Re: [SRU] update for powertop in 6.0.1
Hi, On Sun, 2012-07-01 at 18:11 +0200, Salvatore Bonaccorso wrote: > On Sun, Jul 01, 2012 at 04:41:35PM +0100, Adam D. Barratt wrote: > > On Sun, 2012-07-01 at 14:57 +0200, Salvatore Bonaccorso wrote: > > > As discussed last week shortly on IRC, attached is now the debdiff > > > (only changed changelog entry to reflect NMU) for powertop. > > > > > > (I have added win...@debian.org to recipient lists) > > > > > > Can I upload this? > > > > Please go ahead; thanks. > > Thanks, just uploaded some minutes ago. Flagged for acceptance; thanks. Regards, Adam -- To UNSUBSCRIBE, email to debian-release-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/1341259483.8481.10.ca...@jacala.jungle.funky-badger.org
Re: [SRU] update for powertop in 6.0.1
Hi Adam On Sun, Jul 01, 2012 at 04:41:35PM +0100, Adam D. Barratt wrote: > On Sun, 2012-07-01 at 14:57 +0200, Salvatore Bonaccorso wrote: > > As discussed last week shortly on IRC, attached is now the debdiff > > (only changed changelog entry to reflect NMU) for powertop. > > > > (I have added win...@debian.org to recipient lists) > > > > Can I upload this? > > Please go ahead; thanks. Thanks, just uploaded some minutes ago. Regards, Salvatore signature.asc Description: Digital signature
Re: [SRU] update for powertop in 6.0.1
On Sun, 2012-07-01 at 14:57 +0200, Salvatore Bonaccorso wrote: > As discussed last week shortly on IRC, attached is now the debdiff > (only changed changelog entry to reflect NMU) for powertop. > > (I have added win...@debian.org to recipient lists) > > Can I upload this? Please go ahead; thanks. Regards, Adam -- To UNSUBSCRIBE, email to debian-release-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/1341157295.3721.5.ca...@jacala.jungle.funky-badger.org
Re: [SRU] update for powertop in 6.0.1
Hi Adam
On Tue, Jun 26, 2012 at 07:31:00PM +0200, Salvatore Bonaccorso wrote:
> Hi
>
> On Fri, Apr 06, 2012 at 09:42:54PM +0100, Adam D. Barratt wrote:
> > On Tue, 2011-11-22 at 21:26 +, Adam D. Barratt wrote:
> > > On Wed, 2011-10-19 at 18:18 +0200, Philipp Kern wrote:
> > > > On Thu, Mar 17, 2011 at 01:39:03PM +0100, Patrick Winnertz wrote:
> > > > > > Impressive design. What would happen if it skips half of the
> > > > > > "config"? I
> > > > > > agree that not segfaulting is incredibly better than segfaulting,
> > > > > > but like
> > > > > > this the remaining bunch of lines will be silently ignored?
> > > > > I've rewritten the patch from Mel Gorman, now you'll get a printf
> > > > > output on
> > > > > the console if the config file is too large (and powertop will
> > > > > abort).
> > > > > Please have a look on the updated debdiff attached to this mail.
> > > >
> > > > Bug reports are much easier to track. Please go ahead and sorry for
> > > > the delay.
> > >
> > > Ping?
> >
> > Re-ping.
>
> Is there something one can help for getting this into a stable point
> release?
>
> @Release-Team: I know you will be more busy for the wheezy release
> now!
As discussed last week shortly on IRC, attached is now the debdiff
(only changed changelog entry to reflect NMU) for powertop.
(I have added win...@debian.org to recipient lists)
Can I upload this?
Regards,
Salvatore
diff -u powertop-1.11/debian/changelog powertop-1.11/debian/changelog
--- powertop-1.11/debian/changelog
+++ powertop-1.11/debian/changelog
@@ -1,3 +1,14 @@
+powertop (1.11-1+squeeze1) stable; urgency=low
+
+ * Non-maintainer upload.
+ * Fix segfault on newer kernels with large config files.
