Message ID | 20100618003306.GA3463@kryten (mailing list archive) |
---|---|
State | Accepted, archived |
Commit | 9ae7da5f64bf6a5f427ebdc24b651b3148f5b8d9 |
Delegated to: | Benjamin Herrenschmidt |
Headers | show |
On Fri, Jun 18, 2010 at 10:33:06AM +1000, Anton Blanchard wrote: > > If we configure with CONFIG_SMP=n or set NR_CPUS less than the number of > SMT threads we will set the max cores property to 0 in the > ibm,client-architecture-support structure. On new versions of firmware that > understand this property it obliges and terminates our partition. > > Use DIV_ROUND_UP so we handle not only the CONFIG_SMP=n case but also the > case where NR_CPUS isn't a multiple of the number of SMT threads. Thank you, Anton!!! Acked-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com> (Will test as soon as the system that hit this becomes available.) > Signed-off-by: Anton Blanchard <anton@samba.org> > --- > > Index: linux-2.6/arch/powerpc/kernel/prom_init.c > =================================================================== > --- linux-2.6.orig/arch/powerpc/kernel/prom_init.c 2010-06-17 09:08:20.000000000 +1000 > +++ linux-2.6/arch/powerpc/kernel/prom_init.c 2010-06-17 09:10:02.000000000 +1000 > @@ -872,7 +872,7 @@ static void __init prom_send_capabilitie > "ibm_architecture_vec structure inconsistent: 0x%x !\n", > *cores); > } else { > - *cores = NR_CPUS / prom_count_smt_threads(); > + *cores = DIV_ROUND_UP(NR_CPUS, prom_count_smt_threads()); > prom_printf("Max number of cores passed to firmware: 0x%x\n", > (unsigned long)*cores); > }
Index: linux-2.6/arch/powerpc/kernel/prom_init.c =================================================================== --- linux-2.6.orig/arch/powerpc/kernel/prom_init.c 2010-06-17 09:08:20.000000000 +1000 +++ linux-2.6/arch/powerpc/kernel/prom_init.c 2010-06-17 09:10:02.000000000 +1000 @@ -872,7 +872,7 @@ static void __init prom_send_capabilitie "ibm_architecture_vec structure inconsistent: 0x%x !\n", *cores); } else { - *cores = NR_CPUS / prom_count_smt_threads(); + *cores = DIV_ROUND_UP(NR_CPUS, prom_count_smt_threads()); prom_printf("Max number of cores passed to firmware: 0x%x\n", (unsigned long)*cores); }
If we configure with CONFIG_SMP=n or set NR_CPUS less than the number of SMT threads we will set the max cores property to 0 in the ibm,client-architecture-support structure. On new versions of firmware that understand this property it obliges and terminates our partition. Use DIV_ROUND_UP so we handle not only the CONFIG_SMP=n case but also the case where NR_CPUS isn't a multiple of the number of SMT threads. Signed-off-by: Anton Blanchard <anton@samba.org> ---