Bug#829417: dosemu: *** stack smashing detected ***: /usr/bin/xdosemu terminated
Hello, At the top of debian/rules addding the following: DEB_BUILD_MAINT_OPTIONS=hardening=-all export DEB_BUILD_MAINT_OPTIONS stops crashes from occurring by disabling the protection mechanisms provided by GCC. This does not truly fix the issue because dosemu itself should be modified to handle cases where GCC's protection mechanisms are used.
Bug#829417: dosemu: *** stack smashing detected ***: /usr/bin/xdosemu terminated
On x86_64 running Stretch, I also get this when trying to run unzip32 from DJGPP which is at the following URL: http://www.delorie.com/pub/djgpp/current/unzip32.exe Changing the CPU type to "full" allows the command to run however it quickly crashes with: ERROR: Fault in dosemu code, in_dpmi=1 ERROR: cpu exception in dosemu code outside of DPMI client! trapno: 0x0e errorcode: 0x0004 cr2: 0xa028601d eip: 0x564da024d887 esp: 0x7ffe84229960 eflags: 0x00010246 cs: 0x0033 ds: 0x es: 0x ss: 0x002b ERROR: Please report the contents of ~/.dosemu/boot.log at http://sourceforge.net/tracker/?atid=457447_id=49784=browse It would be even more helpful if would recompile DOSEMU and reproduce this bug with "debug on" in compiletime-settings. Page fault: read instruction to linear address: 0xa028601d CPU was in user mode Exception was caused by non-available page I have additionally attached the boot.log file. CONF: config variable parser_version_3 set CONF: config variable c_system set CONF: Parsing built-in dosemu.conf file. CONF: config variable version_3_style_used set CONF: Parsing built-in global.conf file. CONF: config variable version_3_style_used unset CONF: config variable version_3_style_used set CONF: opened include file /etc/dosemu/dosemu.conf CONF: closed include file /etc/dosemu/dosemu.conf CONF: mapping driver = 'auto' debug flags: -a+cw CONF: Disabling use of pentium timer CONF: dosbanner on CONF: timer freq=18, update=54925 CONF: CPU set to 586 CONF: JIT CPUEMU set to 4 for 586 CONF: 8192k bytes EMS memory CONF: EMS-frame = 0xe400 CONF: DPMI-Server on (0x5000) CONF: DPMI base addr = 0x CONF: PM DOS API Translator on CONF: No DJGPP NULL deref checks: off CONF: 8192k bytes XMS memory CONF: dosemu not running on console CONF: time mode = 'bios' SER: directory /var/lock namestub LCK.. binary No MOUSE: no device specified, type 0 using internaldriver: yes, emulate3buttons: no baudrate: 0 CONF: Keyboard-layout keyb-user CONF: Warning: floppy /dev/fd0 not accessible, disabled CONF: fastfloppy = 1 CONF: IPX support off CONF(LPT0) f: (null) c: lpr -l t: 20 port: 0 CONF(LPT1) f: (null) c: lpr -l -P lpt2 t: 20 port: 0 CONF: not allowing speaker port access CONF: Packet Driver enabled. device: /home/steven/.dosemu/drives/c type 4 h: -1 s: -1 t: -1 drive C: device: /home/steven/.dosemu/drives/d type 4 h: -1 s: -1 t: -1 drive D: CONF: cdrom MSCD0001 on /dev/cdrom CONF: config variable c_system unset Linux kernel 4.9.0; CPU speed is 213400 Hz CPU-EMU speed is 2134 MHz CONF: mostly running as USER: uid=1000 (cached 1000) gid=1000 (cached 1000) DBG_FD already set DOSEMU-1.4.0.8 is coming up on Linux version 4.9.0-3-amd64 #1 SMP Debian 4.9.30-1 (2017-06-04) x86_64 Compiled with GCC version 6.3.0 -m64 WARN: vm86plus service not available in your kernel WARN: using CPU emulation for vm86() CONF: reserving 640Kb at 0x0 for 'd' (Base DOS memory (first 640K)) CONF: reserving 48Kb at 0xF4000 for 'r' (Dosemu reserved area) CONF: reserving 128Kb at 0xA for 'v' (Video memory) PKT: Cannot open raw sockets: Operation not permitted WARN: using non-zero memory base address 0x10f000. WARN: You can use the better-tested zero based setup using WARN: sysctl -w vm.mmap_min_addr=0 WARN: as root, or by changing the vm.mmap_min_addr