| Message ID | 1476608111-28948-1-git-send-email-amir73il@gmail.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Sun, Oct 16, 2016 at 11:55:11AM +0300, Amir Goldstein wrote: > xfs_io -M will start by spawning an idle thread. > The purpose of this idle thread is to test io from a multi threaded > process. With single threaded process, the file table is not shared > and file structs are not reference counted. I can see this option being confusing to those who do not understand what is actually be done here. i.e. we'll get complaints along the lines of "-M say xfs_io does multithreaded IO, but I only see IO from one thread!". Can we name this "-i" to mean "run an idle-loop", which will stop people trying to use it for something it doesn't do? > --- a/io/Makefile > +++ b/io/Makefile .... > --- a/io/init.c > +++ b/io/init.c .... Missing a man page update to document this new option. Cheers, Dave.
On Mon, Oct 17, 2016 at 12:28 AM, Dave Chinner <david@fromorbit.com> wrote: > On Sun, Oct 16, 2016 at 11:55:11AM +0300, Amir Goldstein wrote: >> xfs_io -M will start by spawning an idle thread. >> The purpose of this idle thread is to test io from a multi threaded >> process. With single threaded process, the file table is not shared >> and file structs are not reference counted. > > I can see this option being confusing to those who do not understand > what is actually be done here. i.e. we'll get complaints along the > lines of "-M say xfs_io does multithreaded IO, but I only see IO > from one thread!". Can we name this "-i" to mean "run an idle-loop", > which will stop people trying to use it for something it doesn't do? > I considered those 2 options. I see what you are saying. Will go for -i in v3. >> --- a/io/Makefile >> +++ b/io/Makefile > .... >> --- a/io/init.c >> +++ b/io/init.c > .... > > Missing a man page update to document this new option. > > Cheers, > > Dave. > -- > Dave Chinner > david@fromorbit.com -- To unsubscribe from this list: send the line "unsubscribe fstests" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/io/Makefile b/io/Makefile index 1997ca9..032a8c7 100644 --- a/io/Makefile +++ b/io/Makefile @@ -13,7 +13,7 @@ CFILES = init.c \ mmap.c open.c parent.c pread.c prealloc.c pwrite.c seek.c shutdown.c \ sync.c truncate.c reflink.c -LLDLIBS = $(LIBXCMD) $(LIBHANDLE) +LLDLIBS = $(LIBXCMD) $(LIBHANDLE) $(LIBPTHREAD) LTDEPENDENCIES = $(LIBXCMD) $(LIBHANDLE) LLDFLAGS = -static-libtool-libs diff --git a/io/init.c b/io/init.c index 6b88cc6..8a34ceb 100644 --- a/io/init.c +++ b/io/init.c @@ -16,6 +16,7 @@ * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA */ +#include <pthread.h> #include "platform_defs.h" #include "command.h" #include "input.h" @@ -25,6 +26,7 @@ char *progname; int exitcode; int expert; +int multithreaded; size_t pagesize; struct timeval stopwatch; @@ -141,7 +143,7 @@ init( pagesize = getpagesize(); gettimeofday(&stopwatch, NULL); - while ((c = getopt(argc, argv, "ac:dFfmp:nrRstTVx")) != EOF) { + while ((c = getopt(argc, argv, "ac:dFfmMp:nrRstTVx")) != EOF) { switch (c) { case 'a': flags |= IO_APPEND; @@ -166,6 +168,9 @@ init( exit(1); } break; + case 'M': + multithreaded = 1; + break; case 'n': flags |= IO_NONBLOCK; break; @@ -213,12 +218,38 @@ init( add_check_command(init_check_command); } +/* + * The purpose of this idle thread is to test io from a multi threaded process. + * With single threaded process, the file table is not shared and file structs + * are not reference counted. So in order to detect file struct reference + * leaks, spawning an idle thread will do the trick. + */ +void * +idle_loop(void *arg) +{ + for (;;) + pause(); +} + +void +start_idle_thread() +{ + pthread_t t; + + if (pthread_create(&t, NULL, idle_loop, NULL)) { + fprintf(stderr, "Error creating idle thread\n"); + exit(1); + } +} + int main( int argc, char **argv) { init(argc, argv); + if (multithreaded) + start_idle_thread(); command_loop(); return exitcode; }