RE: [PATCH v2 02/13] block/iscsi:Remove redundant statement in iscsi_open()

["Chenqun (kuhn)" <kuhn.chenqun@huawei.com>]

>-----Original Message-----
>From: Kevin Wolf [mailto:kwolf@redhat.com]
>Sent: Thursday, February 27, 2020 6:31 PM
>To: Chenqun (kuhn) <kuhn.chenqun@huawei.com>
>Cc: qemu-devel@nongnu.org; qemu-trivial@nongnu.org;
>peter.maydell@linaro.org; Zhanghailiang <zhang.zhanghailiang@huawei.com>;
>Euler Robot <euler.robot@huawei.com>; Ronnie Sahlberg
><ronniesahlberg@gmail.com>; Paolo Bonzini <pbonzini@redhat.com>; Peter
>Lieven <pl@kamp.de>; Max Reitz <mreitz@redhat.com>
>Subject: Re: [PATCH v2 02/13] block/iscsi:Remove redundant statement in
>iscsi_open()
>
>Am 27.02.2020 um 02:49 hat Chenqun (kuhn) geschrieben:
>> >-----Original Message-----
>> >From: Kevin Wolf [mailto:kwolf@redhat.com]
>> >Sent: Wednesday, February 26, 2020 5:55 PM
>> >To: Chenqun (kuhn) <kuhn.chenqun@huawei.com>
>> >Cc: qemu-devel@nongnu.org; qemu-trivial@nongnu.org;
>> >peter.maydell@linaro.org; Zhanghailiang
>> ><zhang.zhanghailiang@huawei.com>; Euler Robot
>> ><euler.robot@huawei.com>; Ronnie Sahlberg
><ronniesahlberg@gmail.com>;
>> >Paolo Bonzini <pbonzini@redhat.com>; Peter Lieven <pl@kamp.de>; Max
>> >Reitz <mreitz@redhat.com>
>> >Subject: Re: [PATCH v2 02/13] block/iscsi:Remove redundant statement
>> >in
>> >iscsi_open()
>> >
>> >Am 26.02.2020 um 09:46 hat kuhn.chenqun@huawei.com geschrieben:
>> >> From: Chen Qun <kuhn.chenqun@huawei.com>
>> >>
>> >> Clang static code analyzer show warning:
>> >>   block/iscsi.c:1920:9: warning: Value stored to 'flags' is never read
>> >>         flags &= ~BDRV_O_RDWR;
>> >>         ^        ~~~~~~~~~~~~
>> >>
>> >> Reported-by: Euler Robot <euler.robot@huawei.com>
>> >> Signed-off-by: Chen Qun <kuhn.chenqun@huawei.com>
>> >
>> >Hmm, I'm not so sure about this one because if we remove the line,
>> >flags will be inconsistent with bs->open_flags. It feels like setting
>> >a trap for anyone who wants to add code using flags in the future.
>> Hi Kevin,
>> I find it exists since 8f3bf50d34037266.   :  )
>
>Yes, it has existed from the start with auto-read-only.
>
>> It's not a big deal,  just upset clang static code analyzer.
>> As you said, it could be a trap for the future.
>
>What's interesting is that we do have one user of the flags later in the function,
>but it uses bs->open_flags instead:
>
>    ret = iscsi_allocmap_init(iscsilun, bs->open_flags);
>
>Maybe this should be using flags? (The value of the bits we're interested in is
>the same, but when flags is passed as a parameter, I would expect it to be
>used.)
>
Hi Kevin,
I have a question: are 'flags' exactly the same as 'bs-> open_flags'? 
In the function bdrv_open_common() at block.c file,  the existence of statement( open_flags = bdrv_open_flags(bs, bs->open_flags); ) makes them a little different.
Will this place affect them inconsistently ?

Is it safer if we assign bs-> open_flags to flags?
Modify just like：
@@ -1917,7 +1917,7 @@ static int iscsi_open(BlockDriverState *bs, QDict *options, int flags,
         if (ret < 0) {
             goto out;
         }
-        flags &= ~BDRV_O_RDWR;
+        flags = bs->open_flags;
     }

     iscsi_readcapacity_sync(iscsilun, &local_err);
@@ -2002,7 +2002,7 @@ static int iscsi_open(BlockDriverState *bs, QDict *options, int flags,
         iscsilun->cluster_size = iscsilun->bl.opt_unmap_gran *
             iscsilun->block_size;
         if (iscsilun->lbprz) {
-            ret = iscsi_allocmap_init(iscsilun, bs->open_flags);
+            ret = iscsi_allocmap_init(iscsilun, flags);
         }
     }

Thanks.