setting in WARN: /etc/sysctl.conf or a file in /etc/sysctl.d/ to 0. CONF: reserving 8256Kb at 0x10 for 'x' (Extended memory (HMA+XMS)) Registering HWRAM, type=e base=0x4197c000 size=0x40 CONF: reserving 4096Kb at 0x4197C000 for 'e' (VGAEMU LFB) CONF: reserving 12Kb at 0xC for 'V' (VGAEMU Video BIOS) SERIAL $Id$ CONF: detected layout is "us" CONF: detected alternate layout: us CONF: reserving 16Kb at 0xE4000 for 'E' (EMS page frame) CONF: reserving 16Kb at 0xE8000 for 'E' (EMS page frame) CONF: reserving 16Kb at 0xEC000 for 'E' (EMS page frame) CONF: reserving 16Kb at 0xF for 'E' (EMS page frame) CONF: reserving 132Kb at 0xC3000 for 'U' (Upper Memory Block (UMB, XMS 3.0)) TIME: using 9154 usec for updating ALRM timer === ENTER CPU-EMU === * Fault out of DOSEMU code, cs:eip=33:564da028bd16, cr2=11cf, fault_cnt=1 * Fault out of DOSEMU code, cs:eip=33:564da02884ec, cr2=11db, fault_cnt=1 * Fault out of DOSEMU code, cs:eip=33:564da02884ec, cr2=11db, fault_cnt=1 * Fault out of DOSEMU code, cs:eip=33:564da02884ec, cr2=11db, fault_cnt=1 * Fault out of DOSEMU code, cs:eip=33:564da02884ec, cr2=11db, fault_cnt=1 * Fault out of DOSEMU code, cs:eip=33:564da028bd16, cr2=11cf, fault_cnt=1 ERROR: Fault in dosemu code, in_dpmi=1 ERROR: cpu exception in dosemu code outside of DPMI client! trapno: 0x0e errorcode: 0x0004 cr2: 0xa028601d eip: 0x564da024d887 esp: 0x7ffe84229960 eflags: 0x00010246 cs: 0x0033 ds: 0x es: 0x ss: 0x002b ERROR: Please report the contents of ~/.dosemu/boot.log at http://sourceforge.net/tracker/?atid=457447_id=49784=browse It would be even more helpful if would recompile DOSEMU and
Bug#844617: g++-mingw-w64: C++ linking fails with cannot find -latomic.
On Thu, 17 Nov 2016 21:04:49 +0100 Stephen Kitt <sk...@debian.org> wrote: > Hi Steven, > > On Thu, 17 Nov 2016 11:20:59 -0500, Steven Gawroriski > <ste...@multiphasicapps.net> wrote: > > During the linking stage of even the most simplest source code > > (just a main function), the C++ linking fails because it cannot > > find `-latomic`. This effectively makes the C++ compiler useless > > because it cannot link any binaries. > > Could you try on another architecture? I can link C++ programs on all > the architectures I have easy access to (amd64, i386, armhf, arm64) > and the resulting programs run correctly on Windows. > > Thanks, > > Stephen Hello, I do not have another architecture running sid. However, it appears that the compiler is configured with `--disable-libatomic`, if this makes a difference. Using built-in specs. COLLECT_GCC=i686-w64-mingw32-g++ COLLECT_LTO_WRAPPER=/usr/lib/gcc/i686-w64-mingw32/6.1-win32/lto-wrapper Target: i686-w64-mingw32 Configured with: ../../src/configure --build=powerpc-linux-gnu --prefix=/usr --includedir='/usr/include' --mandir='/usr/share/man' --infodir='/usr/share/info' --sysconfdir=/etc --localstatedir=/var --disable-silent-rules --libdir='/usr/lib/powerpc-linux-gnu' --libexecdir='/usr/lib/powerpc-linux-gnu' --disable-maintainer-mode --disable-dependency-tracking --prefix=/usr --enable-shared --enable-static --disable-multilib --with-system-zlib --libexecdir=/usr/lib --without-included-gettext --libdir=/usr/lib --enable-libstdcxx-time=yes --with-tune=generic --enable-version-specific-runtime-libs --enable-fully-dynamic-string --enable-libgomp --enable-languages=c,c++,fortran,objc,obj-c++ --enable-lto --with-plugin-ld --enable-threads=win32 --program-suffix=-win32 --program-prefix=i686-w64-mingw32- --target=i686-w64-mingw32 --with-as=/usr/bin/i686-w64-mingw32-as --with-ld=/usr/bin/i686-w64-mingw32-ld --disable-libatomic Thread model: win32 gcc version 6.1.1 20160815 (GCC) Thanks. pgpa7pE3kYErU.pgp Description: OpenPGP digital signature
Bug#844617: g++-mingw-w64: C++ linking fails with cannot find -latomic.
