Thanks, but I seem to remember you wrote earlier that an EXTRA modification was required for firmware 53.05, as executing the mass storage command on that firmware enables access to the partition with the movies from Windows, but it does not show up like a virtual drive. I understood you applied a fix for that too?
The strange thing is that if you send the monitor command wl 801c9a84 0xa in FW 53.05 and press play, use a tools to copy all sectors of the physical disk on a 128KB SD card, it will mount without any problems.
Peter
brite_eye- 11-25-2007
Thanks, but I seem to remember you wrote earlier that an EXTRA modification was required for firmware 53.05, as executing the mass storage command on that firmware enables access to the partition with the movies from Windows, but it does not show up like a virtual drive. I understood you applied a fix for that too?
PD goofed up mass storage on 53.12s - probably accidentally rather than by willful intent. It requires extra software on a 53.12:
forumer.com/viewtopic.php?p=43205#43205" target="_blank">http://camerahacks.10.forumer.com/viewtopic.php?p=43205#43205
Simply searching for code patterns (using instructions without addresses) can quickly reveal the proper offsets for both unlock and mass storage patches. But there may need to be some tweaks for the smaller block flash on earlier models before m230.
Are there any other members capable of identifying the patch locations on a 53.12?
Can you or (anyone else) provide patch locations in a whole image flash dump for 53.05 firmware? It should then be an easy task for several to recompile Duggasco's code.
palbeda- 11-26-2007
Can you or (anyone else) provide patch locations in a whole image flash dump for 53.05 firmware? It should then be an easy task for several to recompile Duggasco's code.
Yes, I expect I can. I noted this weekend most pieces of the puzzle are already in the different postings.
Peter
palbeda- 11-29-2007
OK, at least it became clear to me what the problem is in FW 53.05.
As already suggested by other members on this forum, the problem with mass storage on the FW 53.05 is that the first sector that is shown to Windows is the bootsector, while it should be the partitiontable.
I think formatting the storage from DOS or Linux is just too risky. It is possible that the camera would be "confused" when it finds a partitiontable where it expects a bootrecord and would refuse to startup, I was not really comfortable to try this.
I made a temporary fix (to -*test*-('") this theorie) by creating a partition entry with a diskeditor on position 1BE of the first sector seen by Windows. Told DOS to look for the first sector of the disk, first head, first cyclinder, FAT16 and entered the total number of sectors.
That works! Now the disk pops-up in Windows and I was able to read the movies just fine from windows.
Note that it is not a real fix, if you choose "Format Flash" in OPS, the "partitiontable" is cleared and the disk becomes unreadable again...
Peter
brite_eye- 11-29-2007
So why does it work on 53.14?
Could it be that 53.14 firmware points its mass storage interface to the extended partition table (1 sector before the no_name partition) which contains a single partition table entry pointing to no_name?
palbeda- 12-01-2007
Yes, in the dump I have from fw 53.14 you can see there is indeed a partition tabel one sector before the bootsector, which is clearly missing in FW 53.05. Probably it will work in fw 53.05 if enable mass storage, you create a primary partition and format it with FAT16. But it remains very risky in my opinion....
Peter
brite_eye- 12-01-2007
Yes, in the dump I have from fw 53.14 you can see there is indeed a partition tabel one sector before the bootsector, which is clearly missing in FW 53.05
That is odd, when I look at my complete flash dumps created by using OPS download flash/whole image - both the 53.05 and 53.14 no_name partitions are preceeded by an extended partition with a single partition entry. However I did not check for a valid partition entry nor have I (and apparently no one else) checked to see where actual MIPS code is pointing.
Peter - Please verify that you have "good whole" image dumps and that first partition table at 0x5be does not have its second partition entry pointing to an extended partition table just before no_name on your 53.05.
palbeda- 12-04-2007
Let me tell you what I did.
- I have enabled Mass storage on my 53.05
- Then with a hexeditor I look at the Physical disk
- A physical disk should start with a partition table, a logical disk with a boot record
- If I look at the disk: I clearly see a boot record, which Windows interprets as an empty partition table as the positions that should contain the information about partition is empty
- If I add partition information and define the same sector the position of the bootsector, Windows will see the partition and all video files can be read
I will give the full flashdump an extra look if maybe the sector before contains a partition table. So it should have partition info and the magic number at the end of the sector. If that is the case, PD made a n+1 error, but I seem to remember this was not the case. I will double check.
Peter
P.s. I have the executables with this code for fw 53.05 ready, but I will not use it until I have been able to verify the numbers that were changed for the FW 53.05.
brite_eye- 12-04-2007
After further thought, going through the trouble of creating a massdugg and lockfree for older models 53.05 and 53.12 probably is a waste of time. I suggest those working on that switch to automating a complete firmware upload process that will allow upgrading to the la-*test*-('") firmware for model 230s - currently 53.18 .
I previously documented my process for upgrading a 53.12 to a 53.14:
forumer.com/viewtopic.php?p=42951#42951" target="_blank">http://camerahacks.10.forumer.com/viewtopic.php?p=42951#42951
Forumer™ is Voted #1 Free Forum Hosting provider
Build your own community today with the largest message board hosting company.