+Thanks to Mel Gorman (Closes: #610101)
+NMU upload prepared with debdiff created by Patrick Winnertz
+ and found on debian-release mailing list.
+http://lists.debian.org/debian-release/2011/03/msg00292.html
+
+ -- Salvatore Bonaccorso Sun, 01 Jul 2012 13:12:05 +0200
+
powertop (1.11-1) unstable; urgency=low
* New upstream release
@@ -188 +198,0 @@
-
diff -u powertop-1.11/debian/patches/00list powertop-1.11/debian/patches/00list
--- powertop-1.11/debian/patches/00list
+++ powertop-1.11/debian/patches/00list
@@ -1,0 +2 @@
+increase-configlimit.dpatch
only in patch2:
unchanged:
--- powertop-1.11.orig/debian/patches/increase-configlimit.dpatch
+++ powertop-1.11/debian/patches/increase-configlimit.dpatch
@@ -0,0 +1,59 @@
+#! /bin/sh /usr/share/dpatch/dpatch-run
+## increase-configlimit.patch.dpatch by
+##
+## All lines beginning with `## DP:' are a description of the patch.
+## DP: No description.
+
+@DPATCH@
+diff -urNad '--exclude=CVS' '--exclude=.svn' '--exclude=.git' '--exclude=.arch' '--exclude=.hg' '--exclude=_darcs' '--exclude=.bzr' powertop-1.11~/config.c powertop-1.11/config.c
+--- powertop-1.11~/config.c 2011-03-11 19:06:28.0 +0100
powertop-1.11/config.c 2011-03-17 13:23:23.969122826 +0100
+@@ -33,9 +33,14 @@
+ #include "powertop.h"
+
+ /* static arrays are not nice programming.. but they're easy */
+-static char configlines[5000][100];
++static char configlines[MAXCONFIGLINES][100];
+ static int configcount;
+
++static void cleanup_curses_w_errormsg(void) {
++ endwin();
++ printf(_("Your kernel config exceeds the limit of %i lines, aborting!\n"),MAXCONFIGLINES);
++}
++
+ static void read_kernel_config(void)
+ {
+ FILE *file;
+@@ -49,6 +54,10 @@
+ char line[100];
+ if (fgets(line, 100, file) == NULL)
+ break;
++ if (configcount >= MAXCONFIGLINES) {
++atexit(cleanup_curses_w_errormsg);
++exit(EXIT_FAILURE);
++ }
+ strcpy(configlines[configcount++], line);
+ }
+ pclose(file);
+@@ -77,6 +86,10 @@
+ char line[100];
+ if (fgets(line, 100, file) == NULL)
+ break;
++ if (configcount >= MAXCONFIGLINES) {
++ atexit(cleanup_curses_w_errormsg);
++ exit(EXIT_FAILURE);
++ }
+ strcpy(configlines[configcount++], line);
+ }
+ fclose(file);
+diff -urNad '--exclude=CVS' '--exclude=.svn' '--exclude=.git' '--exclude=.arch' '--exclude=.hg' '--exclude=_darcs' '--exclude=.bzr' powertop-1.11~/powertop.h powertop-1.11/powertop.h
+--- powertop-1.11~/powertop.h 2008-12-30 19:52:54.0 +0100
powertop-1.11/powertop.h 2011-03-17 13:23:35.981182381 +0100
+@@ -88,6 +88,7 @@
+
+ #define _(STRING)gettext(STRING)
+
++#define MAXCONFIGLINES 1
+
+ #define PT_COLOR_DEFAULT1
+ #define PT_COLOR_HEADER_BAR 2
signature.asc
Description: Digital signature
Re: [SRU] update for powertop in 6.0.1
Hi On Fri, Apr 06, 2012 at 09:42:54PM +0100, Adam D. Barratt wrote: > On Tue, 2011-11-22 at 21:26 +, Adam D. Barratt wrote: > > On Wed, 2011-10-19 at 18:18 +0200, Philipp Kern wrote: > > > On Thu, Mar 17, 2011 at 01:39:03PM +0100, Patrick Winnertz wrote: > > > > > Impressive design. What would happen if it skips half of the > > > > > "config"? I > > > > > agree that not segfaulting is incredibly better than segfaulting, but > > > > > like > > > > > this the remaining bunch of lines will be silently ignored? > > > > I've rewritten the patch from Mel Gorman, now you'll get a printf > > > > output on > > > > the console if the config file is too large (and powertop will abort). > > > > Please have a look on the updated debdiff attached to this mail. > > > > > > Bug reports are much easier to track. Please go ahead and sorry for > > > the delay. > > > > Ping? > > Re-ping. Is there something one can help for getting this into a stable point release? @Release-Team: I know you will be more busy for the wheezy release now! Regards, Salvatore signature.asc Description: Digital signature
