[PATCH 1/4] powerpc/tm: Add commandline option to disable hardware transactional memory

[Breno Leitao]

On Sat, Oct 21, 2017 at 11:58:47AM +1100, Michael Neuling wrote:
> On Fri, 2017-10-20 at 09:45 -0200, Breno Leitao wrote:
> > Mikey, Cyril,
> > 
> > On Thu, Oct 12, 2017 at 09:17:16PM +1100, Michael Ellerman wrote:
> > > From: Cyril Bur <cyrilbur at gmail.com>
> > > 
> > > Currently the kernel relies on firmware to inform it whether or not the
> > > CPU supports HTM and as long as the kernel was built with
> > > CONFIG_PPC_TRANSACTIONAL_MEM=y then it will allow userspace to make
> > > use of the facility.
> > > 
> > > There may be situations where it would be advantageous for the kernel
> > > to not allow userspace to use HTM, currently the only way to achieve
> > > this is to recompile the kernel with CONFIG_PPC_TRANSACTIONAL_MEM=n.
> > > 
> > > This patch adds a simple commandline option so that HTM can be
> > > disabled at boot time.
> > > 
> > > Signed-off-by: Cyril Bur <cyrilbur at gmail.com>
> > > [mpe: Simplify to a bool, move to prom.c, put doco in the right place.
> > >  Always disable, regardless of initial state, to avoid user confusion.]
> > > Signed-off-by: Michael Ellerman <mpe at ellerman.id.au>
> > > ---
> > >  Documentation/admin-guide/kernel-parameters.txt |  4 ++++
> > >  arch/powerpc/kernel/prom.c                      | 31
> > > +++++++++++++++++++++++++
> > >  2 files changed, 35 insertions(+)
> > > 
> > > diff --git a/Documentation/admin-guide/kernel-parameters.txt
> > > b/Documentation/admin-guide/kernel-parameters.txt
> > > index 05496622b4ef..ef03e6e16bdb 100644
> > > --- a/Documentation/admin-guide/kernel-parameters.txt
> > > +++ b/Documentation/admin-guide/kernel-parameters.txt
> > > @@ -3185,6 +3185,10 @@
> > >  			allowed (eg
> > > kernel_enable_fpu()/kernel_disable_fpu()).
> > >  			There is some performance impact when enabling
> > > this.
> > >  
> > > +	ppc_tm=		[PPC]
> > > +			Format: {"off"}
> > > +			Disable Hardware Transactional Memory
> > > +
> > >  	print-fatal-signals=
> > >  			[KNL] debug: print fatal signals
> > >  
> > > diff --git a/arch/powerpc/kernel/prom.c b/arch/powerpc/kernel/prom.c
> > > index f83056297441..d9bd6555f980 100644
> > > --- a/arch/powerpc/kernel/prom.c
> > > +++ b/arch/powerpc/kernel/prom.c
> > > @@ -658,6 +658,35 @@ static void __init early_reserve_mem(void)
> > >  #endif
> > >  }
> > >  
> > > +#ifdef CONFIG_PPC_TRANSACTIONAL_MEM
> > > +static bool tm_disabled __initdata;
> > 
> > I think the name 'tm_disabled' might cause more confusion on the TM
> > code. Mainly because we already have tm_enable() and tm_enabled()
> > functions which are related to the MSR register and TM bit, and, with
> > your new variable, tm_enabled() and tm_disabled are not going to be
> > exclusionary. Neither tm_enable() with be able to toggle the tm_disabled
> > value.
> 
> Got a proposal for better names?

That is the hardest part, but I thought about something as:

 * tm_disabled_on_boot
 * tm_off
 * tm_explicit_disabled
 * tm_feature_disabled
 * tm_enablement_disabled

I think these names, although a bit longer, might avoid the confusion
with tm_enable/tm_enabled nomenclature.