Partition Manager
Pages: 1 ... 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29
Jon Abbott (1421) 2652 posts |
I’ve uploaded a revised 0.97 if you’d like to try again. |
Anton Reiser (471) 63 posts |
The revised 0.97 version works. Thanks.
|
Jon Abbott (1421) 2652 posts |
Thanks for retesting, I will however need to see the full debug log to figure out where it went wrong. |
Raik (463) 2061 posts |
Sorry. Easter vacation. We are not at home. I can not test before weekend. :-( |
Anton Reiser (471) 63 posts |
|
Jon Abbott (1421) 2652 posts |
Thanks Anton, I’ll review and try to figure out what the issue is. |
Jon Abbott (1421) 2652 posts |
I’ve made a few tweaks to the 48-bit LBA code for ADFS4 and reuploaded if someone would like to try it on a test drive. I don’t think it will get it working, but there’s some additional debug output that might help. Anton, your debug output is odd. Your primary drive has an MBR partition table, but no partitions and the FileCore middle zone didn’t seem to be where it was expected – I can’t explain why though. Your secondary drive, with the 4K sectors – all the discop’s appeared to fail. Finally, the SD card appears to only have an 8GB Linux partition – it that correct? |
Anton Reiser (471) 63 posts |
Jon, as far as I know there is nothing special with the primary drive. Yes, the secondary is a 4Kn 2 TB drive, FileCore formated. SD card 8GB Linux, correct. With the actual PartMan I get the Error ‘Corruption of stack’ |
Jon Abbott (1421) 2652 posts |
Up until that point it was working correctly. It found “ADFS::TIK” on the primary drive and had started reading the secondary 4K drive. I can’t tell exactly where the stack corruption occurred, but it looks like it was caused shortly after reading LBA 1 from the 4K drive to check for an MBR partition table. I’ll recheck all the code for potential buffer-overruns when reading 4K sectors, as memory corruption is the logical explanation. |
Jon Abbott (1421) 2652 posts |
Anton, I’ve corrected a few potential buffer overruns and uploaded a revised build for testing. |
Anton Reiser (471) 63 posts |
Jon, there is something strange there. Apart from that, the FileCore at the primary drive is now recognised, from the 4Kn not. |
Jon Abbott (1421) 2652 posts |
Thanks for the debug log, I can’t explain the odd behaviour but I can see why it failed to identify Filecore on the 4K drive. I’ll review all the code it called and double-check I’ve not missed something obvious. |
Jon Abbott (1421) 2652 posts |
I didn’t spot anything obvious that might have caused the locks (although I can guarantee its due to a bug in the way BASIC handles LOCAL ERROR ), so I’ve removed the error handler from the ADFS code. I’ll probably need to remove all the LOCAL ERROR handlers to prevent hard locks completely. I have also modified the FileCore identify code to support 4K sectors in the latest build if you’d like to see if it identifies FileCore on the 4K drive. |
nemo (145) 2569 posts |
Jon blamed
The much-missed Steve Drain asked you to clarify what you meant to no effect, so I’m going to do the same. There certainly is a long-standing bug concerning
Nope. Not seeing that, and there’s no link between the ‘LOCALness’ of
Too vague to be helpful, but again I can’t see the link.
Do you mean undimensioned array references e.g. a(), or something else? All array references are
No it doesn’t. But your code can certainly corrupt variables and lock up. And beware of this construction:
That will zero the global Is it possible that |
Anton Reiser (471) 63 posts |
Jon, with your latest version PartMgr (mostly) starts. New observation (not tested yet): closeing a filer window (4Kn disc) after PartMgr is running or (sucessfuly) quit → machine hangs. Tomorrow I will disconnect the 4Kn drive to check if this is realy 4Kn related. |
Jon Abbott (1421) 2652 posts |
That doesn’t sound good. Removing the 4K drive would certainly confirm if the issue is 4K specific or ADFS_IDEUserOp related. From your last debug log, all the reads succeeded so on the face of it, the code appears to be working. I suppose it’s possible that by using ADFS_IDEUserOp to read from the drive its somehow left either ADFS or SATADriver in an odd state. Did it identify the FileCore partition on the 4K drive?
Nemo, you posted your reply in the wrong thread. |
Anton Reiser (471) 63 posts |
The 4Kn drive removed, the machine does not hang anymore. I think, this should be solved before looking into the 4Kn thing. |
Jon Abbott (1421) 2652 posts |
Thanks, it sounds like my use of ADFS_IDEUserOp is somehow messing up ADFS. I’ll comment out the code in PM for the time being and will pull the source code for both ADFS4 and SATADriver and check through the relevant code line-by-line. Once I have some tests, I’ll email you direct. I’ve made a note so you can remove your email from the post. |
Raik (463) 2061 posts |
Have try 0.97. Not sure what “Reserved” means in the Filecorepartition. The debug file you find here. |
Jon Abbott (1421) 2652 posts |
Reserved is a “Protective partition” that covers the FileCore portion of the disc. Thanks for the debug log, it managed to identify SSD_FAT this time on ADFS 0:1 which means the 48-bit LBA reads worked. As it failed to identify ROS_Daten on ADFS 0:2, there’s still some issues with identifying FileCore on 4K drives that I need to resolve. |
Jon Abbott (1421) 2652 posts |
I’ve posted a build with further 4K sector tweaks, If you’d like to give it another go. |
Anton Reiser (471) 63 posts |
The new version works fine here. No side effects as previous noticed. |
Jon Abbott (1421) 2652 posts |
Thanks for confirming. Who wants to be the brave soul that tries to format drives under ADFS4? |
Raik (463) 2061 posts |
Not sure what it means. I can not see “handling differences” for me. Is usable like the other devices.
ADFS4 is not the Problem. My bootdevice is ADFS7…
Have do any write/read actions (round about 20GB). No problems. Thanks a lot. Very nice work. |
Chris Mahoney (1684) 2165 posts |
ADFS4 as in “version 4 of ADFS”, not “ADFS drive number 4” :) |
Pages: 1 ... 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29