| Message ID | 1516112253-14480-18-git-send-email-pbonzini@redhat.com |
|---|---|
| State | New |
| Headers | show |
| Series | [PULL,01/51] scsi-generic: Add share-rw option | expand |
On Tue, 16 Jan 2018 15:16:59 +0100 Paolo Bonzini <pbonzini@redhat.com> wrote: > From: Marc-André Lureau <marcandre.lureau@redhat.com> > > Fixes leaks such as: > > Direct leak of 2 byte(s) in 1 object(s) allocated from: > #0 0x7eff58beb850 in malloc (/lib64/libasan.so.4+0xde850) > #1 0x7eff57942f0c in g_malloc ../glib/gmem.c:94 > #2 0x7eff579431cf in g_malloc_n ../glib/gmem.c:331 > #3 0x7eff5795f6eb in g_strdup ../glib/gstrfuncs.c:363 > #4 0x55db720f1d46 in readline_hist_add /home/elmarco/src/qq/util/readline.c:258 > #5 0x55db720f2d34 in readline_handle_byte /home/elmarco/src/qq/util/readline.c:387 > #6 0x55db71539d00 in monitor_read /home/elmarco/src/qq/monitor.c:3896 > #7 0x55db71f9be35 in qemu_chr_be_write_impl /home/elmarco/src/qq/chardev/char.c:167 > #8 0x55db71f9bed3 in qemu_chr_be_write /home/elmarco/src/qq/chardev/char.c:179 > #9 0x55db71fa013c in fd_chr_read /home/elmarco/src/qq/chardev/char-fd.c:66 > #10 0x55db71fe18a8 in qio_channel_fd_source_dispatch /home/elmarco/src/qq/io/channel-watch.c:84 > #11 0x7eff5793a90b in g_main_dispatch ../glib/gmain.c:3182 > #12 0x7eff5793b7ac in g_main_context_dispatch ../glib/gmain.c:3847 > #13 0x55db720af3bd in glib_pollfds_poll /home/elmarco/src/qq/util/main-loop.c:214 > #14 0x55db720af505 in os_host_main_loop_wait /home/elmarco/src/qq/util/main-loop.c:261 > #15 0x55db720af6d6 in main_loop_wait /home/elmarco/src/qq/util/main-loop.c:515 > #16 0x55db7184e0de in main_loop /home/elmarco/src/qq/vl.c:1995 > #17 0x55db7185e956 in main /home/elmarco/src/qq/vl.c:4914 > #18 0x7eff4ea17039 in __libc_start_main (/lib64/libc.so.6+0x21039) > > (while at it, use g_new0(ReadLineState), it's a bit easier to read) > > Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com> > Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com> > Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> > Message-Id: <20180104160523.22995-11-marcandre.lureau@redhat.com> > Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> > --- I'm having some trouble with this patch, using b05631954d6d: # /usr/local/bin/qemu-system-x86_64 -m 1G -nodefaults -net none -monitor stdio -serial none -parallel none -nographic QEMU 2.11.50 monitor - type 'help' for more information (qemu) sys<tab> system_powerdown system_reset system_wakeup (qemu) system_p<tab> # resulting in system_powerdown (qemu) quit Segmentation fault (core dumped) gdb shows: Thread 1 "qemu-system-x86" received signal SIGSEGV, Segmentation fault. 0x00007f7d64d82927 in malloc () from /lib64/libc.so.6 (gdb) bt #0 0x00007f7d64d82927 in malloc () at /lib64/libc.so.6 #1 0x00007f7d6ef68359 in g_malloc () at /lib64/libglib-2.0.so.0 #2 0x00007f7d6ef83004 in g_strsplit () at /lib64/libglib-2.0.so.0 #3 0x000055e5ac0d549d in container_get (root=0x55e5ad570ee0, path=path@entry=0x55e5ac2fa0f8 "/chardevs") at qom/container.c:34 #4 0x000055e5ac14d102 in get_chardevs_root () at chardev/char.c:43 #5 0x000055e5ac14ec4d in qemu_chr_cleanup () at chardev/char.c:1107 #6 0x000055e5abeff1c4 in main (argc=<optimized out>, argv=<optimized out>, envp=<optimized out>) at vl.c:4780 Reverting this patch, commit e5dc1a6c6c435, I don't see the issue. Thanks, Alex
