Re: [PATCH 1/2] include: add new file for cache-related definitions

2014-01-08 Thread Thomas Schwinge
On Thu, 09 Jan 2014 01:30:42 +, Ivan Shmakov wrote: > > Justus Winter <4win...@informatik.uni-hamburg.de> writes: > > > Date: Tue, 7 Jan 2014 15:51:27 +0100 > > […] > > > --- /dev/null > > +++ b/include/cache.h > > @@ -0,0 +1,25 @@ > > +/* > > + * Copyright (c) 2013 Free Software

Re: [PATCH 1/2] include: add new file for cache-related definitions

2014-01-08 Thread Ivan Shmakov
> Justus Winter <4win...@informatik.uni-hamburg.de> writes: > Date: Tue, 7 Jan 2014 15:51:27 +0100 […] > --- /dev/null > +++ b/include/cache.h > @@ -0,0 +1,25 @@ > +/* > + * Copyright (c) 2013 Free Software Foundation. … Which makes me wonder if the boilerplate has to be upda

Re: [PATCH 2/2] Align kmem_cache objects using __cacheline_aligned

2014-01-08 Thread Richard Braun
On Thu, Jan 09, 2014 at 12:40:49AM +0100, Samuel Thibault wrote: > Samuel Thibault, le Thu 09 Jan 2014 00:39:30 +0100, a écrit : > > Mmm, perhaps it would be more useful to define a typedef with the > > attribute, and use it instead of struct kmem_cache? > > Mmm, from slab.h it would seem that you

Re: [PATCH 2/2] Align kmem_cache objects using __cacheline_aligned

2014-01-08 Thread Samuel Thibault
Samuel Thibault, le Thu 09 Jan 2014 00:39:30 +0100, a écrit : > Justus Winter, le Tue 07 Jan 2014 15:51:28 +0100, a écrit : > > * device/dev_lookup.c: Align kmem_cache objects using __cacheline_aligned. > > Mmm, perhaps it would be more useful to define a typedef with the > attribute, and use it i

Re: [PATCH 1/2] include: add new file for cache-related definitions

2014-01-08 Thread Samuel Thibault
Justus Winter, le Tue 07 Jan 2014 15:51:27 +0100, a écrit : > * include/cache.h (__cacheline_aligned): This macro can be used to > align statically allocated objects so that they start at a cache line. Ack. You could already use it in kern/slab.h > --- > include/cache.h | 25

Re: [PATCH 2/2] Align kmem_cache objects using __cacheline_aligned

2014-01-08 Thread Samuel Thibault
Justus Winter, le Tue 07 Jan 2014 15:51:28 +0100, a écrit : > * device/dev_lookup.c: Align kmem_cache objects using __cacheline_aligned. Mmm, perhaps it would be more useful to define a typedef with the attribute, and use it instead of struct kmem_cache? Samuel

Re: [RFC] kern: simple futex for gnumach (version 10)

2014-01-08 Thread Samuel Thibault
Richard Braun, le Mon 06 Jan 2014 15:00:35 +0100, a écrit : > After some thinking, this would actually not work, as physical pages can > be paged out. Augh, I hadn't thought about that indeed. Samuel

Re: [PATCH] Make STARTUP_WITH_SHELL a runtime toggle -- add new "set/show startup-with-shell" option.

2014-01-08 Thread Thomas Schwinge
Hi! On Thu, 24 Oct 2013 16:17:21 +0100, Pedro Alves wrote: > Here's what I pushed ..., and what made the Hurd port pretty unhappy. ;-) > gdb/ > 2013-10-24 Pedro Alves > > * NEWS (New options): Mention set/show startup-with-shell. > * config/alpha/nm-osf3.h (START_INFERIOR_TRAPS

Re: [RFC] kern: simple futex for gnumach (version 11)

2014-01-08 Thread Richard Braun
On Wed, Jan 08, 2014 at 08:43:28PM +0100, Marin Ramesa wrote: > First threads were blocked and awakened when called from userspace. :) > > Timed waits work as expected. Private and shared futexes always block threads > when called from futex_wait(). > > This is now ready for test. What needs to

[RFC] kern: simple futex for gnumach (version 11)

2014-01-08 Thread Marin Ramesa
First threads were blocked and awakened when called from userspace. :) Timed waits work as expected. Private and shared futexes always block threads when called from futex_wait(). This is now ready for test. What needs to be tested is: 1. If sync circle actually wakes the threads when sent a wa