that a problem? Because foobar is not a block device, but grub-probe claims that it will print a device if '--target=device' is given. I think this function could be called from the other part of this file which performs similar checks (if this functionality is to be kept, that is). Yes, I am fine with this. I know that the call to probe() is not supposed to be reentrant, but I'd prefer not to break reentrancy if it can be easily avoided; it is possible that probe() needs to recurse onto itself in the future (because of RAID/LVM). OK, but should I keep it uninitialized? Please remember to fix that in later versions of the patch ;-) Sure, my copyright assignment paper is on it's way... Is it possible to share code with 10_linux.in here? Only if os-prober is installed. But then, os-prober does not check for kernels on / and /target. Maybe this can be simplified with echo something | read a b c d feature? For cosmetic reasons, yes. ;) Cheers, Fabian -- Dipl.-Phys. Fabian Greffrath Ruhr-Universität Bochum Lehrstuhl für Energieanlagen und Energieprozesstechnik (LEAT) Universitätsstr. 150, IB 3/134 D-44780 Bochum Telefon: +49 (0)234 / 32-26334 Fax: +49 (0)234 / 32-14227 E-Mail: [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#462218: Bug#461442: detection of other OSes in update-grub
On Mon, Feb 04, 2008 at 09:43:58AM +0100, Fabian Greffrath wrote:
the given argument if it is a block device and returns NULL else. This
was necessary, because else you could force grub-probe to print 'foobar'
if run as 'grub-probe --target=device --device foobar'.
Why is that a problem?
--- grub2-1.95+20080201~/util/getroot.c 2008-01-12 16:11:56.0
+0100
+++ grub2-1.95+20080201/util/getroot.c2008-02-03 21:56:29.0
+0100
@@ -327,3 +327,17 @@
return grub_dev;
}
+
+char *
+grub_util_check_block_device (const char *blk_dev)
+{
+ struct stat st;
+
+ if (stat (blk_dev, st) 0)
+grub_util_error (Cannot stat `%s', blk_dev);
+
+ if (S_ISBLK (st.st_mode))
+return strdup(blk_dev);
Missing space here. ^
+ else
+return 0;
+}
I think this function could be called from the other part of this file which
performs similar checks (if this functionality is to be kept, that is).
diff -urNad grub2-1.95+20080201~/util/grub-probe.c
grub2-1.95+20080201/util/grub-probe.c
--- grub2-1.95+20080201~/util/grub-probe.c2008-01-25 23:33:57.0
+0100
+++ grub2-1.95+20080201/util/grub-probe.c 2008-02-03 19:30:50.0
+0100
@@ -50,6 +50,7 @@
};
int print = PRINT_FS;
+unsigned int argument_is_device = 0;
I know that the call to probe() is not supposed to be reentrant, but I'd
prefer not to break reentrancy if it can be easily avoided; it is possible
that probe() needs to recurse onto itself in the future (because of RAID/LVM).
#! /bin/sh -e
# update-grub helper script.
# insert copyright and license blurb here
Please remember to fix that in later versions of the patch ;-)
linux)
if [ -x `which linux-boot-prober 2/dev/null` ] ; then
LINUXPROBED=`linux-boot-prober ${DEVICE} | tr ' ' '|' | paste -s
-d ' '`
fi
Is it possible to share code with 10_linux.in here?
if [ -n ${LINUXPROBED} ] ; then
for LINUX in ${LINUXPROBED} ; do
LROOT=`echo ${LINUX} | cut -d ':' -f 1`
LBOOT=`echo ${LINUX} | cut -d ':' -f 2`
LLABEL=`echo ${LINUX} | cut -d ':' -f 3 | tr '|' ' '`
LKERNEL=`echo ${LINUX} | cut -d ':' -f 4`
LINITRD=`echo ${LINUX} | cut -d ':' -f 5`
LPARAMS=`echo ${LINUX} | cut -d ':' -f 6 | tr '|' ' '`
Maybe this can be simplified with echo something | read a b c d feature?