Re: [SRU] update for powertop in 6.0.1
On Tue, 2011-11-22 at 21:26 +, Adam D. Barratt wrote: > On Wed, 2011-10-19 at 18:18 +0200, Philipp Kern wrote: > > On Thu, Mar 17, 2011 at 01:39:03PM +0100, Patrick Winnertz wrote: > > > > Impressive design. What would happen if it skips half of the "config"? > > > > I > > > > agree that not segfaulting is incredibly better than segfaulting, but > > > > like > > > > this the remaining bunch of lines will be silently ignored? > > > I've rewritten the patch from Mel Gorman, now you'll get a printf output > > > on > > > the console if the config file is too large (and powertop will abort). > > > Please have a look on the updated debdiff attached to this mail. > > > > Bug reports are much easier to track. Please go ahead and sorry for > > the delay. > > Ping? Re-ping. Regards, Adam -- To UNSUBSCRIBE, email to debian-release-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/1333744974.24386.45.ca...@jacala.jungle.funky-badger.org
Re: [SRU] update for powertop in 6.0.1
On Wed, 2011-10-19 at 18:18 +0200, Philipp Kern wrote: > On Thu, Mar 17, 2011 at 01:39:03PM +0100, Patrick Winnertz wrote: > > > Impressive design. What would happen if it skips half of the "config"? I > > > agree that not segfaulting is incredibly better than segfaulting, but like > > > this the remaining bunch of lines will be silently ignored? > > I've rewritten the patch from Mel Gorman, now you'll get a printf output on > > the console if the config file is too large (and powertop will abort). > > Please have a look on the updated debdiff attached to this mail. > > Bug reports are much easier to track. Please go ahead and sorry for > the delay. Ping? Regards, Adam -- To UNSUBSCRIBE, email to debian-release-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/1321997218.16755.11.ca...@hathi.jungle.funky-badger.org
Re: [SRU] update for powertop in 6.0.1
On Thu, Mar 17, 2011 at 01:39:03PM +0100, Patrick Winnertz wrote: > > Impressive design. What would happen if it skips half of the "config"? I > > agree that not segfaulting is incredibly better than segfaulting, but like > > this the remaining bunch of lines will be silently ignored? > I've rewritten the patch from Mel Gorman, now you'll get a printf output on > the console if the config file is too large (and powertop will abort). > Please have a look on the updated debdiff attached to this mail. Bug reports are much easier to track. Please go ahead and sorry for the delay. Kind regards Philipp Kern signature.asc Description: Digital signature
Re: [SRU] update for powertop in 6.0.1
Hey,
> Impressive design. What would happen if it skips half of the "config"? I
> agree that not segfaulting is incredibly better than segfaulting, but like
> this the remaining bunch of lines will be silently ignored?
I've rewritten the patch from Mel Gorman, now you'll get a printf output on
the console if the config file is too large (and powertop will abort).
Please have a look on the updated debdiff attached to this mail.
Greetings
Winnie
diff -u powertop-1.11/debian/changelog powertop-1.11/debian/changelog
--- powertop-1.11/debian/changelog
+++ powertop-1.11/debian/changelog
@@ -1,3 +1,10 @@
+powertop (1.11-1+squeeze1) stable; urgency=low
+
+ * Fix segfault on newer kernels with large config files.
