Bricked my 5.28(?) install on my Pi 4, what now?

Hello,
I was trying out Risc OS on my Raspberry Pi 4. I tried installing a few programs via PackMan, including what looked to be a port of Firefox, as well as a program for checking the CPU temperature and clock speed and the like. Firefox seemed to install OK. For the CPU program, I decided it was important enough to get it’s own listing in the “Apps” directory. Big mistake, because when it installed, it gave me a nasty error about how it couldn’t install right. I simply figured the app wouldn’t work and moved on trying others I installed. When I got around to Firefox, my whole system froze and couldn’t be recovered (the mouse moved, but none of the GUI would respond, and hitting f12 to go to a command prompt would not work). This was nothing new, since I had tried other programs that ended up freezing my system in the past, such as the Rise of the Triad shareware release, and I chalked it up to them maybe not being tested for the Pi 4 or whatever. However, when I tried to boot up my system again, it gave a short, two line error message:

Error: DataAbort: Abort on data transfer at &FC1FDAE8 (Error Number &80000002)
*Error: Abort on data transfer at &FC1FDAE8 (Error Number &80000002)
*

On the one hand, it’s sort of lame that installing one package has the potential to brick your entire system. On the other hand, I appreciate that the error messages are so short and concise that I can just type them in by hand after taking a photograph of them on my phone instead of finding some way to copy and paste the whole system log.

Speaking of copying and pasting from the file system, I realized I couldn’t boot into Risc OS after about a few minutes of waiting for something to happen and hoping I didn’t brick it, so I decided to just flash my OS image. One issue, though. I plan on writing an article on this OS on my website, and I was taking notes on the OS, as well as a few screenshots, directly on the OS rather than on a separate computer, since I figured that even if something went wrong, I could just pull the data off the SD card and reformat. Thing is, I’m having trouble figuring out the “Pull the data off the SD card” part now. I plug the SD card into my little old Laptop running Debian Linux, I run lsblk to see what filesystems are available, and I notice two different ones for the SD card. One is easily mountable and is just the boot partition, without any of the actual files I use, and the other, I assume, is the RiscOS partition with my files and programs and the like. I want to mount that partition so that I can grab the little text file I was taking notes on, as well as a screenshot I took earlier, but the issue is, I can’t mount it, the mount command won’t let me. I assume it’s an issue with the filesystem format; I know RiscOS uses SDFS, for instance, and I know it’s obscure enough that Mount won’t just handle it automatically. I can’t find a -t argument for mount to mount the SDFS partition, so I can’t mount it and retrieve my file.
What now? Are my notes just gone? It’s not a huge deal if they are, since they’re just spur of the moment things I rattled off into a text file with the intention of putting into my article later, but I would really like it back. It’s infuriating that I “have” the files but I just really can’t get to them.

EDIT: Also, I hit enter without asking another important question: how can I take more steps to prevent crashes like this in the future? My experience with RiscOS has been pleasant albeit quirky when it works, but it does crash every now and again with certain programs, I have been unable to get a single Packman program working up to this point, and it has some weird behavior (for example, after turning on the computer, neither my mouse nor my keyboard work until plugged in a second time, even if they’re plugged in since boot). However, when I use Risc OS on it’s own, such as by sticking with the vanilla programs, it works much better. Therefore, I can’t help but conclude I’m doing something horribly wrong to my system whenever I attempt to use Packman or something to install some programs. What am I doing wrong? What can I learn in order to use Risc OS without breaking it within a day?

What now? Are my notes just gone?

Write a fresh OS image to another SD, boot from it, Dismount SDFS::0 and then hot swap the SD to access your original files.

My experience with RiscOS has been pleasant albeit quirky when it works, but it does crash every now and again with certain programs, I have been unable to get a single Packman program working up to this point

Unfortunately you’re testing the OS on recent hardware which has only recently had support added for it. Developers have not had time to test/fix their software, so many apps may fail due to CPU instruction differences. The further you go back in hardware, the better the support, so an original Pi for example is one of the most stable experiences.

As Jon said, the Pi4 is most likely to give you issues when trying random bits of software because not everything has been updated for its new cpu. If you have an older Pi try using that – the extra cores and memory are kind of pointless with RISC OS anyway! Good use of an old Pi1 or 2 that isn’t really fast enough for desktop Linux.

Having said that, completely bricking it is odd! Are you sure you don’t have power supply issues? That might also explain the keyb/mouse problems at startup.

If you get a software lockup where everything stops except the mouse pointer, then try pressing Alt-Break. That should pop up a dialogue box offering to kill the errant program.

For mounting on Linux, IIRC the filesystem type to use is adfs.