--
Robert Millan
GPLv2 I know my rights; I want my phone call!
DRM What use is a phone call… if you are unable to speak?
(as seen on /.)
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#462218: Bug#461442: detection of other OSes in update-grub
Fabian Greffrath [EMAIL PROTECTED] writes:
Hi!
Robert Millan schrieb:
If you want the former, feel free to send a patch to grub-devel. For the
latter, look at my questions and send some feedback.
please find attached a patch that adds a new parameter '--device, -d'
to grub-probe. If this parameter is set, grub-probe expects the given
argument to be a block device. All of the '--target' parameters work
with this option as well. If the '--device' parameter is not set,
grub-probe will work as before.
Furthermore I had to add a new public function to getroot.c that
returns the given argument if it is a block device and returns NULL
else. This was necessary, because else you could force grub-probe to
print 'foobar' if run as 'grub-probe --target=device --device foobar'.
Thanks for this patch, however you didn't really describe the problem
you are solving. I am not sure if this patch replaces your patch.
This thread was initiated with a follow up to a mail Robert sent
off-list. So can you please describe what you are doing and why,
otherwise this is simply a piece of code to me that I have to review
;-)
So please explain what this is, I cannot accept patches if I do not
know why I should accept them.
I consider this (i.e. adding a new option to grub-probe) a better
solution than writing a separate tool only for this purpose.
The second file that I attached is the 30_os-prober script making use
of the new grub-probe feature. Please note that the 'case ... in' part
is only run if 'grub-probe --device' actually returns something
usefull.
I am looking forward to your feedback.
Here it is ;-)
First of all, a ChangeLog entry is missing.
Is this code only yours or did you use code from other places?
diff -urNad grub2-1.95+20080201~/include/grub/util/getroot.h
grub2-1.95+20080201/include/grub/util/getroot.h
--- grub2-1.95+20080201~/include/grub/util/getroot.h 2008-01-12
16:11:56.0 +0100
+++ grub2-1.95+20080201/include/grub/util/getroot.h 2008-02-03
18:14:11.0 +0100
@@ -29,5 +29,6 @@
char *grub_get_prefix (const char *dir);
int grub_util_get_dev_abstraction (const char *os_dev);
char *grub_util_get_grub_dev (const char *os_dev);
+char *grub_util_check_block_device (const char *blk_dev);
#endif /* ! GRUB_UTIL_GETROOT_HEADER */
diff -urNad grub2-1.95+20080201~/util/getroot.c
grub2-1.95+20080201/util/getroot.c
--- grub2-1.95+20080201~/util/getroot.c 2008-01-12 16:11:56.0
+0100
+++ grub2-1.95+20080201/util/getroot.c2008-02-03 21:56:29.0
+0100
@@ -327,3 +327,17 @@
return grub_dev;
}
+
+char *
+grub_util_check_block_device (const char *blk_dev)
+{
+ struct stat st;
+
+ if (stat (blk_dev, st) 0)
+grub_util_error (Cannot stat `%s', blk_dev);
+
+ if (S_ISBLK (st.st_mode))
+return strdup(blk_dev);
strdup (blk_dev);
+ else
+return 0;
+}
diff -urNad grub2-1.95+20080201~/util/grub-probe.c
grub2-1.95+20080201/util/grub-probe.c
--- grub2-1.95+20080201~/util/grub-probe.c2008-01-25 23:33:57.0
+0100
+++ grub2-1.95+20080201/util/grub-probe.c 2008-02-03 19:30:50.0
+0100
@@ -50,6 +50,7 @@
};
int print = PRINT_FS;
+unsigned int argument_is_device = 0;
Shouldn't this be static?