+Thanks to Mel Gorman (Closes: #610101)
+
+ -- Patrick Winnertz Fri, 11 Mar 2011 11:22:33 +0100
+
powertop (1.11-1) unstable; urgency=low
* New upstream release
diff -u powertop-1.11/debian/patches/00list powertop-1.11/debian/patches/00list
--- powertop-1.11/debian/patches/00list
+++ powertop-1.11/debian/patches/00list
@@ -1,0 +2 @@
+increase-configlimit.dpatch
only in patch2:
unchanged:
--- powertop-1.11.orig/debian/patches/increase-configlimit.dpatch
+++ powertop-1.11/debian/patches/increase-configlimit.dpatch
@@ -0,0 +1,59 @@
+#! /bin/sh /usr/share/dpatch/dpatch-run
+## increase-configlimit.patch.dpatch by
+##
+## All lines beginning with `## DP:' are a description of the patch.
+## DP: No description.
+
+@DPATCH@
+diff -urNad '--exclude=CVS' '--exclude=.svn' '--exclude=.git' '--exclude=.arch' '--exclude=.hg' '--exclude=_darcs' '--exclude=.bzr' powertop-1.11~/config.c powertop-1.11/config.c
+--- powertop-1.11~/config.c 2011-03-11 19:06:28.0 +0100
powertop-1.11/config.c 2011-03-17 13:23:23.969122826 +0100
+@@ -33,9 +33,14 @@
+ #include "powertop.h"
+
+ /* static arrays are not nice programming.. but they're easy */
+-static char configlines[5000][100];
++static char configlines[MAXCONFIGLINES][100];
+ static int configcount;
+
++static void cleanup_curses_w_errormsg(void) {
++ endwin();
++ printf(_("Your kernel config exceeds the limit of %i lines, aborting!\n"),MAXCONFIGLINES);
++}
++
+ static void read_kernel_config(void)
+ {
+ FILE *file;
+@@ -49,6 +54,10 @@
+ char line[100];
+ if (fgets(line, 100, file) == NULL)
+ break;
++ if (configcount >= MAXCONFIGLINES) {
++atexit(cleanup_curses_w_errormsg);
++exit(EXIT_FAILURE);
++ }
+ strcpy(configlines[configcount++], line);
+ }
+ pclose(file);
+@@ -77,6 +86,10 @@
+ char line[100];
+ if (fgets(line, 100, file) == NULL)
+ break;
++ if (configcount >= MAXCONFIGLINES) {
++ atexit(cleanup_curses_w_errormsg);
++ exit(EXIT_FAILURE);
++ }
+ strcpy(configlines[configcount++], line);
+ }
+ fclose(file);
+diff -urNad '--exclude=CVS' '--exclude=.svn' '--exclude=.git' '--exclude=.arch' '--exclude=.hg' '--exclude=_darcs' '--exclude=.bzr' powertop-1.11~/powertop.h powertop-1.11/powertop.h
+--- powertop-1.11~/powertop.h 2008-12-30 19:52:54.0 +0100
powertop-1.11/powertop.h 2011-03-17 13:23:35.981182381 +0100
+@@ -88,6 +88,7 @@
+
+ #define _(STRING)gettext(STRING)
+
++#define MAXCONFIGLINES 1
+
+ #define PT_COLOR_DEFAULT1
+ #define PT_COLOR_HEADER_BAR 2
Re: [SRU] update for powertop in 6.0.1
On Fri, Mar 11, 2011 at 11:59:51AM +0100, Patrick Winnertz wrote: > Hey, > > I would like to close one important bug in powertop, which affects the usage > on newer kernels (e.g. 2.6.37), as powertop will segfault on such kernels > due to too small buffer (see #610101). > > Please see my attached patch for powertop 1.11-1+squeeze1 Impressive design. What would happen if it skips half of the "config"? I agree that not segfaulting is incredibly better than segfaulting, but like this the remaining bunch of lines will be silently ignored? Kind regards Philipp Kern signature.asc Description: Digital signature
