Re: Problem with kernel patching
On Sun, 03 Jun 2012 21:35:12 +0100, dobrima...@yahoo.pl wrote: (please, no html formatted posts, thanks :-) ) Hello, I have a problem with patching kernel 2.6-2.6.32. I'm trying to add PVUSB [1] support in Xen. According to [1] I should add a patch [2] to kernel. I was trying to do it with steps described in [3]. I had to make some changes in the patch file because in the meantime some other changes appeared. After my changes the patch file looks like [4]. When I try to execute make -f debian/rules source-all it fails at applying my patch stage. The errors appearing until I add a/ i b/ prefixes at lines 2, 3, 2045, 2046, 2060, 2061,4390 and 4391. I'm not sure what is the purpose of using it there. (...) Mmm... it's a rather large patch. I would ask for feedback at the pkg- xen-devel mailing list: http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-xen-devel Greetings, -- Camaleón -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/jqii8t$d82$1...@dough.gmane.org
Problem with kernel patching
Hello, I have a problem with patching kernel 2.6-2.6.32. I'm trying to add PVUSB [1] support in Xen. According to [1] I should add a patch [2] to kernel. I was trying to do it with steps described in [3]. I had to make some changes in the patch file because in the meantime some other changes appeared. After my changes the patch file looks like [4]. When I try to execute make -f debian/rules source-all it fails at applying my patch stage. The errors appearing until I add a/ i b/ prefixes at lines 2, 3, 2045, 2046, 2060, 2061,4390 and 4391. I'm not sure what is the purpose of using it there. Now it goes without any errors, but the files which should be modified, remain intact, and that's my problem actually. I can't figure out why it happens. I will be grateful for any help. [1] http://wiki.xen.org/wiki/Xen_USB_Passthrough [2] http://members.iinet.net.au/~nathanael/pvusb.diff [3] http://wiki.debian.org/HowToRebuildAnOfficialDebianKernelPackage [4] http://wklej.org/id/759234/
Re: kernel patching
On Sat, Oct 15, 2005 at 09:42:02AM +0530, L.V.Gandhi wrote: I already had linux kernel linux-2.6.13.2.tar.bz2. I untaared in /usr/src. and sym link to linux. I downloaded patch for 13.3 and 13.3 gz files and gunzipped. I did dry run patch. I got error. Look at the patch file - you should see a line which modifies EXTRAVERSION right near the top. For example: $ bzcat patch-2.6.13.4.bz2 |head diff --git a/Makefile b/Makefile --- a/Makefile +++ b/Makefile @@ -1,7 +1,7 @@ VERSION = 2 PATCHLEVEL = 6 SUBLEVEL = 13 -EXTRAVERSION = +EXTRAVERSION = .4 NAME=Woozy Numbat The patch files are from the base release, not the last EXTRAVERSION - so you need linux-2.6.13 to patch rather than linux-2.6.13.2. Sorry! -- Jon Dowland http://jon.dowland.name/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: kernel patching
On 10/15/05, Jon Dowland [EMAIL PROTECTED] wrote: The patch files are from the base release, not the last EXTRAVERSION - so you need linux-2.6.13 to patch rather than linux-2.6.13.2. Sorry! Thanks for the info. -- L.V.Gandhi http://lvgandhi.tripod.com/ linux user No.205042
kernel patching
I already had linux kernel linux-2.6.13.2.tar.bz2. I untaared in /usr/src. and sym link to linux. I downloaded patch for 13.3 and 13.3 gz files and gunzipped. I did dry run patch. I got error. lvghomepc:/usr/src# ln -s linux-2.6.13.2 linux lvghomepc:/usr/src# cd linux lvghomepc:/usr/src/linux# cp /mnt/wing/Linux/patch-2.6.13.* . lvghomepc:/usr/src/linux# patch -p1 --dry-run patch-2.6.13.3 patching file Makefile Hunk #1 FAILED at 1. 1 out of 1 hunk FAILED -- saving rejects to file Makefile.rej patching file arch/i386/pci/common.c Reversed (or previously applied) patch detected! Assume -R? [n] I have stooped here. Why there is failure even for vanilla kernel? What to do now? -- L.V.Gandhi http://lvgandhi.tripod.com/ linux user No.205042
Re: Kernel patching + compilation for Cobalt RAQ 500....
