STM32 RODATA LENGTH #16411

[pmjdebruijn]

ldscripts: STM32F103RC_SKR_MINI RODATA LENGTH (#16411)

The SKR mini E3 boards use FLASH_EEPROM_EMULATION, which reserves a
4KB section of flash (at the end) which is used the emulate the
presence of an EEPROM on the board.

This does however imply that the maximum allowed firmware size is
decreased by 4KB as well, in addition to the 28KB for the bootloader.

When this 4KB is not subtracted from the RODATA LENGTH, it's possible
to do a successful Marlin firmware build, which works fine on the board,
until someone saves settings, after which the board can exhibit
erratic behavior.

This also decreases RODATA for three additional boards that I have identified
that are very likely affected by this issue:

• SKR E3 DIP
• Fysetc AIO II
• MKS Robin Mini

[sjasonsmith]

You may want to consider whether that 4K should be reduced to 2K. An earlier flash implementation in the STM32F1 HAL only stored data with 50% efficiency, which is no longer the case. That means that in December the effective EEPROM size increased from 2K to 4K. I don't know how much EEPROM space is actually useful.

Of course reducing that allocated size would come with some risk as well. I don't know whether there is any way to verify at build time that the EEPROM size in the files you modify here agree with the sizes in the compiled pins file.

[pmjdebruijn]

My commits fix current build bugs.

I don't intend on changing any of engineering decisions that have been made at this point, as those changes are significantly more involved.

[sjasonsmith]

I don't intend on changing any of engineering decisions that have been made at this point, as those changes are significantly more involved.

That is understandable. I assumed since you were working with these 256K boards you may be excited to get that 2K back.

Perhaps someone else can chime in on whether they think that size change should be made. If so, it could be done outside this PR or an issue created to remind someone to get around to it.

[thinkyhead]

This is ok as a wide safety margin, for now. If there's a way that 2K can be recovered, please submit a followup.

[sjasonsmith]

This is ok as a wide safety margin, for now. If there's a way that 2K can be recovered, please submit a followup.

Prior to filing an issue to change that, I looked at various other boards and it seems 4K is the normal EEPROM size across everything I checked, so I no longer think there is an issue.

[damaltor]

Hi, i would like to dig this up: The SKR Mini E3 v2.0 nw has a real EEPROM on board and no longer needs FLASH_EEPROM_EMULATION. Can the 4kb subtraction be removed safely?

[thisiskeithb]

Hi, i would like to dig this up: The SKR Mini E3 v2.0 nw has a real EEPROM on board and no longer needs FLASH_EEPROM_EMULATION. Can the 4kb subtraction be removed safely?

We'd need to create a new set of environments since all the SKR Mini E3 boards share the same four scripts and only the E3 Mini V2 has a real EEPROM.

For best results getting help with configuration and troubleshooting, please use the following resources:

• MarlinFW.org Marlin Documentation
• RepRap.org Marlin Forum
• Tom's 3D Forums
• Facebook Group "Marlin Firmware"
• Facebook Group "Marlin Firmware for 3D Printers"
• Marlin Configuration on YouTube
• Marlin Discord server. Join link: https://discord.gg/n5NJ59y

[damaltor]

Thank you!
I did remove the 4k and compiled. it works fine so far.
Keith, i read your notes and comments all over the place, you and the others do a great job!

[pmjdebruijn]

@damaltor this might interest you as well: #19427

[damaltor]

That is really helpful. Thank you!