FixSD patch
David Feugey (2125) 2709 posts |
Hi all. I have a problem with some of my old Pi. The problem is SDFS. It detects SD card changes. If the SDFS module did not try to detect them, it would work. I have not a lot of time for this. So, if someone can remove the SD Card detection on SDFS and compile a ROM One with a SD card detector broken or not :) |
Jon Abbott (1421) 2651 posts |
Have you updated the firmware? I believe recent builds need a fairly up to date firmware to match. |
David Feugey (2125) 2709 posts |
It’s really a SDFS feature, not a firmware bug. RISC OS boots well, but when SDFS reloads during boot, it activates the ‘detect if card is removed’ feature. As it does not work (hardware problem), the system stops booting with a “disc not found error”. A problem discussed here before. Under Linux, there is no problem, as the system does not try to check if the SD card is changed while the OS is running. And so the faulty Pi under RISC OS works perfectly under Linux. The idea would be to have an alternative ROM with this code removed (then, you won’t be able to remove the SD card safely while using this version of RISC OS). Or a kernel boot option. I know that some old Pi B have the problem (I have some). And I suspect that there will be more of them. Sometimes, the connector breaks just when you insert a new card. Nota: I use RC14 on RPB 1 Model B. Standard image, with standard firmware. |
Jon Abbott (1421) 2651 posts |
I think I understand you now. Some of your Pi’s have a faulty SD card slot (the SD sense doesn’t work), so fail to boot due to the way RISCOS handles the issue. I’m assuming they’re out of warranty? |
David Feugey (2125) 2709 posts |
It was never covered by warranty, since it’s not a problem under normal (IE Raspbian) use. |
Jon Abbott (1421) 2651 posts |
I don’t think it’s a feature, sounds like a fault to me as the SD sense should work. I have every model of Pi ever released and haven’t seen this issue on any of them. From the Pi B+, all Pi’s were manufactured in Wales. Prior to this, I believe they were manufactured by several facilities in China. Perhaps one batch or manufacturer are slightly different, possibly defective? Unfortunately I don’t have one to hand to look if there’s any manufacture or batch info on them. It’s probably worth updating the Firmware/RISCOS to the latest on a fault one – just to rule them out, then arrange to ship it to one of the OS dev’s, so they can test and confirm it’s not a bug in the OS that’s gone unnoticed until now. EDIT: This might be a common issue, as I’ve just found this on the foundation site:
And regarding confirming it’s a fault on the Pi, the foundation says:
|