... @sgi.com> wrote: > Subject: [PATCH] x86_64: resize NR_IRQS for large machines > > From: Alan ... > MAX_IO_APICS, use MAX_IO_APICS to set NR_IRQS, otherwise use NR_CPUS. > > Signed-off-by: ... to be a number of other patches that you need to get these ...

linux-kernel - Ingo Molnar - Mar 25 2008 - 12:19

... @sgi.com> wrote: > > > Subject: [PATCH] x86_64: resize NR_IRQS for large machines > > > > From: ... > MAX_IO_APICS, use MAX_IO_APICS to set NR_IRQS, otherwise use NR_CPUS. > > > > Signed-off-by ... > be a number of other patches that you need to get these ...

linux-kernel - Alan Mayer - Mar 25 2008 - 12:24

Hi! > Subject: [PATCH] x86_64: resize NR_IRQS for large machines > > From: Alan ... that are dimensioned > by NR_IRQS get very large, especially the irq_desc ... > MAX_IO_APICS, use MAX_IO_APICS > to set NR_IRQS, otherwise use NR_CPUS. > > Signed-off- ...

linux-kernel - Pavel Machek - Mar 25 2008 - 19:10

... 2008, Pavel Machek wrote: > Hi! > > > Subject: [PATCH] x86_64: resize NR_IRQS for large machines > > > > From: Alan Mayer < ... count is > MAX_IO_APICS, use MAX_IO_APICS > > to set NR_IRQS, otherwise use NR_CPUS. > > > > Signed-off-by: Alan Mayer ...

linux-kernel - Alan Mayer - Mar 26 2008 - 14:49

... @sgi.com> wrote: > On machines with very large numbers of cpus, ... (NR_VECTORS + (32 * NR_CPUS)) > +#else > +#define NR_IRQS (NR_VECTORS + (32 * MAX_IO_APICS)) > +#endif > #define ... current status quo of huge [NR_IRQS] arrays either. Patches that make ...

linux-kernel - Ingo Molnar - Mar 26 2008 - 17:23

... wrote: > > > On machines with very large numbers of cpus, tables ... from the current status quo of huge > [NR_IRQS] arrays either. Patches that make NR_IRQS a ... it has to be something worth fighting for. --ajm Lately it occurs to me, ...

linux-kernel - Alan Mayer - Mar 26 2008 - 17:40

... instead just do something like > > > > unsigned int NR_IRQS __read_mostly; > > > > and then just set it ... for doing it > > the way we're doing it, but quite frankly, I ... on large machines. So the current structure really is not optimal for anyone. ...

linux-kernel - Eric W. Biederman - Apr 12 2008 - 16:12

... and a Kconfig switch > (say CONFIG_ARCH_HAS_DYNAMIC_NR_IRQS) for architectures to toggle. If > the ... arrays that are dimensioned by NR_IRQS. It also initializes these dynamically ... and most drivers to use nr_irqs. This patch will also introduce a ...

linux-kernel - Alan Mayer - Mar 27 2008 - 12:16

... a Kconfig switch > > (say CONFIG_ARCH_HAS_DYNAMIC_NR_IRQS) for architectures to toggle. If > > the ... arrays that are dimensioned by NR_IRQS. It also initializes > these dynamically ... most drivers to use nr_irqs. > This patch will also introduce a ...

linux-kernel - Ingo Molnar - Mar 27 2008 - 12:33

... the current status quo of huge > > [NR_IRQS] arrays either. Patches that make ... and a Kconfig switch (say CONFIG_ARCH_HAS_DYNAMIC_NR_IRQS) for architectures to toggle. If the ... drivers can be converted to nr_irqs. The switch might initially even ...

linux-kernel - Ingo Molnar - Mar 26 2008 - 18:24

... instead just do something like unsigned int NR_IRQS __read_mostly; and then just set it early ... are statically allocated and depend on NR_IRQ's (notably "irq_desc"), but don't you ... * be performance reasons for doing it the way we're doing it, but ...

linux-kernel - Linus Torvalds - Mar 26 2008 - 12:29

... instead just do something like > > unsigned int NR_IRQS __read_mostly; > > and then just set it ... > statically allocated and depend on NR_IRQ's (notably "irq_desc"), but > don't ... be performance reasons for doing it > the way we're doing it, but ...

linux-kernel - Alan Mayer - Mar 26 2008 - 14:55

... substantial amount of code written - while > in a more gradual approach it can be converted one by one perhaps) - and > it's hard for users and fellow kernel hackers. > > Ingo > I think I'll take a crack at it. Doing it in phases means I can ...

linux-kernel - Alan Mayer - Mar 27 2008 - 12:53

... one by one perhaps) - and it's hard for > > users and fellow kernel hackers. > > I ... . Stay tuned. great! If your initial target for this is x86 (which has certainly the ... 'd be glad to host and test your patches in x86.git/latest - even if they ...

linux-kernel - Ingo Molnar - Mar 28 2008 - 05:32