Package: g++-mingw-w64 Version: 6.1.1-12+19.1 Severity: grave Justification: renders package unusable Dear Maintainer, During the linking stage of even the most simplest source code (just a main function), the C++ linking fails because it cannot find `-latomic`. This effectively makes the C++ compiler useless because it cannot link any binaries. -- System Information: Debian Release: stretch/sid APT prefers unstable APT policy: (500, 'unstable') Architecture: powerpc (ppc) Kernel: Linux 4.4.27.161027.1 Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/dash Init: systemd (via /run/systemd/system) Versions of packages g++-mingw-w64 depends on: ii g++-mingw-w64-i6866.1.1-12+19.1 ii g++-mingw-w64-x86-64 6.1.1-12+19.1 g++-mingw-w64 recommends no packages. g++-mingw-w64 suggests no packages. -- no debconf information
Bug#841853: inkscape: Mouse cursor is a white outline on PowerPC
On Mon, 7 Nov 2016 16:28:13 + Mattia Rizzolowrote: > ouch, that must be awful! It seems to be fine now. Last Saturday (November 5th) downloading the source package completed rather quickly (10 seconds perhaps?). > Great! > I also noticed that they committed the patch already in both trunk and > 0.92.x (that, who knows, they might even release some day) > > I'll see about uploading an updated debian package with this too. That would be cool. Thanks.
Bug#841853: inkscape: Mouse cursor is a white outline on PowerPC
On Sun, 6 Nov 2016 15:00:39 + Mattia Rizzolowrote: > Thanks for doing it. > > Thanks for the patch! Thanks. > Though, I'd like to ask you to foward this patch upstream. It's > enough to open a MR against > https://code.launchpad.net/~inkscape.dev/inkscape/trunk for it. > > Could you do it? > Since today my internet connection is unusually and abysmally slow, checking out of the Inkscape repository was essentially moving at dial-up speeds. So instead I have submitted a bug report with the attached patch, along with a reference to this bug. * https://bugs.launchpad.net/inkscape/+bug/1639611 > ↑↑↑ be aware of these 4 trailing whitespaces, some people don't like > them. I removed the trailing whitespace from the patch before uploading it to their bug tracker. The spaces were made by auto-indent.