void
grub_putchar (int c)
@@ -84,9 +85,18 @@
int abstraction_type;
grub_device_t dev = NULL;
- device_name = grub_guess_root_device (path);
+ if (argument_is_device)
+device_name = grub_util_check_block_device (path);
+ else
+device_name = grub_guess_root_device (path);
+
if (! device_name)
-grub_util_error (cannot find a device for %s.\n, path);
+{
+ if (argument_is_device)
+grub_util_error (%s is not a block device.\n, path);
+ else
+grub_util_error (cannot find a device for %s.\n, path);
+}
if (print == PRINT_DEVICE)
{
@@ -201,6 +211,7 @@
static struct option options[] =
{
+{device, no_argument, 0, 'd'},
{device-map, required_argument, 0, 'm'},
{target, required_argument, 0, 't'},
{help, no_argument, 0, 'h'},
@@ -217,10 +228,11 @@
Try ``grub-probe --help'' for more information.\n);
else
printf (\
-Usage: grub-probe [OPTION]... PATH\n\
+Usage: grub-probe [OPTION]... [PATH|DEVICE]\n\
\n\
-Probe device information for a given path.\n\
+Probe device information for a given path or device.\n\
\n\
+ -d, --device given argument is a system device, not a path\n\
-m, --device-map=FILE use FILE as the device map [default=%s]\n\
-t, --target=(fs|drive|device|partmap|abstraction)\n\
print filesystem module, GRUB drive, system
device, partition map module or abstraction module [default=fs]\n\
@@ -246,13 +258,17 @@
/* Check for options. */
while (1)
{
- int c = getopt_long (argc, argv, m:t:hVv, options, 0);
+ int c = getopt_long (argc,
Bug#462218: Bug#461442: detection of other OSes in update-grub
On Thu, Jan 31, 2008 at 09:31:57AM +0100, Fabian Greffrath wrote: Robert Millan schrieb: Suggestions? For the time being I could duplicate the convert() function from Debian's grub-installer to convert the device names. However, I believe that there are plenty of cases in which you could use a generic device name conversion tool... Sorry, I got confused. The generic linux-device / grub-drive conversion tool is no problem. I think it's fine to add it. The problem is if you need grub-probe to work with an input different than a filesystem path. If you want the former, feel free to send a patch to grub-devel. For the latter, look at my questions and send some feedback. Or, if nobody wants to send feedback, I guess I'll have to propose something myself. But I prefer if this is discussed (given that we have time). -- Robert Millan GPLv2 I know my rights; I want my phone call! DRM What use is a phone call… if you are unable to speak? (as seen on /.) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#462218: Bug#461442: detection of other OSes in update-grub
On Thu, Jan 31, 2008 at 03:26:24PM +0100, Fabian Greffrath wrote: Robert Millan schrieb: Sorry, I got confused. The generic linux-device / grub-drive conversion tool is no problem. I think it's fine to add it. Fine! The problem is if you need grub-probe to work with an input different than a filesystem path. I won't need grub-probe at all as soon as the aforementioned tool exists. (My initial idea was to integrate the functionality into grub-probe, but a separate tool is fine as well.) If you want the former, feel free to send a patch to grub-devel. Hey, I'm allready proud of my little shell script... ;) For the latter, look at my questions and send some feedback. Sorry, Robert, but which questions? It is described earlier in this thread, but it doesn't matter. If adding a drive/device conversion utility is enough for you, please can you send a patch for it? You could even merge it with the os-prober one and send it altogether, if that's more practical for you. Thank you! -- Robert Millan GPLv2 I know my rights; I want my phone call! DRM What use is a phone call… if you are unable to speak? (as seen on /.) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#462218: Bug#461442: detection of other OSes in update-grub
