Re: [Qemu-devel] [PATCH 09/26] iscsi: migrate iscsi driver QemuOptionParameter usage

[Leandro Dorileo]

On Fri, Mar 21, 2014 at 02:42:35PM +0100, Peter Lieven wrote:
> On 21.03.2014 14:31, Leandro Dorileo wrote:
> >On Fri, Mar 21, 2014 at 07:43:44AM +0100, Peter Lieven wrote:
> >>On 21.03.2014 01:13, Leandro Dorileo wrote:
> >>>Do the directly migration from QemuOptionParameter to QemuOpts on
> >>>iscsi block driver.
> >>>
> >>>Signed-off-by: Leandro Dorileo <address@hidden>
> >>>---
> >>>  block/iscsi.c | 32 ++++++++++++++++----------------
> >>>  1 file changed, 16 insertions(+), 16 deletions(-)
> >>>
> >>>diff --git a/block/iscsi.c b/block/iscsi.c
> >>>index b490e98..85252e7 100644
> >>>--- a/block/iscsi.c
> >>>+++ b/block/iscsi.c
> >>>@@ -1125,7 +1125,7 @@ static int iscsi_open(BlockDriverState *bs, QDict 
> >>>*options, int flags,
> >>>      QemuOpts *opts;
> >>>      Error *local_err = NULL;
> >>>      const char *filename;
> >>>-    int i, ret;
> >>>+    int i, ret = 0;
> >>why? is there a chance that ret remains uninitialized?
> >Yep, my compiler tells me so:
> >
> >block/iscsi.c:1128:12: error: ‘ret’ may be used uninitialized in this 
> >function [-Werror=maybe-uninitialized]
> >
> >
> >>>      if ((BDRV_SECTOR_SIZE % 512) != 0) {
> >>>          error_setg(errp, "iSCSI: Invalid BDRV_SECTOR_SIZE. "
> >>>@@ -1382,8 +1382,7 @@ static int iscsi_truncate(BlockDriverState *bs, 
> >>>int64_t offset)
> >>>      return 0;
> >>>  }
> >>>-static int iscsi_create(const char *filename, QEMUOptionParameter 
> >>>*options,
> >>>-                        Error **errp)
> >>>+static int iscsi_create(const char *filename, QemuOpts *options, Error 
> >>>**errp)
> >>>  {
> >>>      int ret = 0;
> >>>      int64_t total_size = 0;
> >>>@@ -1393,12 +1392,9 @@ static int iscsi_create(const char *filename, 
> >>>QEMUOptionParameter *options,
> >>>      bs = bdrv_new("");
> >>>-    /* Read out options */
> >>>-    while (options && options->name) {
> >>>-        if (!strcmp(options->name, "size")) {
> >>>-            total_size = options->value.n / BDRV_SECTOR_SIZE;
> >>>-        }
> >>>-        options++;
> >>>+    total_size = qemu_opt_get_size(options, BLOCK_OPT_SIZE, 0);
> >>>+    if (total_size) {
> >>>+        total_size = total_size / BDRV_SECTOR_SIZE;
> >>>      }
> >>you don't need the if condition. 0 / BDRV_SECTOR_SIZE = 0.
> >>
> >I'm not sure, bdrv_img_create() will set BLOCK_OPT_SIZE with img_size, we 
> >have no guarantee on the
> >value passed to bdrv_img_create(), we don't check img_size value there, 
> >having said that can't
> >we run on division by zero here? The previous code wasn't checking it but I 
> >wonder if the problem
> >wasn't there already.
> 
> division by zero is x / 0 not 0 / x.
> 
> 0 / x = 0
> x / 0 = undef
> 

Yep, true.

-- 
Leandro Dorileo