Bug#841853: inkscape: Mouse cursor is a white outline on PowerPC
Taking a look at this myself. The issue is related to cursors and is either an issue in: * Inkscape * GDK * XRender Since the star cursor in Inkscape shows the color to be drawn, including alpha. Using the color `FFDD8833` (which is kind of a tan color) appears on the star cursor itself as a sky blue. The actual color being shown on the cursor is `#3388DD` (or ARGB #FF3388DD). Changing in Inkscape the `FF` (which should be red, but is alpha) makes the blue in the cursor a bit transparent. I am definitely certain that the issue is in Inkscape only. Other programs such as GIMP and anything else do not have any cursor issues at all from what I can tell. If GDK or XRender were at fault, likely more than just a single program would be broken. The only place in the Inkscape code which loads cursors is `sp_cursor_pixbuf_from_xpm`. The `RGBA` struct has a static cast to a `guint32` which bitshifts in values accordingly. However part of the function `sp_cursor_pixbuf_from_xpm` contains the following: guint32 *pixmap_buffer = new guint32[width * height]; However, before returning this buffer is `reinterpret_cast`ed such as the following: return gdk_pixbuf_new_from_data(reinterpret_cast<guchar*>(pixmap_buffer), GDK_COLORSPACE_RGB, TRUE, 8, width, height, width * sizeof(guint32), free_cursor_data, NULL); Looking at the documentation for `gdk_pixbuf_new_from_data`, the data for this call is just specified as _Image data in 8-bit/sample packed format_. I wrote the attached patch which essentially byte swaps the value before this function returns. The cursors are drawn correctly and Inkscape can now be used much more easily. If a contributor agreement is required then I would accept giving away my copyright for the code contained in this patch. Description: Swap before return sp_cursor_pixbuf_from_xpm on big endian. This byte swaps before the return in sp_cursor_pixbuf_from_xpm on big endian systems so that the cursor is made visible on these systems. Author: Steven Gawroriski <ste...@multiphasicapps.net> --- Origin: other Bug-Debian: https://bugs.debian.org/841853 --- inkscape-0.91.orig/src/sp-cursor.cpp +++ inkscape-0.91/src/sp-cursor.cpp @@ -106,6 +106,14 @@ GdkPixbuf *sp_cursor_pixbuf_from_xpm(gch pixmap_buffer[y * width + x] = (it == colorMap.end()) ? 0u : it->second; } } + +#if G_BYTE_ORDER == G_BIG_ENDIAN +for (int i = 0, n = width * height; i < n; i++) +{ +guint32 v = pixmap_buffer[i]; +pixmap_buffer[i] = ((v & 0xFF) << 24) | (((v >> 8) & 0xFF) << 16) | (((v >> 16) & 0xFF) << 8) | ((v >> 24) & 0xFF); +} +#endif return gdk_pixbuf_new_from_data(reinterpret_cast<guchar*>(pixmap_buffer), GDK_COLORSPACE_RGB, TRUE, 8, width, height, width * sizeof(guint32), free_cursor_data, NULL); }
Bug#841853: inkscape: Mouse cursor is a white outline on PowerPC
Package: inkscape Version: 0.91-11 Severity: important Dear Maintainer, On PowerPC, cursors in Inkscape are just a white outline. On other systems there should be black in the center or at least in some situations a black outline around the white outline. The parts that should be black are invisible. This makes Inkscape very difficult to use because on very bright colors, the mouse cursor is near-impossible or impossible to see. Since this system is big endian, it is likely that the channels are being read incorrectly (FF00 or similar becomes 00FF). -- System Information: Debian Release: stretch/sid APT prefers testing APT policy: (500, 'testing') Architecture: powerpc (ppc) Foreign Architectures: i386 Kernel: Linux 4.8.3.161021.1 Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/dash Init: systemd (via /run/systemd/system) Versions of packages inkscape depends on: ii libaspell150.60.7~20110707-3+b1 ii libatk1.0-02.22.0-1 ii libatkmm-1.6-1v5 2.24.2-2 ii libc6 2.24-3 ii libcairo2 