Robert Millan schrieb: Sorry, I got confused. The generic linux-device / grub-drive conversion tool is no problem. I think it's fine to add it. Fine! The problem is if you need grub-probe to work with an input different than a filesystem path. I won't need grub-probe at all as soon as the aforementioned tool exists. (My initial idea was to integrate the functionality into grub-probe, but a separate tool is fine as well.) If you want the former, feel free to send a patch to grub-devel. Hey, I'm allready proud of my little shell script... ;) For the latter, look at my questions and send some feedback. Sorry, Robert, but which questions? -- Dipl.-Phys. Fabian Greffrath Ruhr-Universität Bochum Lehrstuhl für Energieanlagen und Energieprozesstechnik (LEAT) Universitätsstr. 150, IB 3/134 D-44780 Bochum Telefon: +49 (0)234 / 32-26334 Fax: +49 (0)234 / 32-14227 E-Mail: [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#462218: Bug#461442: detection of other OSes in update-grub
Robert Millan schrieb: Suggestions? For the time being I could duplicate the convert() function from Debian's grub-installer to convert the device names. However, I believe that there are plenty of cases in which you could use a generic device name conversion tool... -- Dipl.-Phys. Fabian Greffrath Ruhr-Universität Bochum Lehrstuhl für Energieanlagen und Energieprozesstechnik (LEAT) Universitätsstr. 150, IB 3/134 D-44780 Bochum Telefon: +49 (0)234 / 32-26334 Fax: +49 (0)234 / 32-14227 E-Mail: [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#462218: Bug#461442: detection of other OSes in update-grub
Fabian Greffrath [EMAIL PROTECTED] writes: Robert Millan schrieb: Suggestions? For the time being I could duplicate the convert() function from Debian's grub-installer to convert the device names. It could be nice so it could be tested on Debian while it's not accepted upstream. However, I believe that there are plenty of cases in which you could use a generic device name conversion tool... Sure but at least we can try it now and keep working in upstream to remove this change. -- O T A V I OS A L V A D O R - E-mail: [EMAIL PROTECTED] UIN: 5906116 GNU/Linux User: 239058 GPG ID: 49A5F855 Home Page: http://otavio.ossystems.com.br - Microsoft sells you Windows ... Linux gives you the whole house. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#462218: Bug#461442: detection of other OSes in update-grub
This implies that we'll have to support 6 different kinds of conversions between path, drive and device, some of which might even be impossible. Besides, perhaps it'd be cleaner to split this conversion to a separate tool, so that grub-probe only operates on devices, and the other tool converts paths to devices (but NOT necessarily devices to paths!). What does everyone think about this? Yes, it is really necessary! How about this proposal: $ grub-convert Usage: grub-convert [-d|-g] [DEVICE|DRIVE] Convert back and forth between Linux devices and GRUB drives. -d, --device=DEVICE convert DEVICE into a GRUB-drive [e.g. /dev/hda1 - (hd0,1)] -g, --drive=DRIVE convert DRIVE into a Linux device [e.g. (hd0,1) - /dev/hda1] -h, --helpdisplay this message and exit -V, --version print version information and exit Cheers, Fabian -- Dipl.-Phys. Fabian Greffrath Ruhr-Universität Bochum Lehrstuhl für Energieanlagen und Energieprozesstechnik (LEAT) Universitätsstr. 150, IB 3/134 D-44780 Bochum Telefon: +49 (0)234 / 32-26334 Fax: +49 (0)234 / 32-14227 E-Mail: [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#462218: Bug#461442: detection of other OSes in update-grub
On Wed, Jan 30, 2008 at 11:41:59AM +0100, Fabian Greffrath wrote: This implies that we'll have to support 6 different kinds of conversions between path, drive and device, some of which might even be impossible. Besides, perhaps it'd be cleaner to split this conversion to a separate tool, so that grub-probe only operates on devices, and the other tool converts paths to devices (but NOT necessarily devices to paths!). Actually, this can't work now. After I added a sanity check (sorry, this was necessary) to verify file reading work properly, everything is more intermangled. I'm not sure what would be the right approach. Perhaps a flag in grub-probe that means please skip grub_guess_root_device, argument is a device, not a path! ? This will let the sanity check know when it can't tell which file you want to read. Or perhaps something more radical that splits everything in different layers (of exposed user commands) ? But that could clutter the user with so many commands. Suggestions? How about this proposal: $ grub-convert Usage: grub-convert [-d|-g] [DEVICE|DRIVE] -- Robert Millan GPLv2 I know my rights; I want my phone call! DRM What use is a phone call… if you are unable to speak? (as seen on /.) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#461442: detection of other OSes in update-grub
