RTEMS | bsps/mips/malta: Six defects of the BSP and the MIPS port (#5742)

[Sebastian Huber (@sebhub)]

Issue created by Sebastian Huber: https://gitlab.rtems.org/rtems/rtos/rtems/-/work_items/5742



The Malta board cannot run the test suite. Six defects stand in the way.

The BSP links the software CPU counter of `bsps/shared/dev/cpucounter/`.
`_CPU_Counter_read()` of that counter increments a static variable per call, so
nothing in it measures time. The clock driver sets
`CLOCK_DRIVER_USE_DUMMY_TIMECOUNTER`, so the uptime advances once per clock
tick and nothing resolves an interval below the tick.

`Clock_driver_support_at_tick()` sets the compare register ahead of the count
register, so every period gains the latency of the interrupt. Measured over 100
ticks, a period of 100000 counts takes 100277.

`bsp_start()` sets the status register to `SR_FR | SR_PE`. `SR_FR` selects the
64 bit floating point registers, while the ELF attributes of every executable
state a CPR1 size of 32. A double loses half of its bits on the way to memory
and back. The same write clears `SR_CU1`, and `SR_PE` is a bit of the R3000
which the MIPS32 status register does not carry.

`mips_interrupt_mask()` returns `0x0000ff00`, which enables line 7 of the
processor. That line belongs to the compare register. Count and compare are
both zero after a reset, so the line is asserted from the first instruction. An
executable which needs no clock driver takes the interrupt, finds no handler
and terminates with `RTEMS_FATAL_SOURCE_SPURIOUS_INTERRUPT`.

The BSP links the stubs of `bsps/mips/shared/irq/irqstubs.c`, so every
interrupt controller directive fails. Its dispatcher prints a line for six of
the eight sources of the processor rather than passing them to the interrupt
manager. Its `tm27.h` is the default header, which raises no interrupt.

The context switch of the MIPS port keeps the coprocessor enable of the
floating point unit as a bit of the task. `setjmp()` of a hard float ABI saves
the callee saved registers of the unit, so a task without
`RTEMS_FLOATING_POINT` takes a coprocessor unusable exception inside
`setjmp()`.

Found while checking whether the Malta board can join a simulator test
programme. This description was created with Claude Code assistance.

-- 
View it on GitLab: https://gitlab.rtems.org/rtems/rtos/rtems/-/work_items/5742
You're receiving this email because of your account on gitlab.rtems.org. Unsubscribe from this thread: https://gitlab.rtems.org/-/sent_notifications/5-3aqcbtaf4w40que919gic7pfp-1d/unsubscribe | Manage all notifications: https://gitlab.rtems.org/-/profile/notifications | Help: https://gitlab.rtems.org/help


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rtems.org/pipermail/bugs/attachments/20260903/4b63440e/attachment.htm>