Hi Ginnie,

Virginia Wray wrote:
> Hi Jan -
>
> Thank you for the code review. I've incorporated your suggestions, as 
> well as Dave's.
> I've generated another webrev. Could you take a look and see if I've 
> addressed your comments?

Yes - this is what I had in my mind :-)
The changes look good (modulo Dave's comments).

> I checked with Mary, and the test she ran on Solaris 10 was with a ufs 
> file system. I should have
> mentioned in my original code review request the tests I was having 
> her run.

ok - that is fine. Thank you for clarifying this.

Jan

>
> Revised webrev:
> http://cr.opensolaris.org/~ginnie/4279-2/
>
> thx,
>
> ginnie
>
>
>
> On 08/20/09 03:25, Jan Damborsky wrote:
>> Hi Ginnie,
>>
>> please see my comments below (only those
>> not already covered by Dave are mentioned).
>>
>> Thank you,
>> Jan
>>
>>
>> td_mg.c
>> -------
>> Since fstyp_ident() allows checking for particular filesystem,
>> you might avoid using strcmp() and simplifying the code
>> as well as making life for fstyp_ident() easier, since it wouldn't
>> need to enumerate trough all known filesystem modules - e.g.
>>
>> 1232         if ((status = fstyp_ident(fstyp_handle, NULL, &fstype)) 
>> == 0) {
>> 1233                 td_debug_print(LS_DBGLVL_INFO, \
>> 1234                         "td_fstyp_check():fstype is %s\n", fstype);
>> 1235                 if ((status = strcmp(fstype, fs)) == 0) {
>> 1236                         fstyp_fini(fstyp_handle);
>> 1237                         (void) close(fd);
>> 1238                         return (status);
>> 1239                 }
>> 1240         }
>>
>> ->
>>
>> if ((status = fstyp_ident(fstyp_handle, fs, &fstype)) == 0) {
>>
>>    /* filesystem matches */
>>
>> } else {
>>
>>    /*
>>     * filesystem is not that type, capture output of
>>     * fstyp_strerror(fstyp_handle, status)
>>     */
>>
>> }
>>
>>
>> Speaking about testing, I can see from bug report that error messages
>> disappeared after fix was applied - this covers 'ZFS scenario'.
>>
>> With respect to 'UFS scenario' (the code path where there is UFS 
>> filesystem
>> present on slice), if it was not already covered, I might recommend 
>> to run the
>> installer on system with Solaris instance on UFS and check that the code
>> path is activated (td_fstyp_check() correctly recognizes UFS filesystem)
>> and installer reports Solaris instance on it.
>>
>>
>>
>> Virginia Wray wrote:
>>> Hi -
>>>
>>> Could I get a code review for the following:
>>> http://defect.opensolaris.org/bz/show_bug.cgi?id=4279
>>>
>>> Webrev:
>>> http://cr.opensolaris.org/~ginnie/4279/
>>>
>>> thx,
>>>
>>
>


Reply via email to