On 01/02/2018 19:00, Alex Williamson wrote: > On Tue, 16 Jan 2018 15:16:59 +0100 > Paolo Bonzini <pbonzini@redhat.com> wrote: > >> From: Marc-André Lureau <marcandre.lureau@redhat.com> >> >> Fixes leaks such as: >> >> Direct leak of 2 byte(s) in 1 object(s) allocated from: >> #0 0x7eff58beb850 in malloc (/lib64/libasan.so.4+0xde850) >> #1 0x7eff57942f0c in g_malloc ../glib/gmem.c:94 >> #2 0x7eff579431cf in g_malloc_n ../glib/gmem.c:331 >> #3 0x7eff5795f6eb in g_strdup ../glib/gstrfuncs.c:363 >> #4 0x55db720f1d46 in readline_hist_add /home/elmarco/src/qq/util/readline.c:258 >> #5 0x55db720f2d34 in readline_handle_byte /home/elmarco/src/qq/util/readline.c:387 >> #6 0x55db71539d00 in monitor_read /home/elmarco/src/qq/monitor.c:3896 >> #7 0x55db71f9be35 in qemu_chr_be_write_impl /home/elmarco/src/qq/chardev/char.c:167 >> #8 0x55db71f9bed3 in qemu_chr_be_write /home/elmarco/src/qq/chardev/char.c:179 >> #9 0x55db71fa013c in fd_chr_read /home/elmarco/src/qq/chardev/char-fd.c:66 >> #10 0x55db71fe18a8 in qio_channel_fd_source_dispatch /home/elmarco/src/qq/io/channel-watch.c:84 >> #11 0x7eff5793a90b in g_main_dispatch ../glib/gmain.c:3182 >> #12 0x7eff5793b7ac in g_main_context_dispatch ../glib/gmain.c:3847 >> #13 0x55db720af3bd in glib_pollfds_poll /home/elmarco/src/qq/util/main-loop.c:214 >> #14 0x55db720af505 in os_host_main_loop_wait /home/elmarco/src/qq/util/main-loop.c:261 >> #15 0x55db720af6d6 in main_loop_wait /home/elmarco/src/qq/util/main-loop.c:515 >> #16 0x55db7184e0de in main_loop /home/elmarco/src/qq/vl.c:1995 >> #17 0x55db7185e956 in main /home/elmarco/src/qq/vl.c:4914 >> #18 0x7eff4ea17039 in __libc_start_main (/lib64/libc.so.6+0x21039) >> >> (while at it, use g_new0(ReadLineState), it's a bit easier to read) >> >> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com> >> Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com> >> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> >> Message-Id: <20180104160523.22995-11-marcandre.lureau@redhat.com> >> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> >> --- > > I'm having some trouble with this patch, using b05631954d6d: > > # /usr/local/bin/qemu-system-x86_64 -m 1G -nodefaults -net none -monitor stdio -serial none -parallel none -nographic > QEMU 2.11.50 monitor - type 'help' for more information > (qemu) sys<tab> > system_powerdown system_reset system_wakeup > (qemu) system_p<tab> # resulting in system_powerdown > (qemu) quit > Segmentation fault (core dumped) > > gdb shows: > > Thread 1 "qemu-system-x86" received signal SIGSEGV, Segmentation fault. > 0x00007f7d64d82927 in malloc () from /lib64/libc.so.6 > (gdb) bt > #0 0x00007f7d64d82927 in malloc () at /lib64/libc.so.6 > #1 0x00007f7d6ef68359 in g_malloc () at /lib64/libglib-2.0.so.0 > #2 0x00007f7d6ef83004 in g_strsplit () at /lib64/libglib-2.0.so.0 > #3 0x000055e5ac0d549d in container_get (root=0x55e5ad570ee0, path=path@entry=0x55e5ac2fa0f8 "/chardevs") at qom/container.c:34 > #4 0x000055e5ac14d102 in get_chardevs_root () at chardev/char.c:43 > #5 0x000055e5ac14ec4d in qemu_chr_cleanup () at chardev/char.c:1107 > #6 0x000055e5abeff1c4 in main (argc=<optimized out>, argv=<optimized out>, envp=<optimized out>) at vl.c:4780 > > Reverting this patch, commit e5dc1a6c6c435, I don't see the issue. > Thanks, Yeah, I have a fix queued. Unfortunately, I don't have the usual setup to do pre-pull-request sets here so it will have to wait for next Monday. Paolo
On Thu, 1 Feb 2018 19:10:06 -0500 Paolo Bonzini <pbonzini@redhat.com> wrote: > On 01/02/2018 19:00, Alex Williamson wrote: > > On Tue, 16 Jan 2018 15:16:59 +0100 > > Paolo Bonzini <pbonzini@redhat.com> wrote: > > > >> From: Marc-André Lureau <marcandre.lureau@redhat.com> > >> > >> Fixes leaks such as: > >> > >> Direct leak of 2 byte(s) in 1 object(s) allocated from: > >> #0 0x7eff58beb850 in malloc (/lib64/libasan.so.4+0xde850) > >> #1 0x7eff57942f0c in g_malloc ../glib/gmem.c:94 > >> #2 0x7eff579431cf in g_malloc_n ../glib/gmem.c:331 > >> #3 0x7eff5795f6eb in g_strdup ../glib/gstrfuncs.c:363 > >> #4 0x55db720f1d46 in readline_hist_add /home/elmarco/src/qq/util/readline.c:258 > >> #5 0x55db720f2d34 in readline_handle_byte /home/elmarco/src/qq/util/readline.c:387 > >> #6 0x55db71539d00 in monitor_read /home/elmarco/src/qq/monitor.c:3896 > >> #7 0x55db71f9be35 in qemu_chr_be_write_impl /home/elmarco/src/qq/chardev/char.c:167 > >> #8 0x55db71f9bed3 in qemu_chr_be_write /home/elmarco/src/qq/chardev/char.c:179 > >> #9 0x55db71fa013c in fd_chr_read /home/elmarco/src/qq/chardev/char-fd.c:66 > >> #10 0x55db71fe18a8 in qio_channel_fd_source_dispatch /home/elmarco/src/qq/io/channel-watch.c:84 > >> #11 0x7eff5793a90b in g_main_dispatch ../glib/gmain.c:3182 > >> #12 0x7eff5793b7ac in g_main_context_dispatch ../glib/gmain.c:3847 > >> #13 0x55db720af3bd in glib_pollfds_poll /home/elmarco/src/qq/util/main-loop.c:214 > >> #14 0x55db720af505 in os_host_main_loop_wait /home/elmarco/src/qq/util/main-loop.c:261 > >> #15 0x55db720af6d6 in main_loop_wait /home/elmarco/src/qq/util/main-loop.c:515 > >> #16 0x55db7184e0de in main_loop /home/elmarco/src/qq/vl.c:1995 > >> #17 0x55db7185e956 in main /home/elmarco/src/qq/vl.c:4914 > >> #18 0x7eff4ea17039 in __libc_start_main (/lib64/libc.so.6+0x21039) > >> > >> (while at it, use g_new0(ReadLineState), it's a bit easier to read) > >> > >> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com> > >> Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com> > >> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> > >> Message-Id: <20180104160523.22995-11-marcandre.lureau@redhat.com> > >> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> > >> --- > > > > I'm having