[PATCH] Cygwin: chown: make sure ctime gets updated when necessary

[PATCH] Cygwin: chown: make sure ctime gets updated when necessary

[Ken Brown]

Following POSIX, ensure that ctime is updated if chown succeeds,
unless the new owner is specified as (uid_t)-1 and the new group is
specified as (gid_t)-1.  Previously, ctime was unchanged whenever the
owner and group were both unchanged.

Aside from POSIX compliance, this fix makes gnulib report that chown
works on Cygwin.  This improves the efficiency of packages like GNU
tar that use gnulib's chown module.  Previously such packages would
use a gnulib replacement for chown on Cygwin.
---
 winsup/cygwin/fhandler_disk_file.cc | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/winsup/cygwin/fhandler_disk_file.cc b/winsup/cygwin/fhandler_disk_file.cc
index 07f9c513a..72d259579 100644
--- a/winsup/cygwin/fhandler_disk_file.cc
+++ b/winsup/cygwin/fhandler_disk_file.cc
@@ -863,6 +863,7 @@ fhandler_disk_file::fchown (uid_t uid, gid_t gid)
   tmp_pathbuf tp;
   aclent_t *aclp;
   int nentries;
+  bool noop = true;
 
   if (!pc.has_acls ())
     {
@@ -887,11 +888,18 @@ fhandler_disk_file::fchown (uid_t uid, gid_t gid)
 				    aclp, MAX_ACL_ENTRIES)) < 0)
     goto out;
 
+  /* According to POSIX, chown can be a no-op if uid is (uid_t)-1 and
+     gid is (gid_t)-1.  Otherwise, even if uid and gid are unchanged,
+     we must ensure that ctime is updated. */
   if (uid == ILLEGAL_UID)
     uid = old_uid;
+  else
+    noop = false;
   if (gid == ILLEGAL_GID)
     gid = old_gid;
-  if (uid == old_uid && gid == old_gid)
+  else
+    noop = false;
+  if (noop)
     {
       ret = 0;
       goto out;
-- 
2.30.0

Re: [PATCH] Cygwin: chown: make sure ctime gets updated when necessary

[Corinna Vinschen]

On Jan 25 12:24, Ken Brown via Cygwin-patches wrote:
> Following POSIX, ensure that ctime is updated if chown succeeds,
> unless the new owner is specified as (uid_t)-1 and the new group is
> specified as (gid_t)-1.  Previously, ctime was unchanged whenever the
> owner and group were both unchanged.
> 
> Aside from POSIX compliance, this fix makes gnulib report that chown
> works on Cygwin.  This improves the efficiency of packages like GNU
> tar that use gnulib's chown module.  Previously such packages would
> use a gnulib replacement for chown on Cygwin.
> ---
>  winsup/cygwin/fhandler_disk_file.cc | 10 +++++++++-
>  1 file changed, 9 insertions(+), 1 deletion(-)
> 
> diff --git a/winsup/cygwin/fhandler_disk_file.cc b/winsup/cygwin/fhandler_disk_file.cc
> index 07f9c513a..72d259579 100644
> --- a/winsup/cygwin/fhandler_disk_file.cc
> +++ b/winsup/cygwin/fhandler_disk_file.cc
> @@ -863,6 +863,7 @@ fhandler_disk_file::fchown (uid_t uid, gid_t gid)
>    tmp_pathbuf tp;
>    aclent_t *aclp;
>    int nentries;
> +  bool noop = true;
>  
>    if (!pc.has_acls ())
>      {
> @@ -887,11 +888,18 @@ fhandler_disk_file::fchown (uid_t uid, gid_t gid)
>  				    aclp, MAX_ACL_ENTRIES)) < 0)
>      goto out;
>  
> +  /* According to POSIX, chown can be a no-op if uid is (uid_t)-1 and
> +     gid is (gid_t)-1.  Otherwise, even if uid and gid are unchanged,
> +     we must ensure that ctime is updated. */
>    if (uid == ILLEGAL_UID)
>      uid = old_uid;
> +  else
> +    noop = false;
>    if (gid == ILLEGAL_GID)
>      gid = old_gid;
> -  if (uid == old_uid && gid == old_gid)

Basically ok, but why not just

     if (uid == ILLEGAL_UID && gid == ILLEGAL_GID)

instead of the noop var?


Corinna

Re: [PATCH] Cygwin: chown: make sure ctime gets updated when necessary

[Ken Brown]

