[RESEND] monitor: Fix return type ofmonitor_fdset_dup_fd_find

monitor_fdset_dup_fd_find_remove() and monitor_fdset_dup_fd_find()
returns mon_fdset->id which is int64_t. Downcast from int64_t to int leads to
a bug with removing fd from fdset which id >= 2^32.
So, fix return types for these function.

Signed-off-by: Yury Kotov <yury-kotov@yandex-team.ru>
---
 include/monitor/monitor.h | 2 +-
 monitor.c                 | 4 ++--
 stubs/fdset.c             | 2 +-
 3 files changed, 4 insertions(+), 4 deletions(-)

 }
--=20
2.21.0

Yury Kotov <yury-kotov@yandex-team.ru> writes:

> monitor_fdset_dup_fd_find_remove() and monitor_fdset_dup_fd_find()
> returns mon_fdset->id which is int64_t. Downcast from int64_t to int leads to

Grammar nits:

    s/returns/return/
    s/Downcast/Downcasting/

> a bug with removing fd from fdset which id >= 2^32.

s/which/with/

> So, fix return types for these function.
>
> Signed-off-by: Yury Kotov <yury-kotov@yandex-team.ru>
> ---

If I feed your message to git-am, I get

    Applying: monitor: Fix return type of monitor_fdset_dup_fd_find
    error: corrupt patch at line 12
    Patch failed at 0001 monitor: Fix return type of monitor_fdset_dup_fd_find
    [...]

Did you use git-send-email?

>  include/monitor/monitor.h | 2 +-
>  monitor.c                 | 4 ++--
>  stubs/fdset.c             | 2 +-
>  3 files changed, 4 insertions(+), 4 deletions(-)
>
> diff --git a/include/monitor/monitor.h b/include/monitor/monitor.h
> index c1b40a9cac..2872621afd 100644
> --- a/include/monitor/monitor.h
> +++ b/include/monitor/monitor.h
> @@ -46,7 +46,7 @@ AddfdInfo *monitor_fdset_add_fd(int fd, bool has_fdset_=
> id, int64_t fdset_id,
>  int monitor_fdset_get_fd(int64_t fdset_id, int flags);
>  int monitor_fdset_dup_fd_add(int64_t fdset_id, int dup_fd);
>  void monitor_fdset_dup_fd_remove(int dup_fd);
> -int monitor_fdset_dup_fd_find(int dup_fd);
> +int64_t monitor_fdset_dup_fd_find(int dup_fd);
> =20

Looks mime-damaged.

>  void monitor_vfprintf(FILE *stream,
>                        const char *fmt, va_list ap) GCC_FMT_ATTR(2, 0);
> diff --git a/monitor.c b/monitor.c
> index 4807bbe811..50e6e820d6 100644
> --- a/monitor.c
> +++ b/monitor.c
> @@ -2585,7 +2585,7 @@ err:
>      return -1;
>  }
> =20
> -static int monitor_fdset_dup_fd_find_remove(int dup_fd, bool remove)
> +static int64_t monitor_fdset_dup_fd_find_remove(int dup_fd, bool remove)
>  {
>      MonFdset *mon_fdset;
>      MonFdsetFd *mon_fdset_fd_dup;
> @@ -2613,7 +2613,7 @@ err:
>      return -1;
>  }
> =20
> -int monitor_fdset_dup_fd_find(int dup_fd)
> +int64_t monitor_fdset_dup_fd_find(int dup_fd)
>  {
>      return monitor_fdset_dup_fd_find_remove(dup_fd, false);
>  }
> diff --git a/stubs/fdset.c b/stubs/fdset.c
> index 4f3edf2ea4..a1b8f41f62 100644
> --- a/stubs/fdset.c
> +++ b/stubs/fdset.c
> @@ -7,7 +7,7 @@ int monitor_fdset_dup_fd_add(int64_t fdset_id, int dup_fd=
> )
>      return -1;
>  }
> =20
> -int monitor_fdset_dup_fd_find(int dup_fd)
> +int64_t monitor_fdset_dup_fd_find(int dup_fd)
>  {
>      return -1;
>  }
> --=20
> 2.21.0

The patch is complete because:

* monitor_fdset_dup_fd_find_remove() is used only by
  monitor_fdset_dup_fd_find(), which you fix as well, and
  monitor_fdset_dup_fd_remove(), which ignores the return value.