retitle 461442 please add support for os-prober to detect other OSes in update-grub block 461442 by 462218 tags 462218 + upstream forwarded 462218 http://lists.gnu.org/archive/html/grub-devel/2008-01/msg00446.html thanks -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#462218: Bug#461442: detection of other OSes in update-grub
On Wed, Jan 23, 2008 at 09:55:12AM +0100, Fabian Greffrath wrote: retitle 461442 please add support for os-prober to detect other OSes in update-grub block 461442 by 462218 tags 462218 + upstream forwarded 462218 http://lists.gnu.org/archive/html/grub-devel/2008-01/msg00446.html thanks What we forwarded is the bug about grub-probe change. I'm not sure if it makes sense to merge os-prober support in upstream, although I'm open to discussion. -- Robert Millan GPLv2 I know my rights; I want my phone call! DRM What use is a phone call… if you are unable to speak? (as seen on /.) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#461442: detection of other OSes in update-grub
Robert Millan [EMAIL PROTECTED] writes: On Wed, Jan 23, 2008 at 09:55:12AM +0100, Fabian Greffrath wrote: retitle 461442 please add support for os-prober to detect other OSes in update-grub block 461442 by 462218 tags 462218 + upstream forwarded 462218 http://lists.gnu.org/archive/html/grub-devel/2008-01/msg00446.html thanks What we forwarded is the bug about grub-probe change. I'm not sure if it makes sense to merge os-prober support in upstream, although I'm open to discussion. Ack here. os-probe is too Debian specific (or better, linux specific) and I think it could be use by other distros but shouldn't be part of grub itself. -- O T A V I OS A L V A D O R - E-mail: [EMAIL PROTECTED] UIN: 5906116 GNU/Linux User: 239058 GPG ID: 49A5F855 Home Page: http://otavio.ossystems.com.br - Microsoft sells you Windows ... Linux gives you the whole house. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#461442: detection of other OSes in update-grub
Otavio Salvador schrieb: Ack here. os-probe is too Debian specific (or better, linux specific) and I think it could be use by other distros but shouldn't be part of grub itself. Ack, too. The 'forwarded' tag was only set to the 462218 bug, which is about grub-probe. The os-prober related bug has not been forwarded. I just mentioned os-prober as a motivation for the 'blocks' tag. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#461442: detection of other OSes in update-grub
On Wed, Jan 23, 2008 at 09:28:02AM -0200, Otavio Salvador wrote: Robert Millan [EMAIL PROTECTED] writes: On Wed, Jan 23, 2008 at 09:55:12AM +0100, Fabian Greffrath wrote: retitle 461442 please add support for os-prober to detect other OSes in update-grub block 461442 by 462218 tags 462218 + upstream forwarded 462218 http://lists.gnu.org/archive/html/grub-devel/2008-01/msg00446.html thanks What we forwarded is the bug about grub-probe change. I'm not sure if it makes sense to merge os-prober support in upstream, although I'm open to discussion. Ack here. os-probe is too Debian specific (or better, linux specific) and I think it could be use by other distros but shouldn't be part of grub itself. I was referring to the /etc/grub.d/ glue that Fabian is writing. If os-prober is used by other distros, I don't see why not pushing it upstream. What do you think, Fabian? -- Robert Millan GPLv2 I know my rights; I want my phone call! DRM What use is a phone call… if you are unable to speak? (as seen on /.) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#461442: detection of other OSes in update-grub