Thanks for all the advice, guys! Mounting the whole SD card like Bryan Hogan suggested (not one of the sub partitions, but the whole thing) with adfs did the trick (command I used: “sudo mount -t adfs /dev/mmcblk0 /media/sdcard”, yours will vary depending on what device your SD card is mounted as, try to figure it out with lsblk). I’m unable to access any of the files without using Sudo, but I was at least able to use sudo to copy them into my local documents directory to do with them what I will. Also, although I did not take Jon Abbott’s course of action regarding the “hot swap” maneuver, it would have been helpful since that screenshot I took earlier is still in sprite format, and I’m not aware of a way to convert it to an ordinary image without employing the use of a program built into RiscOS. I can still use a USB stick or something to transfer it over later.

I do have an old 2013 Raspberry Pi 1 B way back in the day (not even model B+ or whatever, just B, I’m talking one with only TWO USB ports, the OG), but I’m at college right now and I’m away from home, where it’s kept. I took the Pi 4 with me because I knew it was the most powerful and I just need it to access my homework online and view it using a monitor (we get the homework as PDFs and the like, but we can do it on paper and then submit answers and work online) I am leaning on the far back of my dorm desk. The idea is, instead of occupying valuable desk space with my laptop, I just have a monitor at the back of my dorm desk (and the keyboard and pi are on a separate cubby that rolls out underneath the desk), giving me nearly twice the desk space to work with when doing homework (especially paper-intensive math homework) compared to looking at the homework on my laptop.

I was interested in RISC OS for this purpose because my original plan was to undervolt my Pi 4 becuase I was getting poor performance on Linux and I knew RISC OS runs well. However, I recently purchased a USB C Power Supply just for the Pi 4 alongside a fan, and I installed both earlier today, so I think that I won’t have so many problems if I go back to Raspbian now. On the other hand, it’s come to my attention that Risc OS is less than optimal for the Pi 4. So, I think I am just going to run Linux on my Pi 4 for the time being, and when I can gain access to my older Pi, I will more than happily continue reviewing Risc OS for the purpose of writing that article I was talking about. If I want to write about this OS, I want to make sure I’m using a device with the best support.

Thanks for all the help and tips, guys!

On the other hand, it’s come to my attention that Risc OS is less than optimal for the Pi 4.

I would not say that. I use it for production, with hundred of software, and it’s as stable as the other Pi.

It’s not RiscOS itself that’s not optimal, it’s some of the apps. Each generation of pi uses a new generation of ARM processor, and these each have differences in instruction set. Apps need (at the least) recompiling with a recent compiler, and then should run on all pi’s. However some apps are no longer in development, and in some cases the source code has been lost.

I’ve found the Pi 4B to have a similar level of stability as the 3B 3B+ and other systems such as the ARMx6 Mini.m when running up to date software (i.e. properly ported to 32 bit, and compiled in the last couple of years). However, unlike the previous machines mentioned, the 4B’s A72 processor is far less tolerant of older software which has not been correctly ported and still contains illegal 26 bit code, but with a 32 bit header tacked on. I’ve found quite a few things like this recently, but if you are new to the platform, and don’t have masses of legacy software, it’s going to be less of a problem. Try to make sure you find the latest version of all software, and if there are issues, report them on the forum.

a program for checking the CPU temperature and clock speed

I assume you are referring to CPUClock. To the best of my knowledge this runs without problem on the Pi4. If there are problems then I would like to know where it is failing.

Users may like to know that there is a test version available from my web site that also switches the GPIO attached fan on and off, if your Pi is fitted with such a fan.

Try to make sure you find the latest version of all software, and if there are issues, report them on the forum.

Is there a ‘best way’ to make sure you’ve found not only the latest but, a version that has been tested or certified for Newer hardware? It’s frustrating and off-putting using trial and error to find a program that has been thoroughly tested and maintained.

Perhaps the developer could include in their program description, in the Store, those platforms that have been tested. This would serve to minimize the difficulties new users experience regardless of the 26 bit history.

Seeing “Rpi” in the description could go a long way toward having confidence in the repository contents. A more extravagant approach would be to further filter the listings with an RPi button so you know it pertains to and has been tested on that specific hardware.

Perhaps the developer could include in their program description, in the Store, those platforms that have been tested.

While it is a bit hit and miss with third party websites (I’m aware of several versions of Samba floating around), anything on the store should run on contemporary hardware unless it states otherwise. If it doesn’t, please report it. (bugs happen)

One of the problems is that some developers have a limited number of machines. Take me, for example. An emulated RiscPC, a Pi2 (ARMv7), and a Beagle xM.
I cannot categorically state that my Manga works on a Pi400. I assume it is enough like a Pi that it will work, but I don’t have a 400 and nobody with one has told me. Maybe those with don’t read manga? Who knows.
Plus, it works on one person’s ARMX6, but not another. Cause currently unknown. Yes, it’s frustrating.

