Partition Manager
Pages: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 ... 29
Alan Adams (2486) 1149 posts |
|
Jon Abbott (1421) 2651 posts |
I’ve checked the logs, SDFS is reporting only one drive. What does the following show:
I’m expecting it to say 1, in which case Partition Manager is not going to see the 2nd SD. Is this the ARMX6? If it’s physical, could you confirm Configure has the correct number of drives set – assuming you can set the number of SD? If not, I might have to ignore SDFS_Drives and try for more regardless. |
Bryan (8467) 468 posts |
No. It is light blue.Everything on Partition Manager works except that the FAT 32 partition name is shown as < unknown> in both the graphic and text As I said in my updated post above:- |
Alan Adams (2486) 1149 posts |
It says 1
yes. I am not sure, however there are two SD cards plugged into the processor board. One is standard, one micro. Configure doesn’t have an SD option. *status shows SDFSdrive 1 Free is interesting: The image has drive 1 below drive 0. Note that drive 1 is called System – it contains recovery data, which proved very useful when I inadvertently formatted the boot SSD. Drive 0 says “space on 64GB” and below size: 60GB. Is this just the difference between 10^9 and 2^30 versions of GB? Or does it indicate partitioning? – if so 60GB + 7GB doesn’t fit into 64GB. |
Bryan (8467) 468 posts |
How about Configure, Discs? Does that show Floppy Discs? |
Alan Adams (2486) 1149 posts |
Yes, set to 0 correctly. |
Will Ling (519) 98 posts |
Hi Jon, I’ve just set up my Iyonix so I could try out PartMgr 0.39, and it reports only a single disc, all unallocated. There are two adfs discs, 40 and 80 GB. I’ve put a screen grab of Free for each, PartMgr output, and debug output here. |
Jon Abbott (1421) 2651 posts |
I’ve now checked my Pi and confirm Configure doesn’t let you set the number of SD drives. I’m not sure if that’s an oversight or by design. It does however mean that SDFS::2 won’t show in Partition Manager or HForm in your case. |
Chris Gransden (337) 1207 posts |
Configure is only used to set the ‘traditional’ fixed discs. USB SCSI, SDFS and ADFS drives get picked up automatically so it’s not necessary to specify the number of drives. |
Alan Adams (2486) 1149 posts |
I tried changing SDFSdrive from 1 to 2. It booted in exactly the same way as before, and Partmgr gave exactly the same display. I suspect this parameter is used to identify the drive to boot from if SDFS is set as the startup file system. Om an ARMX6 it starts from SCSI. |
Jon Abbott (1421) 2651 posts |
v0.40 up. I’ve rewritten SCSIFS to use SCSI_Op, so could folk with machines that use PartMan see if it now reads the GPT table and partitions correctly. SDFS now ignores SDFS_Drive and tries for 8 SD, which will hopefully sort the missing SD on machines that support more than one. |
Alan Adams (2486) 1149 posts |
The SDFS change worked. However the two SCSI filecore drives are no longer showing in the lower part of the window. I guess this might just be because it displays a limited number of lines? |
Jon Abbott (1421) 2651 posts |
Could you upload debug/txt please |
David Pitt (3386) 1248 posts |
PartMgr 0.40 does now find the two filecore partitions on the PartMan partitioned pen but names them both as Filecore2, the first should be Filecore1. Report is here |
Bryan (8467) 468 posts |
Hmmmm, Not so sure about v0.40 My SCSI::4 fulldisc filecore on mSATA now shows the missing drivename from the SCSI::0 SDcard. ie two drives wrong instead of one. Debug file uploaded in Tests today @ 17:30 Update 17:45 I have just added my RiscPC disc to this Pi4 system and I now get a BASIC error Number Too Big at line 85 in Volume_Info_FAT |
Jon Abbott (1421) 2651 posts |
I should probably add that SCSI_Op isn’t documented – something I’ve raised many, many times. As a consequence, I’ve been unable to find out how to get the Access Key for drives, so am using MAGIC_ACCESS_KEY from the OS source code. This is certainly not how it should be done. Once I have your debug/txt’s, I can bugfix it! |
Andrew McCarthy (3688) 605 posts |