Hi Johnathan, I'm confused, are You trying to install a debain patch on a vanilla kernel or the other way round? I cannot find a cobalt patch with my woody package list, so it's not in the current stable distribution? Anyway, debian kernel packages are usually somehow patched and You might run into trouble, once You use a vanilla kernel source against a debian patch or the other way round. Switch both to debian packages (recommended) or the other way round. I'm using the testing distribution to compile the kernel upon. There is no Debian kernel source package for 2.4.23 so it's a vanilla unpatched kernel source, but the patch package only supports kernel 2.4.23?! I think I'll have to contact the patch maintainer... -Cheers Max. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Kernel patching + compilation for Cobalt RAQ 500....
Hi Folks, Has anyone recently built a kernel for the i386 based Cobalt RAQ500? I'm trying to get debian up and running on two of these units, and I really want to use a 'proper' debian package to install the kernel from. I've installed the 2.4.23 kernel source (no package available for this) and the kernel-patch-2.4-cobalt package. I tried running /usr/src/kernel-patches/all/apply whilst in /usr/src/linux-2.4.23 but that fails, so I tried manually applying /usr/src/kernel-patches/all/cobalt/cobalt-2.4.23.patch.gz after uncompressing it and it seems to apply fine apart from failing to patch one of the bluetooth drivers. The real problem occurs when I try to compile the kernel... ruler.c:50: error: syntax error before '*' token ruler.c:50: warning: type defaults to `int' in declaration of `busprocs' ruler.c:50: warning: data definition has no type or storage class ruler.c: In function `do_busproc': ruler.c:128: error: called object is not a function ruler.c: In function `cobalt_ruler_register': ruler.c:309: warning: assignment from incompatible pointer type ruler.c:310: warning: assignment from incompatible pointer type ruler.c: In function `cobalt_ruler_unregister': ruler.c:347: warning: assignment from incompatible pointer type then it dies :( Has anyone managed to compile with this patch? should I raise a bug report? -Cheers Max. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Kernel patching + compilation for Cobalt RAQ 500....
Hallo, I'm trying to get debian up and running on two of these units, and I really want to use a 'proper' debian package to install the kernel from. I've installed the 2.4.23 kernel source (no package available for this) and the kernel-patch-2.4-cobalt package. I'm confused, are You trying to install a debain patch on a vanilla kernel or the other way round? I cannot find a cobalt patch with my woody package list, so it's not in the current stable distribution? Anyway, debian kernel packages are usually somehow patched and You might run into trouble, once You use a vanilla kernel source against a debian patch or the other way round. Switch both to debian packages (recommended) or the other way round. [...] seems to apply fine apart from failing to patch one of the bluetooth drivers. This is usually either a patch against the wrong kernel version (there were some hefty changes during 2.4 developement) or patch against a patched kernel source, the second patch was not expecting to find. Best regards, js -- Down with categorical imperative! -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Kernel Patching: Updating the debian/changelog
My notes are not perfect, but I believe on one machine that's running 2.4.23 the kernel was built from kernel.org sources but make-kpkg was used to build the kernel. I just downloaded and patched the 2.4.23 source tree with the 2.4.24 patch. Then I ran fakeroot make-kpkg --revision=custom.1.0 kernel_image and the build went ok except for: The changelog says we are creating 2.4.23-xfs-3ware, but I thought the version is 2.4.24-xfs-3ware I'm a bit confused about how the debian/changelog got there -- does make-kpkg add it when building form kernel.org sources? I have notes of hand-editing the changelog before, but I can't see to get the format correct. I know there's a tool to update the changelog, but can't remember what that is. Anyway, the question is: how do I patch the kernel and correctly update the debian/changelog so make-kpkg doesn't puke. -- Bill Moseley [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Kernel Patching: Updating the debian/changelog
On Mon, Jan 05, 2004 at 11:09:09AM -0800, Bill Moseley [EMAIL PROTECTED] wrote: I have notes of hand-editing the changelog before, but I can't see to get the format correct. I know there's a tool to update the changelog, but can't remember what that is. You should use 'dch -i' and add a changelog for your new kernel. Anyway, the question is: how do I patch the kernel and correctly update the debian/changelog so make-kpkg doesn't puke. I think the above will do if you upgrade the version in the changelog to 2.4.24-whatever. Cheers, GCS -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Kernel Patching: Updating the debian/changelog
Hello Bill Moseley ([EMAIL PROTECTED]) wrote: My notes are not perfect, but I believe on one machine that's running 2.4.23 the kernel was built from kernel.org sources but make-kpkg was used to build the kernel. I just downloaded and patched the 2.4.23 source tree with the 2.4.24 patch. Then I ran fakeroot make-kpkg --revision=custom.1.0 kernel_image and the build went ok except for: The changelog says we are creating 2.4.23-xfs-3ware, but I thought the version is 2.4.24-xfs-3ware I'm a bit confused about how the debian/changelog got there -- does make-kpkg add it when building form kernel.org sources? I think so. Run make-kpgk clean and try again. best regards Andreas Janssen -- Andreas Janssen [EMAIL PROTECTED] PGP-Key-ID: 0xDC801674 Registered Linux User #267976 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Kernel patching
Ronald Castillo wrote: Greetings... When I apply a patch to my kernel (from 2.2.12 to 2.2.13), do I need to do any other step after I do the gzip -cd patchxx.gz | patch -p0? Thanks.. The sources are now patched. (assuming you were in the /usr/src directory when you unzipped and patched.) I assume you are using the docs in /usr/src/linux/README. You now need to rebuild the kernel. hth dyer
Kernel patching
Greetings... When I apply a patch to my kernel (from 2.2.12 to 2.2.13), do I need to do any other step after I do the gzip -cd patchxx.gz | patch -p0? Thanks..