Robert Millan [EMAIL PROTECTED] writes: On Wed, Jan 23, 2008 at 09:28:02AM -0200, Otavio Salvador wrote: Robert Millan [EMAIL PROTECTED] writes: On Wed, Jan 23, 2008 at 09:55:12AM +0100, Fabian Greffrath wrote: retitle 461442 please add support for os-prober to detect other OSes in update-grub block 461442 by 462218 tags 462218 + upstream forwarded 462218 http://lists.gnu.org/archive/html/grub-devel/2008-01/msg00446.html thanks What we forwarded is the bug about grub-probe change. I'm not sure if it makes sense to merge os-prober support in upstream, although I'm open to discussion. Ack here. os-probe is too Debian specific (or better, linux specific) and I think it could be use by other distros but shouldn't be part of grub itself. I was referring to the /etc/grub.d/ glue that Fabian is writing. If os-prober is used by other distros, I don't see why not pushing it upstream. What do you think, Fabian? os-probe would be hard to get accepted since it couldn't be relicensed to FSF without much hassle. -- O T A V I OS A L V A D O R - E-mail: [EMAIL PROTECTED] UIN: 5906116 GNU/Linux User: 239058 GPG ID: 49A5F855 Home Page: http://otavio.ossystems.com.br - Microsoft sells you Windows ... Linux gives you the whole house. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#461442: detection of other OSes in update-grub
On Wed, Jan 23, 2008 at 01:10:03PM +0100, Fabian Greffrath wrote: I think, we should keep the script in the Debian package for the time beeing. Before it's included upstream we should see to get os-prober included in other distributions and make sure we can properly add other Linux/Hurd installations to grub.cfg based on os-prober's output. It is difficult to merge things in upstream when they haven't been writen purposely for upstream. This is a decision that has to be taken now. -- Robert Millan GPLv2 I know my rights; I want my phone call! DRM What use is a phone call… if you are unable to speak? (as seen on /.) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#462218: Bug#461442: detection of other OSes in update-grub
On Wed, Jan 23, 2008 at 10:01:59AM -0200, Otavio Salvador wrote: I was referring to the /etc/grub.d/ glue that Fabian is writing. If os-prober is used by other distros, I don't see why not pushing it upstream. What do you think, Fabian? os-probe would be hard to get accepted since it couldn't be relicensed to FSF without much hassle. Sorry, I didn't explain well. I'm NOT referring to os-prober itself, but to the glue that Fabian is writing. -- Robert Millan GPLv2 I know my rights; I want my phone call! DRM What use is a phone call… if you are unable to speak? (as seen on /.) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#462218: Bug#461442: detection of other OSes in update-grub
Robert Millan schrieb: I was referring to the /etc/grub.d/ glue that Fabian is writing. If os-prober is used by other distros, I don't see why not pushing it upstream. What do you think, Fabian? For me personally it makes no difference if the script ends up in Debian or upstream. However, the first thing the script does is if [ -x `which os-prober 2/dev/null` ] ; then (...) so it won't hurt anybody without os-prober installed. I have to stress that my script is currently (well, as soon as #462218 is closed) only able to add chainloaded partitions. For all other kinds of Linux/Hurd installations it would need to remount the partitions to find out where the linux-/initrd-images reside. This is something I want to avoid, because the partitions have allready been mounted by os-prober and I don't want to duplicate it's code for this purpose. Maybe os-prober's output can be modified to be more verbose about other Linux/Hurd systems it detects. However, we will be able to detect the majority of preinstalled OSes from Redmond. ;) I think, we should keep the script in the Debian package for the time beeing. Before it's included upstream we should see to get os-prober included in other distributions and make sure we can properly add other Linux/Hurd installations to grub.cfg based on os-prober's output. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#462218: Bug#461442: detection of other OSes in update-grub
Robert Millan schrieb:
It is difficult to merge things in upstream when they haven't been writen
purposely for upstream. This is a decision that has to be taken now.
I believe I am not the one to decide about this.