This is also not counting changes in the OS that could, depending on the program, change things. I mentioned the other day about spurious “Bad request” errors on opening USB endpoints as files. 5.23 works flawlessly. 5.25 and later, does not.

So, really, what would need to be mentioned is “this specific version of the program on this specific OS version, on this specific version of this specific device”. That’s unwieldy.
Better, I think, to report back problems encountered. Nobody here writes software to laugh at people’s attempts to run it. This RISC OS market is tiny, so writing software is more of a passion than a way to make income. Thus, if something doesn’t work, we’re more likely to be interested in the problem than random developers who have marketing between them and the end users. Whether it’s seen as a moral problem or a technical one, people around these parts are likely to want to know why. Why is it not working? What can be done to fix it?

The benefit of a small marketplace, we take pride in the things we create, and we talk directly to the end users. This isn’t some kind of “contact customer support, get the brush off” situation.

If you find something on Store that isn’t working, get in touch.

get in touch.

For my part, I make no guarantees to be able to fix something however I will usually take a look (which may require additional information from you) because the point of writing software is that it may be useful to people. If it’s not working, well, then it rather fails at the “be useful” part, doesn’t it? ;-)

Plus, seeing reports like “Xyzzy is crap, it crashes all the time” in public forums before anybody has bothered contacting me is kind of depressing. Yes, it has happened (something didn’t like the StrongARM). One of the reasons I have little interest the in newsgroups.

What you ran into wasn’t avoidable. Firefox only worked on an earlier version of RISC OS. There was a small red triangle, with a question mark, next to firefox. That indicates there might be an issue, reference here . Some packages with a red triangle do work, some don’t, but others need to have their !RunImage file patching with this utility

The default fast RISC OS browser is Netsurf.

If it helps, my Pi4 running RISC OS is rock solid. Also !TechWriter
made writing college papers a dream ;)

I use RISC OS daily, Linux when I need a different set of tools. :)

Without knowing the specific version/bundle of RO that Stanford has installed it’s hard to comment, but one question springs to mind:

Have you looked at the RISC OS Direct bundle?

I suspect that after trying that, with the browsers in the bundle you may lose interest in finding out whether a 15+ year old port of Firefox is worth looking at never mind running.
There is allegedly an update of the Direct bundle in the pipeline.

Note: The current Direct bundle is based on an older OS revision, but can be updated pretty easily from all accounts.

There is allegedly an update of the Direct bundle in the pipeline.

That’s true, but do wait for the next version to be released. An official announcement should be incoming. ;)

That’s true, but do wait for the next version to be released. An official announcement should be incoming. ;)

You’re taking all the fun out of his life.
Wait? Wait? You’re implying he’s grown up (unlike the rest of us)

One of the problems is that some developers have a limited number of machines.

I own: Pi zero, Pi zero W, Pi3b+, Pi4b. If ANY developer wants to include these in testing, it would tickle me pink to help the RiscOS Community.

When I began to bloom as a developer, a hefty lot of my DOS programs were released as shareware and later more aptly named freeware. MSdos, PCdos, DRdos to name a few and each of them were updating with money driven frequency. My work was driven by passion and sometimes received overly critical negative reviews. I came to regard these as my most valuable feedback once I realized the reveiwer had also shared his passion with my work.

Better, I think, to report back problems encountered.

To leave testing to the end user is indeed unavoidable, however, to leave testing to only the end user should always be avoided.

One of the reasons I have little interest the in newsgroups.

It’s a fact of life, unfortunately, that ‘no good deed goes unpunished’. Good recognition is the blue ribbon, everything else is just thanks for your participation.

Take a break, when you come back, I’ll give you a high five!

sometimes received overly critical negative reviews. I came to regard these as my most valuable feedback once I realized the reveiwer had also shared his passion with my work.

Critical negative reviews are not the problem. If I screwed up, I screwed up. Or if my way of thinking clashes with somebody else’s way of thinking (let’s just point at “brace style” and Hungarian notation and leave it at that).

If the criticism is justified… sure, I’ll be annoyed, but I’ll get over it. If it’s justified it is justified.

But people slagging me off offensively without even bothering to try to contact me first? That’s just not helpful. But it’s easy to be a dick when you’re looking at a screen and not a person…

however, to leave testing to only the end user should always be avoided.

Uh… where have you been for the last decade? We’re now embracing the lean agile waterfall devops (and any other buzzwords I’ve missed) which is roughly translated as “bring some barely functional badly designed piece of software to market, fix the bugs when the users find ’em”.
With an extension specially for Google: “and if fixing it is too much of a pain in the ass, retire it to the ever-growing graveyard of abandoned good ideas”. ¹

Take a break, when you come back, I’ll give you a high five!