On 1/25/2021 1:57 PM, Corinna Vinschen via Cygwin-patches wrote:
> On Jan 25 12:24, Ken Brown via Cygwin-patches wrote:
>> Following POSIX, ensure that ctime is updated if chown succeeds,
>> unless the new owner is specified as (uid_t)-1 and the new group is
>> specified as (gid_t)-1.  Previously, ctime was unchanged whenever the
>> owner and group were both unchanged.
>>
>> Aside from POSIX compliance, this fix makes gnulib report that chown
>> works on Cygwin.  This improves the efficiency of packages like GNU
>> tar that use gnulib's chown module.  Previously such packages would
>> use a gnulib replacement for chown on Cygwin.
>> ---
>>   winsup/cygwin/fhandler_disk_file.cc | 10 +++++++++-
>>   1 file changed, 9 insertions(+), 1 deletion(-)
>>
>> diff --git a/winsup/cygwin/fhandler_disk_file.cc b/winsup/cygwin/fhandler_disk_file.cc
>> index 07f9c513a..72d259579 100644
>> --- a/winsup/cygwin/fhandler_disk_file.cc
>> +++ b/winsup/cygwin/fhandler_disk_file.cc
>> @@ -863,6 +863,7 @@ fhandler_disk_file::fchown (uid_t uid, gid_t gid)
>>     tmp_pathbuf tp;
>>     aclent_t *aclp;
>>     int nentries;
>> +  bool noop = true;
>>   
>>     if (!pc.has_acls ())
>>       {
>> @@ -887,11 +888,18 @@ fhandler_disk_file::fchown (uid_t uid, gid_t gid)
>>   				    aclp, MAX_ACL_ENTRIES)) < 0)
>>       goto out;
>>   
>> +  /* According to POSIX, chown can be a no-op if uid is (uid_t)-1 and
>> +     gid is (gid_t)-1.  Otherwise, even if uid and gid are unchanged,
>> +     we must ensure that ctime is updated. */
>>     if (uid == ILLEGAL_UID)
>>       uid = old_uid;
>> +  else
>> +    noop = false;
>>     if (gid == ILLEGAL_GID)
>>       gid = old_gid;
>> -  if (uid == old_uid && gid == old_gid)
> 
> Basically ok, but why not just
> 
>       if (uid == ILLEGAL_UID && gid == ILLEGAL_GID)
> 
> instead of the noop var?

I went back and forth on that.  Following your suggestion, the code looks like this:

   if (uid == ILLEGAL_UID && gid == ILLEGAL_GID)
     {
       ret = 0;
       goto out;
     }
   if (uid == ILLEGAL_UID)
     uid = old_uid;
   if (gid == ILLEGAL_GID)
     gid = old_gid;

I was trying to avoid checking uid == ILLEGAL_UID and gid == ILLEGAL_GID twice. 
  But on second thought, it's probably silly to worry about that.  The code is 
cleaner without the noop variable.

Thanks.

Ken

Re: [PATCH] Cygwin: chown: make sure ctime gets updated when necessary

[Corinna Vinschen]

On Jan 25 14:16, Ken Brown via Cygwin-patches wrote:
> On 1/25/2021 1:57 PM, Corinna Vinschen via Cygwin-patches wrote:
> > On Jan 25 12:24, Ken Brown via Cygwin-patches wrote:
> > > Following POSIX, ensure that ctime is updated if chown succeeds,
> > > unless the new owner is specified as (uid_t)-1 and the new group is
> > > specified as (gid_t)-1.  Previously, ctime was unchanged whenever the
> > > owner and group were both unchanged.
> > > 
> > > Aside from POSIX compliance, this fix makes gnulib report that chown
> > > works on Cygwin.  This improves the efficiency of packages like GNU
> > > tar that use gnulib's chown module.  Previously such packages would
> > > use a gnulib replacement for chown on Cygwin.
> > > ---
> > >   winsup/cygwin/fhandler_disk_file.cc | 10 +++++++++-
> > >   1 file changed, 9 insertions(+), 1 deletion(-)
> > > 
> > > diff --git a/winsup/cygwin/fhandler_disk_file.cc b/winsup/cygwin/fhandler_disk_file.cc
> > > index 07f9c513a..72d259579 100644
> > > --- a/winsup/cygwin/fhandler_disk_file.cc
> > > +++ b/winsup/cygwin/fhandler_disk_file.cc
> > > @@ -863,6 +863,7 @@ fhandler_disk_file::fchown (uid_t uid, gid_t gid)
> > >     tmp_pathbuf tp;
> > >     aclent_t *aclp;
> > >     int nentries;
> > > +  bool noop = true;
> > >     if (!pc.has_acls ())
> > >       {
> > > @@ -887,11 +888,18 @@ fhandler_disk_file::fchown (uid_t uid, gid_t gid)
> > >   				    aclp, MAX_ACL_ENTRIES)) < 0)
> > >       goto out;
> > > +  /* According to POSIX, chown can be a no-op if uid is (uid_t)-1 and
> > > +     gid is (gid_t)-1.  Otherwise, even if uid and gid are unchanged,
> > > +     we must ensure that ctime is updated. */
> > >     if (uid == ILLEGAL_UID)
> > >       uid = old_uid;
> > > +  else
> > > +    noop = false;
> > >     if (gid == ILLEGAL_GID)
> > >       gid = old_gid;
> > > -  if (uid == old_uid && gid == old_gid)
> > 
> > Basically ok, but why not just
> > 
> >       if (uid == ILLEGAL_UID && gid == ILLEGAL_GID)
> > 
> > instead of the noop var?
> 
> I went back and forth on that.  Following your suggestion, the code looks like this:
> 
>   if (uid == ILLEGAL_UID && gid == ILLEGAL_GID)
>     {
>       ret = 0;
>       goto out;
>     }
>   if (uid == ILLEGAL_UID)
>     uid = old_uid;
>   if (gid == ILLEGAL_GID)
>     gid = old_gid;
> 
> I was trying to avoid checking uid == ILLEGAL_UID and gid == ILLEGAL_GID
> twice.  But on second thought, it's probably silly to worry about that.  The
> code is cleaner without the noop variable.

Both is ok with me, whatever you think more spiffy here.


Thanks,
Corinna