some trouble with this patch, using b05631954d6d: > > > > # /usr/local/bin/qemu-system-x86_64 -m 1G -nodefaults -net none -monitor stdio -serial none -parallel none -nographic > > QEMU 2.11.50 monitor - type 'help' for more information > > (qemu) sys<tab> > > system_powerdown system_reset system_wakeup > > (qemu) system_p<tab> # resulting in system_powerdown > > (qemu) quit > > Segmentation fault (core dumped) > > > > gdb shows: > > > > Thread 1 "qemu-system-x86" received signal SIGSEGV, Segmentation fault. > > 0x00007f7d64d82927 in malloc () from /lib64/libc.so.6 > > (gdb) bt > > #0 0x00007f7d64d82927 in malloc () at /lib64/libc.so.6 > > #1 0x00007f7d6ef68359 in g_malloc () at /lib64/libglib-2.0.so.0 > > #2 0x00007f7d6ef83004 in g_strsplit () at /lib64/libglib-2.0.so.0 > > #3 0x000055e5ac0d549d in container_get (root=0x55e5ad570ee0, path=path@entry=0x55e5ac2fa0f8 "/chardevs") at qom/container.c:34 > > #4 0x000055e5ac14d102 in get_chardevs_root () at chardev/char.c:43 > > #5 0x000055e5ac14ec4d in qemu_chr_cleanup () at chardev/char.c:1107 > > #6 0x000055e5abeff1c4 in main (argc=<optimized out>, argv=<optimized out>, envp=<optimized out>) at vl.c:4780 > > > > Reverting this patch, commit e5dc1a6c6c435, I don't see the issue. > > Thanks, > > Yeah, I have a fix queued. Unfortunately, I don't have the usual setup > to do pre-pull-request sets here so it will have to wait for next Monday. > > Paolo > The queued fix is: http://patchwork.ozlabs.org/patch/862816/
diff --git a/include/qemu/readline.h b/include/qemu/readline.h index c08cf74..e812583 100644 --- a/include/qemu/readline.h +++ b/include/qemu/readline.h @@ -59,5 +59,6 @@ ReadLineState *readline_init(ReadLinePrintfFunc *printf_func, ReadLineFlushFunc *flush_func, void *opaque, ReadLineCompletionFunc *completion_finder); +void readline_free(ReadLineState *rs); #endif /* READLINE_H */ diff --git a/monitor.c b/monitor.c index d682eee..b9da5e2 100644 --- a/monitor.c +++ b/monitor.c @@ -583,7 +583,7 @@ static void monitor_data_destroy(Monitor *mon) if (monitor_is_qmp(mon)) { json_message_parser_destroy(&mon->qmp.parser); } - g_free(mon->rs); + readline_free(mon->rs); QDECREF(mon->outbuf); qemu_mutex_destroy(&mon->out_lock); } diff --git a/util/readline.c b/util/readline.c index bbdee79..24ec839 100644 --- a/util/readline.c +++ b/util/readline.c @@ -500,12 +500,28 @@ const char *readline_get_history(ReadLineState *rs, unsigned int index) return rs->history[index]; } +void readline_free(ReadLineState *rs) +{ + int i; + + if (!rs) { + return; + } + for (i = 0; i < READLINE_MAX_CMDS; i++) { + g_free(rs->history[i]); + } + for (i = 0; i < READLINE_MAX_COMPLETIONS; i++) { + g_free(rs->completions[i]); + } + g_free(rs); +} + ReadLineState *readline_init(ReadLinePrintfFunc *printf_func, ReadLineFlushFunc *flush_func, void *opaque, ReadLineCompletionFunc *completion_finder) { - ReadLineState *rs = g_malloc0(sizeof(*rs)); + ReadLineState *rs = g_new0(ReadLineState, 1); rs->hist_entry = -1; rs->opaque = opaque;