Please find attached my script proposal. It's only 1.5 kB, tab-indented
and full of comments and paranoid test. I'd like to hand it over to the
GRUB-maintainers (including copyright!) for further nursing. ;)
Cheers,
Fabian
#! /bin/sh -e
# update-grub helper script.
# insert copyright and license blurb here
if [ -x `which os-prober 2/dev/null` ] ; then
# os-prober output contains the partition and the name, a label and
# a boot keyword, separated by colons (:) for each detected OS.
# It may contain space characters and linebreaks,
# e.g. /dev/hda2:Windows XP Professional:Windows:chain.
# Convert space characters to underscores and linebreaks to spaces.
RESULT=`os-prober | tr ' ' '_' | tr '\n' ' '`
fi
if [ x${RESULT} != x ] ; then
for OS in ${RESULT} ; do
# e.g. /dev/hda2:Windows_XP_Professional:Windows:chain
PARTITION=`echo ${OS} | cut -d ':' -f 1`
LONGNAME=`echo ${OS} | cut -d ':' -f 2 | tr '_' ' '`
LABEL=`echo ${OS} | cut -d ':' -f 3`
BOOT=`echo ${OS} | cut -d ':' -f 4`
if [ x${LONGNAME} = x ] ; then
# is this likely to happen?
LONGNAME=${LABEL}
fi
case ${BOOT} in
chain)
echo Found ${LONGNAME} on ${PARTITION} 2
if [ -x `which grub-probe 2/dev/null` ] ;
then
# need to convert partition device to
GRUB drive here!
GRUB_DEVICE=`grub-probe --something
${PARTITION}`
cat EOF
menuentry ${LONGNAME} (on ${PARTITION}) {
set root=${GRUB_DEVICE}
chainloader +1
}
EOF
else # wtf?!
echo Error: Missing grub-probe! 2
fi
;;
macos|macosx)
# can't macos* also be chainloaded?
;;
hurd|linux)
# other Linux/HURD-Systems are not (yet)
supported
# with the given output of os-prober
;;
*)
# is it possible to reach here?
;;
esac
done
fi
Bug#462218: Bug#461442: detection of other OSes in update-grub
On Wed, Jan 23, 2008 at 02:28:36PM +0100, Fabian Greffrath wrote: Robert Millan schrieb: It is difficult to merge things in upstream when they haven't been writen purposely for upstream. This is a decision that has to be taken now. I believe I am not the one to decide about this. Please find attached my script proposal. It's only 1.5 kB, tab-indented and full of comments and paranoid test. I'd like to hand it over to the GRUB-maintainers (including copyright!) for further nursing. ;) Nice. Can you please send this to grub-devel ? (explaining the context) -- Robert Millan GPLv2 I know my rights; I want my phone call! DRM What use is a phone call… if you are unable to speak? (as seen on /.) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#461442: detection of other OSes in update-grub
Robert Millan schrieb: Nice. Can you please send this to grub-devel ? (explaining the context) NB: So you consider it for upstream inclusion? -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#462218: Bug#461442: detection of other OSes in update-grub
On Wed, Jan 23, 2008 at 04:00:44PM +0100, Fabian Greffrath wrote: Robert Millan schrieb: Nice. Can you please send this to grub-devel ? (explaining the context) NB: So you consider it for upstream inclusion? I don't drive the train in upstream, but in principle I think it's fine. -- Robert Millan GPLv2 I know my rights; I want my phone call! DRM What use is a phone call… if you are unable to speak? (as seen on /.) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#461442: detection of other OSes in update-grub
reopen 461442 thanks On Tue, Jan 22, 2008 at 10:15:33AM +0100, Fabian Greffrath wrote: Robert Millan schrieb: - os-prober is hooked in /etc/grub.d/ to run every time. I'll go this way. Patches welcome... I am nearly done. All I need is an algorithm to convert device names into GRUB devices, since os-prober returns partitions like '/dev/hda2'. It's a pity that grub-probe can only return this information for mounted devices. The partitions were allready mounted by os-prober and I do not want to mount them again only to receive their GRUB device name. Is there a better solution? Sure, we can fix grub-probe. This was already needed for something else (but I forgot what ;-)). Can you propose a CLI by which grub-probe would be told to process devices rather than mount points (in grub-devel)? The hard part here is not the code, but coming up with a consistent interface (the code is rather trivial, since the conversion operation is completely isolated). -- Robert Millan GPLv2 I know my rights; I want my phone call! DRM What use is a phone call… if you are unable to speak? (as seen on /.) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#461442: detection of other OSes in update-grub
Robert Millan schrieb:
Sure, we can fix grub-probe. This was already needed for something else
(but I forgot what ;-)).