v.040 Looking good. ;) FAT check, my Transfer partition is reported as W95 FAT32 (LBA) 278 MB under Linux. RISC OS reports is as 265 MB. Partition Manager reports the same size, but thinks its FAT16. Debug txt: Log deleted. |
Bryan (8467) 468 posts |
OK second post in Tests has the debug file when RiscPC disc is added to system and it crashes at line 86. (so,no display) Just to be clear, system currently has 4 discs SD card slot is empty. The Pi4 hardware boots off FAT32 partition found in SDcard in the USB adapter. (seen by RiscOS as Fat32fs::Pi4boot or SCSI::1) SCSI::0 is MBR FAT32 partitioned SD Card in USB adapter |
Julie Stamp (8365) 474 posts |
The device reservation system is described on pages 56 – 59 of the manual I pointed you to. You know what the access key is, because you’ll have set it when you reserve the device. |
Chris Gransden (337) 1207 posts |
Here’s the results with a single USB drive. The whole drive is formatted as filecore with an embedded fat partition. Screenshot from PartMgr 0.14 which picks up the FAT partition. Screenshot from PartMgr 0.39 doesn’t show the FAT partition. Debug file here. Screenshot from PartMgr 0.40 same as 0.39 but doesn’t show the drive details at the bottom. Debug file here. |
Alan Adams (2486) 1149 posts |
Here http://www.adamshome.org.uk/RiscOS/partmgr040-debug.txt. Note that the drives show in the graphical part of the window, but not in the text part at the bottom. |
Doug Webb (190) 1180 posts |
PartMgr 0.40 still only sees one drive on my Iyonix as I still have the 160GB drive that is formatted down to 128GB installed as ADFS::5. Note the Change log in the zip arachive may need chnagings as it jumps from 0.39 to 0.49 for listed chnages :-) |
Doug Webb (190) 1180 posts |
Also with PartMgr 040 ARMX6 SCSI::4 and SCSI::1 are shown as in the screenshot from Chris Gransden and now do not appear in the text either. |
Bryan (8467) 468 posts |
Jon, This program is magic. I am very pleased with it. Even in its current not-quite-finished state, it is letting me see exactly what is on each of my many and various “discs” (SD, mSATA, SSD, MemorySticks, even spinning rust). They have a wide variety of partitions and formats. Risc OS and Windows. I have created a named copy of !PartMgr0-40 (with line 86 commented out) Thank you so much. Keep up the good work. |
Jon Abbott (1421) 2651 posts |
Noted, I’ve grabbed your logs and will investigate.
Yep..total hack on my part and a step backwards, but absolutely necessary to support PartMan. It also sets the framework for how it should bypass partitioning once the Partition bounty completes. It certainly highlights the fact that the bounty must provide a means for physical disk addressing as opposed to partition based addressing. Either there needs to be a MAGIC-NUMBER (which I personally hate!) Or a separate SWI for bypassing *_DiscOp and going physical – which is how I implemented physical/virtual DiscOp in ADFFS. The FAT16 partition on SCSI::5 is being detected as FAT32, which might be down to…
Your log shows the “Transfer” partition as FAT16, it’s even in the FAT header:
This is something I noted might become an issue. I’ve implemented Microsoft’s official FAT detection method where it’s stated that below certain sizes, FAT16 should always be used in preference to FAT32. FAT is a nightmare to detect as the version info in the FAT header isn’t mandatory, so if you’re not using Microsoft’s method you can detect FAT16 as FAT32. I’d say Linux is incorrect in this instance. I think Microsoft got wise to this as Windows 10 now just uses “FAT” and you have to drill down to find out if it’s FAT16/32. I’m contemplating doing likewise.
Thanks. I did read those pages, but the wording on SCSI_Reserve 0 isn’t clear unless you already know “access key” in R8 is something to be made up:
It’s also not suitable for User code as it needs a callback handler for the release, so I will need to implement a helper Module – which isn’t ideal.
Thanks, I’ve grabbed the logs for 0.39 and screenshots and will investigate. The 0.40 log has an access error – but I don’t think I need it. I need to implement the SCSI helper Module before we can progress this one.
Again…I need to implement the SCSI helper Module to resolve this.
Well spotted. |
Pages: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 ... 29