# 30th June 2022, 6:00 - 7:00 UTC+0
Attendees: Edward, Thomas, Stefan, Nikolai, Anastasia, Felix
## Decisions Summary
* Raiden_debug_spi: Thomas will do a review of this programmer and add comments to the bug [https://ticket.coreboot.org/issues/358%5D(https://ticket.coreboot.org/issues...) * As a result of in-depth discussion of lspcon_i2c_spi and realtek_mst_i2c_spi a bunch of bugs will be created (TODO for aklm). Some of them have wider scope like infra improvements. * Sb600spi work has started and needs reviewers attention * [https://review.coreboot.org/c/flashrom/+/65237%5D(https://review.coreboot.or...) * [https://review.coreboot.org/c/flashrom/+/65238%5D(https://review.coreboot.or...) * [https://review.coreboot.org/c/flashrom/+/65239%5D(https://review.coreboot.or...)
## Agenda
* [aklm] Bugs for release “document this programmer”. Is there anything left to do, if yes then what is left? * [raiden_debug_spi](https://github.com/flashrom/flashrom/blob/master/raiden_debug_spi.c) [https://ticket.coreboot.org/issues/358%5D(https://ticket.coreboot.org/issues...) (man page and code comments added - [https://review.coreboot.org/c/flashrom/+/62768%5D(https://review.coreboot.or...)) * Docs [https://chromium.googlesource.com/chromiumos/third_party/hdctools/+/master/d...) * Thomas will review raiden_debug and add all comments to the bug * AI(quasisec) work with Tools team to document the hardware schematics to make it easier to find and nativage by putting all the docs in one place. * AI(quasisec) file bug to convert init code so that usb vid:pid pair are in the deventry. Basically FIX the FIXME in [https://github.com/flashrom/flashrom/blob/master/raiden_debug_spi.c#L352%5D(...) and remove lines [https://github.com/flashrom/flashrom/blob/master/raiden_debug_spi.c#L1501%5D...) * Would be nice to consider moving from using packed structs to serialisation functions. * * lspcon_i2c_spi [https://ticket.coreboot.org/issues/360%5D(https://ticket.coreboot.org/issues...) (man page added) * There is nothing to add to the name it seems: no model * Is there a way to know which chip you are talking? * We can try to improve board_enable infra for i2c, and then we can hook i2c programmer. * Board enable is specific to internal programmer. We could make it more generic to support more than internal. So that it can say “this board supports this chip”, or “no lspcon detected”. * Lspcon should query some kind of board enable (not the current one), and then board enable can come back and say “you are safe to run / or not”. * For the release we need to make sure people can’t break their devices easily. * TODO needs a bug1 (upgrade board enable infra), not release blocking * Nik good suggest to have a programmer param of “I_want_a_brick=true” to avoid accidentally picking the wrong programmer. * Until we have the infrastructure for i2c drivers to allow listing drivers. * TODO needs a bug2 for I_want_brick=true, release blocking. Temp solution until bug1 fixed. * Lspcon_i2c_spi_set_mpu_active in init function, is it board specific? If I have another board, can I run the programmer, or does it leave some gpio open? * TODO update lspcon_i2c_spi_write_aai documentation, why it throws an error straight away, why it is not using default aai function. Release blocking. * Can we find out which specific chips lspcon supports? * TODO to add documentation at the top of the file that only very specific chips supported. Document as much as you can on the top. When it is working and when it is not working. Release blocking * [https://github.com/Pulse-Eight/libcec/issues/280%5D(https://github.com/Pulse...) * Where are hardcoded SPI opcodes? * This is a good argument for converting to opaque. * `SWSPI_WDATA_*` defines appear to be JEDEC commands, at least numerically. So a code comment for now until we know more. * TODO needs a bug (converting to opaque), soft blocking release * realtek_mst_i2c_spi [https://ticket.coreboot.org/issues/361%5D(https://ticket.coreboot.org/issues...) (man page added) * ditto I_wanna_brick. * https://review.coreboot.org/c/flashrom/+/65545 * TODO update the comment on `realtek_mst_i2c_spi_toggle_gpio_88_strap` * https://review.coreboot.org/c/flashrom/+/65546 * [https://ticket.coreboot.org/issues/376%5D(https://ticket.coreboot.org/issues...) (it says review mediatek_i2c_spi) * ditto I_wanna_brick. * **_TL;DR All i2c drivers make the assumption that the user knows what they are doing. However this could be a dangerous assumption._** * Immediate fix to have get_param() “I_want_a_brick=true” || bail. * Long term fix, extend board_enable to be more generic. * AI(quasisec): Consolidate usb_device.c / usbdev.c into something more unified. * usb_device is only used by raiden debug * usbdev is only used by [developer_box](https://github.com/flashrom/flashrom/blob/master/developerbox_spi.c) * TODO needs a bug, not release blocking * [quasisec]: Uploaded some patches to clean up sb600spi. Hard to get support directly from AMD however would like to get a heart beat going reviewing the patches. * [https://review.coreboot.org/c/flashrom/+/65237%5D(https://review.coreboot.or...) * [https://review.coreboot.org/c/flashrom/+/65238%5D(https://review.coreboot.or...) * [https://review.coreboot.org/c/flashrom/+/65239%5D(https://review.coreboot.or...)
-
Anastasia Klimchuk