On Sun, 1 Dec 2024 00:09:21 -0600 Ava V <avarn...@hotmail.com> wrote:
Michael Tokarev: I find your reply to be disrespectful.
This is what it is. Your bug report looked disrespectful to me, due an attempt to turn lack of knowledge into a bug in the software. Maybe I should've expressed my feeling more clearly. Sometimes when I see "bug" reports like this, my wish to help can't stand against another part of me which says "wtf". And sometimes this other part wins. Especially when people think their lack of knowledge is "more important" than others - this adds up. That's why your bug report looked disrespectful to me. I could've replied in a less harsh tone, that's for sure - to show better standards in communications. Back to the issue.
I can and did use search engines. My results indicated that more recent versions of QEMU do support 64MiB flash in the virt machine type. Maybe this older version from stable doesn't support it, or maybe my invocation is wrong. However, I did try multiple types of invocation to get the size of the pflash to be 64MiB, which did not work.
I knew nothing about this very issue until this bug report. It took me about 1 minute to do a google search for "QEMU_EFI qemu aarch64" - the first query with keywords from your report - to find countless guides and examples of the usage. For example 2nd hit: https://gist.github.com/oznu/ac9efae7c24fd1f37f1d933254587aa4 or 4th: https://cdn.kernel.org/pub/linux/kernel/people/will/docs/qemu/qemu-arm64-howto.html This shows how pflash should be used with QEMU_EFI.fd. It really expects a file of 64MiB size.
If there is a version discrepancy between this version of qemu-system-aarch64 and qemu-efi-aarch64, then that is a bug because both of these packages came from Debian Stable. They should work together.
If it doesn't work with 64Mib flash image, that's sure a different issue, not the one you're "reporting" here. There should not be a version disparity here.
At the very least, there is a documentation bug, because there is no explanation for how to invoke qemu with this EFI image. In fact, my invocation already shows that I know how to use a search engine, because using a search engine is the only way to figure out how to use this thing.
qemu emulates a ton of various architectures and boards. Each of them has its own nuances. It's just impractical to document them all - it'll be a huge document, difficult to find information in and difficult to maintain. Yes, qemu documentation can be improved - the main docs, see https://www.qemu.org/documentation/. In there, it might be mentioned that (some?) aarch64 boards expects pflash of size 64MiB, exactly like it's mentioned now for riscv64 (this one expects 32MiB though). Will it be easier to find than the two examples I've shown above? I dunno. If you think there can be an improvement, you can file a bug about lack of documentation instead, - that'd receive definitely better initial response for sure. But I'd give the same reasons anyway, - there's just too many architectures and boards and too many nuances out there. Hope this helps. And feel free to ask on qemu-us...@nongnu.org or even qemu-de...@nongnu.org (these details are usually too specific for more general debian-users@ mailing lists). /mjt
