Re: [PATCH 7/8] hw/m68k/mcf5206.c: Switch to transaction-based ptimer API

[Thomas Huth]

Am Sat, 19 Oct 2019 13:10:27 +0200
schrieb Thomas Huth <address@hidden>:

> Am Sat, 19 Oct 2019 12:48:59 +0200
> schrieb Thomas Huth <address@hidden>:
> 
> > Am Thu, 17 Oct 2019 14:29:04 +0100
> > schrieb Peter Maydell <address@hidden>:
> > 
> > > Switch the mcf5206 code away from bottom-half based ptimers to
> > > the new transaction-based ptimer API.  This just requires adding
> > > begin/commit calls around the various places that modify the
> > > ptimer state, and using the new ptimer_init() function to create
> > > the timer.
> > > 
> > > Signed-off-by: Peter Maydell <address@hidden>
> > > ---
> > >  hw/m68k/mcf5206.c | 9 +++++----
> > >  1 file changed, 5 insertions(+), 4 deletions(-)  
> > 
> > After applying the patch, I now get an assertion:
> > 
> > $ qemu-system-m68k -nographic -kernel
> > ~/tmp/images/image-an5206-big-20000706.bin -M an5206
> > 
> > uClinux/COLDFIRE(m5206)
> > COLDFIRE port done by Greg Ungerer, address@hidden
> > Flat model support (C) 1998,1999 Kenneth Albanowski, D. Jeff Dionne
> > KERNEL -> TEXT=0x010000-0x077cb8 DATA=0x077cb8-0x08b0d0
> > BSS=0x08b0d0-0x0a2d58 KERNEL -> ROMFS=0x0a2d58-0x183b10
> > MEM=0x183b10-0x1fff000 STACK=0x1fff000-0x2000000 qemu-system-m68k:
> > hw/core/ptimer.c:410: ptimer_transaction_begin: Assertion
> > `!s->in_transaction || !s->callback' failed.
> > 
> > Looks like something is still wrong here...

Looking at the code a little bit, I think you missed the early return
in m5206_timer_recalibrate() :

    if ((s->tmr & TMR_RST) == 0)
        return;

That needs a ptimer_transaction_commit() now, too.

 Thomas