Kernel patching
Hi all. (Sorry, forgot to include the meaningful subject: message below) I've just applied 2.0.35, 2.0.36 and 2.0.37 patches to my 2.0.34 kernel. Now as I compile the kernel, I get these errors: setup.c: In function `Cx86model': setup.c:286: `Cx86_mult' undeclared (first use this function) setup.c:286: (Each undeclared identifier is reported only once setup.c:286: for each function it appears in.) setup.c:306: warning: passing arg 1 of `sprintf' makes pointer from integer without a cast setup.c:312: warning: passing arg 1 of `sprintf' makes pointer from integer without a cast setup.c:318: warning: passing arg 1 of `sprintf' makes pointer from integer without a cast setup.c:328: warning: passing arg 1 of `sprintf' makes pointer from integer without a cast setup.c:331: warning: passing arg 1 of `sprintf' makes pointer from integer without a cast setup.c:339: `ext_cpuid' undeclared (first use this function) setup.c:341: warning: passing arg 1 of `sprintf' makes pointer from integer without a cast setup.c:349: warning: passing arg 1 of `sprintf' makes pointer from integer without a cast setup.c: In function `AMDmodel': setup.c:381: `ext_cpuid' undeclared (first use this function) setup.c: In function `get_cpu_mkt_name': setup.c:439: `x86_ext_capability' undeclared (first use this function) setup.c: In function `getmodel': setup.c:458: `ext_cpuid' undeclared (first use this function) setup.c: In function `get_cpuinfo': setup.c:517: `x86_clkmult' undeclared (first use this function) setup.c:517: `Cx86_mult' undeclared (first use this function) setup.c:517: warning: format argument is not a pointer (arg 3) setup.c:551: `x86_ext_capability' undeclared (first use this function) setup.c: At top level: setup.c:53: warning: `Cx86_type' defined but not used make[1]: *** [setup.o] Error 1 make[1]: Leaving directory `/usr/src/kernel-source-2.0.34/arch/i386/kernel' make: *** [linuxsubdirs] Error 2 Any idea what I (the patch?) has done wrong and how to go about it? TIA, Andrew --- Andrei S. Ivanov [EMAIL PROTECTED] [EMAIL PROTECTED] UIN 12402354 http://members.tripod.com/AnSIv --Little things for Linux. http://www.missouri.edu/~c680789 --Computer languages of the world My work in progress. ---
HELP: Kernel patching failed - 2.0.34 to 2.0.36
Hello! When I tried to patch my kernel I encounter a lot of error messages, I'm not sure if I can continue, hoping somebody with more experience can give me some suggestion. I've collected the error messages and attatched it to this mail. All what I did is (as root): cp /download/patch-2.0.36.gz /usr/src cd /usr/src/linux make clean cd .. zcat patch-2.0.36.gz|patch -p0 -s --dry-run patch.out TIA!! BTW: Do you think that it's time to move to kernel 2.2.1? patch.out Description: Binary data
Re: Journalling: Upgrades, Kernel patching, ...
