Alexander Eremin wrote:
>> Alexander Eremin wrote:
>>> Hi all,
>>> pls review changes for bug 10098 (usbcopy does not
>> deal with smedia
>>> mounted usb flash properly)
>>>
>>> webrev: http://cr.opensolaris.org/~alhazred/10098/
>>>
>>> Also fixed that after 'format' new Solaris slice
>> was mounted again
>>> with Nautilus popup window.
>>>
>>> Tested on snv125:
>>> # mount -p
>>> ...
>>> /dev/dsk/c2t0d0p0:1 - /media/NONAME pcfs - no
>>>
>> nosuid,hidden,nofoldcase,clamptime,noatime,timezone=-1
>> 0800
>>> /dev/dsk/c2t0d0s0 - /mnt ufs - no
>>> rw,intr,largefiles,logging,xattr,onerror=panic
>>>
>>> # ./usbcopy osol121.usb
>>> Found the following USB devices:
>>> 0: /dev/rdsk/c2t0d0p0 976.5 MB Multi Flash
>> Reader 1.00
>>> Enter the number of your choice: 0
>>>
>>> WARNING: All data on your USB storage will be lost.
>>> Are you sure you want to install to
>>> Multi Flash Reader 1.00, 976 MB at
>> /dev/rdsk/c2t0d0p0 ? (y/n) y
>>> Copying and verifying image to USB device
>>> Finished 889 MB in 280 seconds (3.1MB/s)
>>> 0 block(s) re-written due to verification failure
>>> Installing grub to USB device /dev/rdsk/c2t0d0s0
>>> Completed copy to USB
>>> #
>>>
>>>
>>> Best regrds,
>>> ::alhazred
>>> _______________________________________________
>>> caiman-discuss mailing list
>>> caiman-discuss at opensolaris.org
>>>
>> http://mail.opensolaris.org/mailman/listinfo/caiman-di
>> scuss
>>
>> Alexander,
>>
>> The code changes seem OK to me but it's not clear to
>> me if your test
>> environment actually exhibits the problem reported in
>> bug 10098. Does
>> it? Maybe I just missing something... ? If I am
>> please just let me know
>> and describe what your testing covers.
>>
>> As reported in the bug, it was not always possible to
>> reproduce this
>> problem I think it would be valuable to do some more
>> testing. To ensure
>> this actually fixes the problem I would like to ask
>> that the test output
>> include a reproduction of the problem reported in bug
>> 10098 using the
>> original usbcopy. Then see it resolved using the
>> usbcopy with your
>> proposed changes.
>>
>>
>> Thank you! Joe
>>
>>
>> _______________________________________________
>> caiman-discuss mailing list
>> caiman-discuss at opensolaris.org
>> http://mail.opensolaris.org/mailman/listinfo/caiman-di
>> scuss
>
> Thanks for reviewing Joe,
> here's only output from "difficult" case when smedia has a pcfs and solaris
> partitions and they all was mounted automatically. Also tested with one and
> two pcfs partitions as reported in bug. After test I haven't them mounted.
> # mount -p
> ...
> /dev/dsk/c2t0d0p0 - /media/NO NAME pcfs - no
> nosuid,hidden,nofoldcase,clamptime,noatime,timezone=-1 0800
> # ./usbcopy osol121.usb
> ...
> # mount -p|grep c2t0d0p0
> #
>
> Second umount after 'format' is required because using usbcopy in fresh
> snv125
> I found that newly created partition is mounting again. Number of seconds for
> sleep also found using some tests, lees than 5 is not enough.
>
> Regards,
> Alex
Thanks you for the reply Alex.
One more question/suggestion.
Would it make sense to replace:
175 # umount new slice
176 sleep 5
177 umount -f $s0bdev > /dev/null 2>&1
With something like:
typeset -i umount_loop=15
typeset -i loop_cnt=0
while [[ ${loop_cnt} -lt ${umount_loop} ]]; do
# umount new slice
umount -f $s0bdev > /dev/null 2>&1
if [[ $? -eq 0 ]]; then
break;
fi
# if the umount failed try again in a second
sleep 1
(( loop_cnt=${loop_cnt} + 1 ))
done
I think it might be safer and help avoid having the user wait any longer
than needed.
What do you think?
Joe