Oh, I mentioned StrongARM. I took a break. From 2002 until 2009 (was here without a phone line). When I got back online: one of the most interesting people had tied in a rather interesting way, and the landscape had greatly changed. A spot of unpleasantness had come and gone, as had the Iyonix. And that was the end of RISC OS.
Or Was It? ;-)

We’ll have to virtual high-five. Social distancing ².

¹ Word to the wise – never depend upon a Google product. Their flagship things make them money. All the interesting little side projects? They’ll be abandoned, eventually. Use Google to Google “Google Graveyard”.

² 2020 was the year I spent 46 years rehearsing for! 😂

people slagging me off offensively without even bothering to try to contact me first? That’s just not helpful.

No argument here! I haven’t read what it is this refers to but, it sounds as if you’re referring to a rant and not a critical review. Do you really want someone who ‘slags you off’ to contact you?

“bring some barely functional badly designed piece of software to market, fix the bugs when the users find ’em”.

I think this business practice first came to market from Microsoft and it is my best reason for having always looked for an alternative to their products.

retire it to the ever-growing graveyard of abandoned good ideas”.

To the best of my knowledge: ALL of my code now resides there!

From 2002 until 2009 … the landscape had greatly changed

I find this an interesting point of discussion if you’re willing to share it on the forum. I’m not certain it’s related but I have sensed an internal strife within some sectors of the community.
Though I may be misunderstood in my efforts to question the contents therein, I am grateful, if only to be polite, and the least likely to condemn others for their opinions.

Do you really want someone who ‘slags you off’ to contact you?

No, not particularly. But it would have been nice to attempt a communication before the insults, just in case, you know, it’s an old version and their particular gripe has been fixed.

first came to market from Microsoft

Nowadays it is practically the way everything is. Even hardware gizmos aren’t immune to this nonsense. I die a little inside when I hear that somebody had to update the firmware in their lightbulb.
Their lightbulb! It has, literally, one f’ing job. And it needs an update.

Once upon a time, stuff came on ROM or CD and it pretty much needed to be correct. Not perfect, nothing is, but the cost of reissuing fixes for big problems could sink a company.
Now? Thanks to the internet and flash memory, stuff can be updated on the fly. While this is great for supporting enhanced features and additions, it’s also bad because it significantly lowers the bar when it comes to making the software in the first place. It’s okay, bugs can be fixed…later on.

if you’re willing to share it on the forum.

Might be a bit of a touchy subject for some.

Basically, there was this version of the OS.
And there was that version of the OS.
And there was some confusion over who owned what.
And who could do what, with what they thought they owned.
And things got a bit nasty.
Quibbles → Quarrels → Incoherent shoutyness.
People picked sides, nailed their flags to the wall and supported their tribe.
Nothing new there, really, just look at politics on both sides of the ocean.

I missed all of that nonsense. By the time I came back, the other version was all but dead because there clearly wasn’t much of a future in legacy systems, and their attempted 32 bit version was unfinished possibly due to starting too late in the game (I get the impression that Gerph pretty much did that one single-handedly?).

For me, it’s also something of a sadness. That a conversion to 32 bit was done twice. That the two versions developed out of step (leading to the Toolbox fiasco). And that one branch is quite a bit more advanced than the other…and guess which one we’re using.

I guess it would have been much better all around if both sides worked together to further the OS itself, but then it’s always easier to find solutions in hindsight. As to the events themselves, probably best to quote the infamous bumper sticker “shit happens” and leave it at that.

And there was some confusion over who owned what.
And who could do what, with what they thought they owned.

Say no more! Not a surprise to hear that the non-participants bare the punishment.
Money is fine, the greed it spawns can be more destructive than the atom-bomb.
I’ll not judge those trying to get ahead, lest I be judged for falling so far behind.
Though it was already a strong suspicion, thank you for bringing that into the light.

it would have been much better all around if both sides worked together to further the OS itself, but then it’s always easier to find solutions in hindsight.

These are words of wisdom.

Re the differnce between the pi400 “limited” version of the OS and the standard 5.28 download.

And yet the download is a larger file than the vanilla download"
!

In what way does the pi400 OS differ from the standard version?

e.g I could not get !SparkFs working on the pi400, whereas it works fine on the pi4b.

George Pearce

In what way does the pi400 OS differ from the standard version?

A swift visual inspection suggests that the major additions were !Python27 (in Programming) and the Pi400Bot background images (in Documents).

I could not get !SparkFs working on the pi400

What was going wrong? Failure to start up, failure when unpacking a zip file, or something else?

In what way does the pi400 OS differ from the standard version?

It works on the Pi400 :-)

You might be better trying out the RISC OS Direct version, which comes with a bigger bundle of software pre-installed – https://www.riscosdev.com/direct/

Disclaimer – this only came out the other day and I haven’t tried it yet!