On Wed, 9 Apr 1997, Jim Pick wrote: Great! I would like to see http://localhost/cgi-bin/dwww?type=dirlocation=/usr/doc done as a multi-column table; it would look better and be easier to search with the eyes. You could add tags conditional on $ENV{'USER_AGENT'} being a table aware browser, and use pre with added spaces and a hard formatted table for non-tableing ones. It will take some codeing for sure. That's a good idea. I hadn't thought of doing that. I'll add it to the to do list. :-) Don't parse UserAgent! Lynx can ignore the tables... Like it does in www.debian.org/Packages/unstable/ -- Nicolás Lichtmaier.- [EMAIL PROTECTED]
Journalling: Upgrades, Kernel patching, ...
I just upgraded BitterSweet to Debian pre1.3, over the modem. It took all night at 33.6, to ftp the files I selected. Everything I'd carefully chosen, a 1 hour job, with `dselect` transfered without a single glitch, from URL:ftp://ftp.debian.org/debian. I selected files from bo, contrib, and non-free. (It said 'unstable', later... I thought I typed 'bo'...? Hm.) There was only about 4 installation errors, I think... (They scroll away too fast, and there is no log kept, AFACT.) Wow! Pretty good, considering I got over a hundred new and upgraded packages! Much easier than compilityouseff. There have been quite a number of improvements since the version I first installed, which was 1.1.n from the InfoMagic set, from which I also demo'd RedHat 3.0.3. In particular, I like the new `bug` program, which lets you easily file a bug report against a software package. You type `bug packagename` at a shell prompt, and edit the report template in your $VISUAL editor (I prefer `gnuclient`.), it gets mailed off to the bug tracking system, which autoresponds with your bug number. I think that `bug` will help to improve the quality of the Debian GNU/Linux distribution and the software that comprises it, since by making it simpler and more efficient to report bugs, people are more likely to, and the maintainers will become aware of problems sooner. I was very impressed by the expediency with which the recently announced `amd` (Auto-Mounter Daemon) security bug was fixed! There have been other cases where We have had bug fixes before the other distributions have had them. Some RedHat people from PLUG (Portland Linux User's Group) seemed very amazed when I talked about running the Corel WordPerfect Java applets inside Linux NetXcrape. Mine always crashes! I have to keep Java disabled!. To have a one or zero floppy CDROM install sounds like a worthy goal! Six floppies to do a base install isn't too awful bad though, and my test the other day showed me that it provides everything needed to ftp, nfs, or plip the rest of the system, if a CDROM, or Zip disk isn't available. There should be a dialog for configuring pppd though; that was lacking. The 56Kbps modems are going to make the ftp installation and upgrade option very viable. (Internet Arena has them now; I'm waiting for a response about an upgrade to my modem from Cardinal. There must be thousands of others like me at both ends of the link.). It is pretty amazing to have things being installed into the machine, while it is up and running, without having to do much of anything to make it happen. (I ran with slackware for 12 months.) To test this a little, I goofed around on another console and in X windows with logging in and out, telnetting to Internet Arena, and running commands (`mc`) that had just been unpacked by `dselect`. I ran `top` for a bit to watch what `dselect` was doing, and then used `gitps -wauxf` to get momentary snapshots of it. Now THAT's a great tool! I like how I can flick over to another console and use XEmacs' ediff to compare two configuration files, merging my local edits into the new package's .dpkg-inst version. I was able to add the transnames fixups back to BitterSweet's /etc/init.d/boot like that, while other packages were still being installed by `dselect` on the foist console. I think that to perform an upgrade on an ISP server would be very doable. I wouldn't even hesitate to upgrade `sendmail`, `qpopper`, `cron`, `innd`, `apache`, `RADIUS`, `majordomo`, `perl`, `libc`, or any other mission critical package, without even changing runlevel, or forcing people to log out first. (I guess I'd tell them, in case there's a goof, so they aren't running anything critical when it happens.) Dpkg works very well! I glanced over the new 'dwww' documantation interface, and its really looking good! I hadn't had the menu package installed previously. The documents under that hierarchy look really nice! Is that the debiandoc DTD that the doc people have been talking about? Great! I would like to see http://localhost/cgi-bin/dwww?type=dirlocation=/usr/doc done as a multi-column table; it would look better and be easier to search with the eyes. You could add tags conditional on $ENV{'USER_AGENT'} being a table aware browser, and use pre with added spaces and a hard formatted table for non-tableing ones. It will take some codeing for sure. `dpkg-repack` sounds very promising. I will try it when I get my second box up and running. I want to find out how easy that make it to klone web servers and X-swerver confounderations. (I will also do further experiments with the transnames kernel patch, while I simutaniously teach myself scheme.) Everything restarts again without too much trouble, right after configuration, incluging X-Windows. The irritating minor (really...) bugs are: * that my locally compiled 3-d XDM login (that I found on the net from rastasia), got overwritten, and the installed one does not
Re: Journalling: Upgrades, Kernel patching, ...