1.14.6-1+b1 ii libcairomm-1.0-1v5 1.12.0-1+b1 ii libcdr-0.1-1 0.1.3-3 ii libexif12 0.6.21-2 ii libfontconfig1 2.11.0-6.7 ii libfreetype6 2.6.3-3+b1 ii libgc1c2 1:7.4.2-8 ii libgcc11:6.2.0-6 ii libgdk-pixbuf2.0-0 2.36.0-1 ii libglib2.0-0 2.50.1-1 ii libglibmm-2.4-1v5 2.50.0-1 ii libgomp1 6.2.0-6 ii libgsl22.2.1+dfsg-1 ii libgtk2.0-02.24.31-1 ii libgtkmm-2.4-1v5 1:2.24.5-1 ii libgtkspell0 2.0.16-1.1 ii libjpeg62-turbo1:1.5.0-1 ii liblcms2-2 2.7-1 ii libmagick++-6.q16-5v5 8:6.8.9.9-7.2+b1 ii libmagickcore-6.q16-2 8:6.8.9.9-7.2+b1 ii libmagickwand-6.q16-2 8:6.8.9.9-7.2+b1 ii libpango-1.0-0 1.40.3-2 ii libpangocairo-1.0-01.40.3-2 ii libpangoft2-1.0-0 1.40.3-2 ii libpangomm-1.4-1v5 2.40.1-3 ii libpng16-161.6.25-2 ii libpoppler-glib8 0.44.0-3 ii libpoppler61 0.44.0-3 ii libpopt0 1.16-10 ii librevenge-0.0-0 0.0.4-6 ii libsigc++-2.0-0v5 2.10.0-1 ii libstdc++6 6.2.0-6 ii libvisio-0.1-1 0.1.5-4 ii libwpg-0.3-3 0.3.1-3 ii libx11-6 2:1.6.3-1 ii libxml22.9.4+dfsg1-2 ii libxslt1.1 1.1.29-1 pn python:any ii zlib1g 1:1.2.8.dfsg-2+b1 Versions of packages inkscape recommends: ii aspell0.60.7~20110707-3+b1 ii fig2dev [transfig]1:3.2.6-3 ii imagemagick 8:6.8.9.9-7.2+b1 ii libimage-magick-perl 8:6.8.9.9-7.2 pn libwmf-bin ii python-lxml 3.6.4-1 ii python-numpy 1:1.11.2-1 ii transfig 1:3.2.6-3 Versions of packages inkscape suggests: pn dia | dia-gnome pn libsvg-perl pn libxml-xql-perl pn pstoedit ii python-uniconvertor 1.1.5-3 ii ruby 1:2.3.0+4 -- no debconf information
Bug#841229: openjdk-8-jre-jamvm: OutOfMemoryError in ConcurrentHashMap
Package: openjdk-8-jre-jamvm Version: 8u102-b14.1-2 Severity: grave Justification: renders package unusable Dear Maintainer, Using JamVM causes the following exception to be thrown running the most simplest programs from "Hello World" to the Java compiler: Error: A JNI error has occurred, please check your installation and try again java.lang.OutOfMemoryError at java.util.concurrent.ConcurrentHashMap.initTable(ConcurrentHashMap.java:2233) at java.util.concurrent.ConcurrentHashMap.putVal(ConcurrentHashMap.java:1017) at java.util.concurrent.ConcurrentHashMap.put(ConcurrentHashMap.java:1006) at sun.util.locale.LocaleObjectCache.put(LocaleObjectCache.java:87) at sun.util.locale.BaseLocale.createInstance(BaseLocale.java:69) at java.util.Locale.createConstant(Locale.java:709) at java.util.Locale.(Locale.java:490) at java.lang.String.toLowerCase(String.java:2670) at java.net.URL.(URL.java:385) at java.net.URL.(URL.java:310) at java.net.URL.(URL.java:333) at sun.net.www.ParseUtil.fileToEncodedURL(ParseUtil.java:272) at sun.misc.Launcher.getFileURL(Launcher.java:484) at sun.misc.Launcher$ExtClassLoader.getExtURLs(Launcher.java:194) at sun.misc.Launcher$ExtClassLoader.(Launcher.java:164) at sun.misc.Launcher$ExtClassLoader$1.run(Launcher.java:148) at sun.misc.Launcher$ExtClassLoader$1.run(Launcher.java:142) at java.security.AccessController.doPrivileged(Native Method) at sun.misc.Launcher$ExtClassLoader.getExtClassLoader(Launcher.java:141) at sun.misc.Launcher.(Launcher.java:71) at sun.misc.Launcher.(Launcher.java:57) at java.lang.ClassLoader.initSystemClassLoader(ClassLoader.java:1451) at java.lang.ClassLoader.getSystemClassLoader(ClassLoader.java:1436) at sun.launcher.LauncherHelper.(LauncherHelper.java:92) HotSpot Zero (which is far slower) is not affected by this issue. -- System Information: Debian Release: stretch/sid APT prefers testing APT policy: (500, 'testing') Architecture: powerpc (ppc) Kernel: Linux 4.7.0-1-powerpc Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/dash Init: systemd (via /run/systemd/system) Versions of packages openjdk-8-jre-jamvm depends on: ii libc6 2.24-3 ii openjdk-8-jre-headless 8u102-b14.1-2 ii zlib1g 1:1.2.8.dfsg-2+b1 openjdk-8-jre-jamvm recommends no packages. openjdk-8-jre-jamvm suggests no packages. -- no debconf information