Hi David, I think I figured it out: I am running in the non-secure world of TrustZone. Turns out ST missed configuring the IO Port I as non-secure in their STM32U5 port of TrustedFirmware-M (all other IO ports are configured as non-secure). I am compiling right now.
Security sucks :-) Thanks for your help Michael Am Fr., 25. März 2022 um 17:50 Uhr schrieb David Sidrane < david.sidr...@nscdg.com>: > Does the CPU busfault if write the GPIOI MODER? Do the values stick when > written? > > -----Original Message----- > From: Michael Jung <mij...@gmx.net> > Sent: Friday, March 25, 2022 9:36 AM > To: dev <dev@nuttx.apache.org> > Subject: Re: STM32U585: Can't enable peripheral clock for IO port I > (GPIOIEN > in RCC_AHB2ENR1) > > Hi David, > > Thanks for your reply. Actually, I am trying to get the B-U585I-IOT02A's > NOR flash chip running, which is connected via OCTOSPI. I can issue a > single command (READID) to the NOR flash and I get a sane reply, but any > follow-on command will just return nonsense (more specifically, the READID > reply will be returned over and over again). This seemed to point to > problems with the nCS signal, which turned out to be on PI5. I then > realized > via JTAG/SVD that GPIOIEN is not set. Single stepping through the code I > see > that the value 0x802001ff is written to RCC_AHB2ENR1, but reading back the > register returns 0x802000ff. > > The stm32l5, which I based the stm32u5 code on, did not have an IO port I. > I must have missed a piece of logic during porting I guess. > > Bye, > Michael > > Am Fr., 25. März 2022 um 17:22 Uhr schrieb David Sidrane < > david.sidr...@nscdg.com>: > > > Hi Michael, > > > > Odd. Did you check the defines? How are you verifying it? Have you > > JTAG-ed it, could it be bad SVD? Have you just printed the value to be > > set and the result ? > > > > David > > > > -----Original Message----- > > From: Michael Jung <mij...@gmx.net> > > Sent: Friday, March 25, 2022 9:07 AM > > To: dev <dev@nuttx.apache.org> > > Subject: STM32U585: Can't enable peripheral clock for IO port I > > (GPIOIEN in > > RCC_AHB2ENR1) > > > > Hello, > > > > This is more about STM32 than it is about NuttX, but I know there are > > experts here: STM32U585 does have 8 GPIOs on Port I. However, > > setting bit > > 8 (GPIOIEN) in RCC_AHB2ENR1 seems to be ignored (i.e. GPIOIEN remains > > at 0). > > For the other eight IO ports (GPIO[A-H]EN) it works just fine. Any > > ideas anyone? > > > > Thanks! > > Michael > > >