[Qemu-devel] [PATCH] gluster: Correctly propagate errors when volume isn't accessible
The docs for glfs_init suggest that the function sets errno on every failure. In fact it doesn't. As other functions such as qemu_gluster_open() in the gluster block code report their errors based on this fact we need to make sure that errno is set on each failure. This fixes a crash of qemu-img/qemu when a gluster brick isn't accessible from given host while the server serving the volume description is. Thread 1 (Thread 0x77fba740 (LWP 203880)): #0 0x777673f8 in glfs_lseek () from /usr/lib64/libgfapi.so.0 #1 0x55574a68 in qemu_gluster_getlength () #2 0x55565742 in refresh_total_sectors () #3 0x5556914f in bdrv_open_common () #4 0x5556e8e8 in bdrv_open () #5 0x5556f02f in bdrv_open_image () #6 0x5556e5f6 in bdrv_open () #7 0x555c5775 in bdrv_new_open () #8 0x555c5b91 in img_info () #9 0x762c9c05 in __libc_start_main () from /lib64/libc.so.6 #10 0x555648ad in _start () --- Notes: I'm also going to report a bug in libgfapi too. block/gluster.c | 7 ++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/block/gluster.c b/block/gluster.c index 8836085..d0726ec 100644 --- a/block/gluster.c +++ b/block/gluster.c @@ -207,6 +207,11 @@ static struct glfs *qemu_gluster_init(GlusterConf *gconf, const char *filename, volume=%s image=%s transport=%s, gconf-server, gconf-port, gconf-volname, gconf-image, gconf-transport); + +/* glfs_init sometimes doesn't set errno although docs suggest that */ +if (errno == 0) +errno = EINVAL; + goto out; } return glfs; @@ -482,7 +487,7 @@ static int qemu_gluster_create(const char *filename, glfs = qemu_gluster_init(gconf, filename, errp); if (!glfs) { -ret = -EINVAL; +ret = -errno; goto out; } -- 1.9.2
Re: [Qemu-devel] [PATCH] gluster: Correctly propagate errors when volume isn't accessible
On Fri, May 09, 2014 at 12:08:10PM +0200, Peter Krempa wrote: The docs for glfs_init suggest that the function sets errno on every failure. In fact it doesn't. As other functions such as qemu_gluster_open() in the gluster block code report their errors based on this fact we need to make sure that errno is set on each failure. This fixes a crash of qemu-img/qemu when a gluster brick isn't accessible from given host while the server serving the volume description is. Thread 1 (Thread 0x77fba740 (LWP 203880)): #0 0x777673f8 in glfs_lseek () from /usr/lib64/libgfapi.so.0 #1 0x55574a68 in qemu_gluster_getlength () #2 0x55565742 in refresh_total_sectors () #3 0x5556914f in bdrv_open_common () #4 0x5556e8e8 in bdrv_open () #5 0x5556f02f in bdrv_open_image () #6 0x5556e5f6 in bdrv_open () #7 0x555c5775 in bdrv_new_open () #8 0x555c5b91 in img_info () #9 0x762c9c05 in __libc_start_main () from /lib64/libc.so.6 #10 0x555648ad in _start () --- Notes: I'm also going to report a bug in libgfapi too. block/gluster.c | 7 ++- 1 file changed, 6 insertions(+), 1 deletion(-) Please use scripts/checkpatch.pl to check coding style in the future. I added {} around the if statement body. QEMU always uses curlies even for 1-statement bodies. Thanks, applied to my block tree: https://github.com/stefanha/qemu/commits/block Stefan
Re: [Qemu-devel] [PATCH] gluster: Correctly propagate errors when volume isn't accessible
On 05/09/14 13:39, Stefan Hajnoczi wrote: On Fri, May 09, 2014 at 12:08:10PM +0200, Peter Krempa wrote: ... Please use scripts/checkpatch.pl to check coding style in the future. I added {} around the if statement body. QEMU always uses curlies even for 1-statement bodies. Ah, right, sorry about that. I'm used to libvirt's coding style which allows it. I'll use the patch checker next time. Thanks, applied to my block tree: https://github.com/stefanha/qemu/commits/block Thanks. Stefan Peter signature.asc Description: OpenPGP digital signature