reduce vector table size

[nerdralph]

Since there`s gaps in the vector table, you can put some of the startup code there to save flash.
https://github.com/nerdralph/stm32f0-blink/blob/master/src/vector_table.S#L29

[deividAlfa]

Why so much hassle to save few dozens of flash bytes?

[nerdralph]

Why so much hassle to save few dozens of flash bytes?

I didn't think copying my vector table asm code would be a hassle.

[deividAlfa]

I can see the bss zeroing, but where's the data initialization?

[IOsetting]

Hi Ralph, is it feasible to add a new startup file e.g. "startup_hk32f030mf4p6_compact.s"? Then user can choose different options in Makefile for compiling.

[nerdralph]

Hi Ralph, is it feasible to add a new startup file e.g. "startup_hk32f030mf4p6_compact.s"? Then user can choose different options in Makefile for compiling.

I suppose, but what would be the reason not to use a smaller init all the time? I think the effort of adding makefile options would be better spent on making a single optimized init.

[IOsetting]

Hi Ralph,
I got your point. In a fixed scenario, a smaller init would be preferred, users will benefit from these hacks.
This is a project template for a rarely known chip, I feel that at this phase, a standard startup file would be best for most users, readable, reliable.
Changing the startup file only need to edit one line in Makefile so no effort needed, if most users feel the smaller init is stable and good in most cases, we can set it as default.

[nerdralph]

That's a reasonable choice to make for someone who doesn't have as much confidence in my code as I do.

…

On Wed, Jan 4, 2023 at 8:45 AM IOsetting ***@***.***> wrote: Hi Ralph, I got your point. In a fixed scenario, a smaller init would be preferred, users will benefit from these hacks. This is a project template for a rarely known chip, I feel that at this phase, a standard startup file would be best for most users, readable, trustable. Changing the startup file only need to edit one line in Makefile so no effort needed, if most users feel the smaller init is stable and good in most cases, we can set it as default. — Reply to this email directly, view it on GitHub <#3 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABKNZ6QGGDVYRRVLNA6YFJDWQVWGZANCNFSM6AAAAAATPECKSI> . You are receiving this because you authored the thread.Message ID: ***@***.***>