Nevertheless it will be very very convenient to have this in grub-probe.
Can you propose a CLI by which grub-probe would be told to process devices
rather than mount points (in grub-devel)? The hard part here is not the code,
but coming up with a consistent interface (the code is rather trivial, since
the conversion operation is completely isolated).
Sure, CCing grub-devel.
We should stay with the '-t drive' option to print the GRUB drive. If no
further option is given, the next item on the command line is expected
to be a path - just like before.
Then, another option, e.g. '-o', should be introduced to specify the
origin of the said item if it is different from path, e.g.:
$ grub-probe
Usage: grub-probe [OPTION]... [ORIGIN]
Probe device information for a given path or, if the -o option is given,
for a given GRUB drive or system device.
-m, --device-map=FILE use FILE as the device map
[default=/boot/grub/device.map]
-t, --target=(fs|drive|device|partmap|abstraction)
print filesystem module, GRUB drive, system
device, partition map module or abstraction module [default=fs]
-o, --origin=(path|drive|device)
expect to read from path, GRUB drive or
system device [default=path]
-h, --helpdisplay this message and exit
-V, --version print version information and exit
-v, --verbose print verbose messages
This way it will be possible to even convert back and forth between GRUB
drives and system devices.
The conversion that is needed to translate os-prober output to GRUB menu
input will read:
GRUB_DEVICE=`grub-probe -t drive -o device ${PARTITION}`.
Hope you like this proposal (of course, origin is subject to change;
it's just the first word that came to my mind that might be fitting)!
Cheers,
Fabian
--
Dipl.-Phys. Fabian Greffrath
Ruhr-Universität Bochum
Lehrstuhl für Energieanlagen und Energieprozesstechnik (LEAT)
Universitätsstr. 150, IB 3/134
D-44780 Bochum
Telefon: +49 (0)234 / 32-26334
Fax: +49 (0)234 / 32-14227
E-Mail: [EMAIL PROTECTED]
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#461442: detection of other OSes in update-grub
Package: grub2 Severity: wishlist Hello, I'd like to know if it is possible (or maybe even planned) to integrate the code into update-grub, that detects other (mostly proprietary) OSes on my harddrive(s) during D-I? I mean, it would be cool to have e.g.WinXP detetcted and added to the GRUB-menu via a simple update-grub. Cheers, Fabian -- Dipl.-Phys. Fabian Greffrath Ruhr-Universität Bochum Lehrstuhl für Energieanlagen und Energieprozesstechnik (LEAT) Universitätsstr. 150, IB 3/134 D-44780 Bochum Telefon: +49 (0)234 / 32-26334 Fax: +49 (0)234 / 32-14227 E-Mail: [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#461442: detection of other OSes in update-grub
Fabian Greffrath [EMAIL PROTECTED] writes: Package: grub2 Severity: wishlist Hello, I'd like to know if it is possible (or maybe even planned) to integrate the code into update-grub, that detects other (mostly proprietary) OSes on my harddrive(s) during D-I? I mean, it would be cool to have e.g.WinXP detetcted and added to the GRUB-menu via a simple update-grub. It's done during d-i using os-probe. -- O T A V I OS A L V A D O R - E-mail: [EMAIL PROTECTED] UIN: 5906116 GNU/Linux User: 239058 GPG ID: 49A5F855 Home Page: http://otavio.ossystems.com.br - Microsoft sells you Windows ... Linux gives you the whole house. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