I glanced over the new 'dwww' documantation interface, and its really looking good! Cool. I hadn't had the menu package installed previously. The documents under that hierarchy look really nice! Is that the debiandoc DTD that the doc people have been talking about? I'm not sure. I think so. Great! I would like to see http://localhost/cgi-bin/dwww?type=dirlocation=/usr/doc done as a multi-column table; it would look better and be easier to search with the eyes. You could add tags conditional on $ENV{'USER_AGENT'} being a table aware browser, and use pre with added spaces and a hard formatted table for non-tableing ones. It will take some codeing for sure. That's a good idea. I hadn't thought of doing that. I'll add it to the to do list. :-) Cheers, - Jim pgpWfsKRW3p05.pgp Description: PGP signature
RE: Journalling: Upgrades, Kernel patching, ...
-BEGIN PGP SIGNED MESSAGE- RE: [EMAIL PROTECTED] Is this a beta testers report??? If so is this the place to send these reports to Debian? I'm a bit confused here since most people subscribed to this list have been using bo, or parts of it for some time now. I've been on this list since around the first of the year and haven't seen this kind of Report before. Was this an accident or is this the list to make such reports? Honest question, not a flame. Because it seems an unlikely place to inform people of things we've been talking about in here for months now. If this be the place for reports of this nature let me know so I know where to submit mine. On 09-Apr-97 Karl M. Hegbloom wrote: I just upgraded BitterSweet to Debian pre1.3, over the modem. It took all night at 33.6, to ftp the files I selected. Everything I'd carefully chosen, a 1 hour job, with `dselect` transfered without a single glitch, from URL:ftp://ftp.debian.org/debian. I selected files from bo, contrib, and non-free. (It said 'unstable', later... I thought I typed 'bo'...? Hm.) There was only about 4 installation errors, I think... (They scroll away too fast, and there is no log kept, AFACT.) Wow! Pretty good, considering I got over a hundred new and upgraded packages! Much easier than compilityouseff. snip The lp.c fix was easy because 2.0.30 added only one statement to the top of a function that had also been altered by the parport patch. CVS didn't know what to do, but when I looked, it was obvious how to make it look. I'm sure of that one. I will try `kernel-package` when I'm done. Have a good one. - -- Rick Jones E-Mail: Rick [EMAIL PROTECTED] Date: 09-Apr-97 Time: 22:59:31 - -- -BEGIN PGP SIGNATURE- Version: 2.6.2 iQCVAwUBM0xXkwi+Ph+i3TgpAQGe8QP+J6PTHDZa99T8LLI+tDlJX0zl4XhBsf2l TcjgW3rCIFhnwUsdSgLkxccq0I+J19+yFUJ54FyI+eIp/JSLmeBGeZIN4uHXpvR7 KI3+sc2vp3Xc77v6VK1VL627LJNG4ZN5InjEyNLKzecQ3hY4bhC5lSxWz+QxUDY8 5ZPZUabGWGM= =niin -END PGP SIGNATURE- Karl M. Hegbloom [EMAIL PROTECTED] http://www.inetarena.com/~karlheg Portland, OR USA Debian GNU 1.2 Linux 2.0.29t You tell me and we'll both know.
RE: Journalling: Upgrades, Kernel patching, ...
Ahem... could all further users please not reply-to-all in further replies to Karl's message? He put all the staff members at Internet Arena in the CC field ( although only a few are members of PLUG ) and we've been recieving replies we don't always need. I think only three of us are subscribed to plug, and one to debian-user. Hopefully nobody over here sent this message already and I missed it... I'd feel REALLY stupid then :-D As much as I am interested in Karl's finds, it can tend to be excessive when there's so much e-mail to answer in limited time, with schoolwork and all. Hope I'm not angering anyone by asking this. My thanks in advance, Ryan J. Myers ( Webmaster ) [EMAIL PROTECTED] Portland, Oregon ( US ) http://www.inetarena.com/~rmyers/ Nolite tes bastardes carborundum. - Margaret Atwood C REALITY.SYS is corrupt, reboot universe? (Y|N) - Anonymous --- The contents of this email are my own opinions ( read: insane rants ) and do not represent the positions of my employers, their subsidiaries, or the little green men that have been following me around all day.