Hi William,
William Schumann wrote:
> Jan,
> The changes appear fine.
>
> I have some reservations about /opt/install-test path name, since it
> seems vague.
The name of target directory complies with filesystem(5) specification
in this point:
...
/usr/sbin
Platform-dependent executables for system administra-
tion, expected to be run only by system administrators.
An approved installation location for bundled Solaris
software. The analogous location for add-on system
software or for applications is /opt/packagename/sbin.
...
I have initially thought about SUNWinstall-test directory, but
as Dave pointed out, IPS packages will drop 'SUNW' soon
and then we would need to rename that directory.
As far as package name itself is concerned, I have tried to
follow current naming scheme used for install packages.
> A 'README' would help with this, and with test program usage
> information, but the README could be added at a later time, after
> careful consideration of its contents.
That is a good point - I think we should create that file
which would contain information about how to use test stuff,
who is the target audience, ...
Since we will need to think more about its content, are you
fine with filing separate bug for this ?
Thank you,
Jan
>
> William
>
> jan damborsky wrote:
>> Hi William,
>>
>> thank you very much for review.
>> I have incorporated all changes you suggest.
>> Could you please take a look at new webrev ?
>>
>> http://cr.opensolaris.org/~dambi/bug-5284
>>
>> Thank you,
>> Jan
>>
>>
>> William Schumann wrote:
>>> Jan,
>>> ti_dm.c:
>>> The usage of macro idm_cyls_to_secs() can result in lost information
>>> when assigning to 64-bit unsigned from 32-bit unsigned parameters,
>>> as is done for assigning to (diskaddr_t) extvtoc.v_part[n].p_start.
>>> Suggest making idm_cyls_to_secs() a function instead so that
>>> typechecking is performed at compile time. Another way would be to
>>> force the size to uint64_t by casting the first parameter either
>>> within the macro or when calling the macro. I think that using
>>> uppercase macro names is in general a better idea so that the
>>> developer is alerted to the differences between functions and macros.
>>>
>>> Makefiles for libti and libtd are generating warnings for multiple
>>> clean targets. Remove clean targets from both Makefiles and add:
>>> CLEANFILES = $(TEST_PROGS)
>>> to utilize the clean targets already provided in
>>> usr/src/lib/Makefile.targ
>>>
>>> In pkginfo.tmpl suggest:
>>> DESC="Test suite for OpenSolaris installation"
>>> NAME="OpenSolaris install test suite"
>>>
>>> /usr/sbin looks like a good destination for test programs.
>>> William
>>> jan damborsky wrote:
>>>> Hi,
>>>>
>>>> based on Dave's comment for bug 5644, I have also created
>>>> new SVR4 package called SUNWinstall-test which is supposed
>>>> to deliver test bits for install technologies.
>>>>
>>>> Right now, it contains Target Discovery and Target Instantiation
>>>> test drivers, it is assumed to be later enriched with test drivers
>>>> for other installer modules (ICT, orchestrator, ...)
>>>>
>>>> I have updated webrev accordingly - it is located at the same place:
>>>>
>>>> http://cr.opensolaris.org/~dambi/bug-5284
>>>>
>>>> I have tested that package successfully installs bits on Sparc
>>>> as well as x86. Test of integration to the nv_osol0904 was done:
>>>>
>>>> $ ssh install-bldsp-11.sfbay
>>>> $ cd
>>>> /net/indiana-build.central/export/home/dambi/ws/install-os-5284/packages/i386/nightly-nd
>>>>
>>>>
>>>> $ /net/paradise.sfbay/export/tools/bin/integrate -l install -r
>>>> nv_osol0904 -C -n SUNWinstall-test
>>>> cannot access parent directories
>>>> ##############################################################################
>>>>
>>>>
>>>>
>>>> (TEST MODE ONLY) Integrating Package 'SUNWinstall-test' to
>>>> nv_osol0904 ...
>>>>
>>>> Sending item information to server for verification ...
>>>>
>>>>
>>>>
>>>> STATUS = PASSED (Test mode only, no item is integrated)
>>>>
>>>> ##############################################################################
>>>>
>>>>
>>>>
>>>> See http://spgweb.sfbay/SPE/SRE/Tools/re_audits.html for audit help.
>>>>
>>>> I have added '-C' option, as SUNWinstall-test is new package,
>>>> entry for it will need to be created.
>>>>
>>>> I am not sure about message "cannot access parent directories" -
>>>> it is displayed only when new package is to be integrated
>>>> ('-C' option is provided). If I test for existing package, that
>>>> message
>>>> is not displayed (results of integration test for all install
>>>> packages attached).
>>>>
>>>>
>>>> Thank you,
>>>> Jan
>>>>
>>>>
>>>> jan damborsky wrote:
>>>>> Hi,
>>>>>
>>>>> could I please ask for reviewing the fix for following bugs ?
>>>>>
>>>>> TI doesn't utilize last cylinder when creating VTOC with default
>>>>> layout
>>>>> http://defect.opensolaris.org/bz/show_bug.cgi?id=5284
>>>>>
>>>>> TD and TI test drivers should be built as part of full build of
>>>>> source gate
>>>>> http://defect.opensolaris.org/bz/show_bug.cgi?id=5644
>>>>>
>>>>>
>>>>> webrev is available at:
>>>>> http://cr.opensolaris.org/~dambi/bug-5284
>>>>>
>>>>> Thank you very much,
>>>>> Jan
>>>>>
>>>>>
>>>>> Modules affected:
>>>>> -----------------
>>>>> * libti, TI & TD Makefiles
>>>>>
>>>>> Testing done:
>>>>> -------------
>>>>> configuration:
>>>>> * HW:
>>>>> - vmware guest (1GB RWM, 650MB RWM) on Linux host
>>>>>
>>>>> * SW:
>>>>> - LiveCD installation based on osol-0811-rc2.iso
>>>>> - LD_PRELOAD set in order to link fixed libti
>>>>> - TI test driver
>>>>>
>>>>> Scenarios tested:
>>>>> -----------------
>>>>> * LiveCD installation of 2008.11 in vmware guest with
>>>>> 1GB RWM, recommended disk size provided for the
>>>>> installation (swap created on ZFS volume), x86
>>>>>
>>>>> * LiveCD installation of 2008.11 in vmware guest with
>>>>> 650MB RWM, recommended disk size provided for the
>>>>> installation (swap created on slice 1), x86
>>>>>
>>>>> * created default VTOC w/o slice 1 dedicated to swap
>>>>> utilizing TI test driver (x86 & Sparc)
>>>>> * created default VTOC w/ slice 1 dedicated to swap
>>>>> utilizing TI test driver (x86 & Sparc)
>>>>>
>>>>>
>>>>> Test results:
>>>>> -------------
>>>>>
>>>>> [1] Without fix
>>>>> - last available cylinder remained unused (x86)
>>>>> - last available cylinder utilized (Sparc)
>>>>>
>>>>> [2] With fix
>>>>> - last available cylinder utilized (x86 & Sparc)
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> caiman-discuss mailing list
>>>>> caiman-discuss at opensolaris.org
>>>>> http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
>>>>
>>>> ------------------------------------------------------------------------
>>>>
>>>>
>>>> _______________________________________________
>>>> caiman-discuss mailing list
>>>> caiman-discuss at opensolaris.org
>>>> http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
>>