* monitor_fdset_dup_fd_find() is used only by qemu_close(), which stores
  the return value in an int64_t.

Reviewed-by: Markus Armbruster <armbru@redhat.com>

On 5/14/19 8:15 AM, Yury Kotov wrote:
> monitor_fdset_dup_fd_find_remove() and monitor_fdset_dup_fd_find()
> returns mon_fdset->id which is int64_t. Downcast from int64_t to int leads to
> a bug with removing fd from fdset which id >= 2^32.
> So, fix return types for these function.

fd's cannot exceed 2^32. We should instead be fixing anything that uses
int64_t with an fd to be properly limited to 32 bits.  That is, I think
the real problem is in qapi/misc.json:

 { 'struct': 'AddfdInfo', 'data': {'fdset-id': 'int', 'fd': 'int'} }
instead of 'fd':'int32'.  For that matter, 'fdset-id' larger than 32
bits is unlikely to be useful (there's no reason to have more fdsets
than you can have possible fds to put in those sets).

NACK to this version, but a v2 that addresses the real problem is
worthwhile.

> +++ b/include/monitor/monitor.h
> @@ -46,7 +46,7 @@ AddfdInfo *monitor_fdset_add_fd(int fd, bool has_fdset_=
> id, int64_t fdset_id,
>  int monitor_fdset_get_fd(int64_t fdset_id, int flags);
>  int monitor_fdset_dup_fd_add(int64_t fdset_id, int dup_fd);
>  void monitor_fdset_dup_fd_remove(int dup_fd);
> -int monitor_fdset_dup_fd_find(int dup_fd);
> +int64_t monitor_fdset_dup_fd_find(int dup_fd);
> =20

Your patch came through corrupted. You may want to double-check how you
are sending them, to ensure they are not mangled.

14.05.2019, 17:05, "Eric Blake" <eblake@redhat.com>:
> On 5/14/19 8:15 AM, Yury Kotov wrote:
>>  monitor_fdset_dup_fd_find_remove() and monitor_fdset_dup_fd_find()
>>  returns mon_fdset->id which is int64_t. Downcast from int64_t to int leads to
>>  a bug with removing fd from fdset which id >= 2^32.
>>  So, fix return types for these function.
>
> fd's cannot exceed 2^32. We should instead be fixing anything that uses
> int64_t with an fd to be properly limited to 32 bits. That is, I think
> the real problem is in qapi/misc.json:
>
>  { 'struct': 'AddfdInfo', 'data': {'fdset-id': 'int', 'fd': 'int'} }
> instead of 'fd':'int32'. For that matter, 'fdset-id' larger than 32
> bits is unlikely to be useful (there's no reason to have more fdsets
> than you can have possible fds to put in those sets).
>
> NACK to this version, but a v2 that addresses the real problem is
> worthwhile.
>

Ok, so I will change fdset_id type int64_t -> int everywhere and
int64_t -> int32_t for qmp commands. Right?

>>  +++ b/include/monitor/monitor.h
>>  @@ -46,7 +46,7 @@ AddfdInfo *monitor_fdset_add_fd(int fd, bool has_fdset_=
>>  id, int64_t fdset_id,
>>   int monitor_fdset_get_fd(int64_t fdset_id, int flags);
>>   int monitor_fdset_dup_fd_add(int64_t fdset_id, int dup_fd);
>>   void monitor_fdset_dup_fd_remove(int dup_fd);
>>  -int monitor_fdset_dup_fd_find(int dup_fd);
>>  +int64_t monitor_fdset_dup_fd_find(int dup_fd);
>>  =20
>
> Your patch came through corrupted. You may want to double-check how you
> are sending them, to ensure they are not mangled.
>

Omg, sorry. I just copy-pasted my previous patch from mail client in raw format
and I did not notice the escape characters, because of which the patch is
incorrect.

Regards,
Yury