On Sat, 11 Aug 2007 02:38:40 +0200, Segher Boessenkool said:
> >> That means GCC cannot compile Linux; it already optimises
> >> some accesses to scalars to smaller accesses when it knows
> >> it is allowed to. Not often though, since it hardly ever
> >> helps in the cost model it employs.
> >
> >
That means GCC cannot compile Linux; it already optimises
some accesses to scalars to smaller accesses when it knows
it is allowed to. Not often though, since it hardly ever
helps in the cost model it employs.
Please give an example code snippet + gcc version + arch
to back this up.
u
On Sat, Aug 11, 2007 at 02:38:40AM +0200, Segher Boessenkool wrote:
> >>That means GCC cannot compile Linux; it already optimises
> >>some accesses to scalars to smaller accesses when it knows
> >>it is allowed to. Not often though, since it hardly ever
> >>helps in the cost model it employs.
> >
That means GCC cannot compile Linux; it already optimises
some accesses to scalars to smaller accesses when it knows
it is allowed to. Not often though, since it hardly ever
helps in the cost model it employs.
Please give an example code snippet + gcc version + arch
to back this up.
u
On Fri, Aug 10, 2007 at 10:07:27PM +0200, Segher Boessenkool wrote:
>
> That means GCC cannot compile Linux; it already optimises
> some accesses to scalars to smaller accesses when it knows
> it is allowed to. Not often though, since it hardly ever
> helps in the cost model it employs.
Please g
On Fri, Aug 10, 2007 at 03:49:03PM -0400, Chris Snook wrote:
> Paul E. McKenney wrote:
> >On Thu, Aug 09, 2007 at 03:24:40PM -0400, Chris Snook wrote:
> >>Paul E. McKenney wrote:
> >>>On Thu, Aug 09, 2007 at 02:13:52PM -0400, Chris Snook wrote:
> Paul E. McKenney wrote:
> >On Thu, Aug 09, 2
The compiler is within its rights to read a 32-bit quantity 16 bits at
at time, even on a 32-bit machine. I would be glad to help pummel any
compiler writer that pulls such a dirty trick, but the C standard
really
does permit this.
Code all over the kernel assumes that 32-bit reads/writes
are
Paul E. McKenney wrote:
On Thu, Aug 09, 2007 at 03:24:40PM -0400, Chris Snook wrote:
Paul E. McKenney wrote:
On Thu, Aug 09, 2007 at 02:13:52PM -0400, Chris Snook wrote:
Paul E. McKenney wrote:
On Thu, Aug 09, 2007 at 01:14:35PM -0400, Chris Snook wrote:
If you'r
On Fri, Aug 10, 2007 at 11:08:20AM +0200, Andi Kleen wrote:
> On Friday 10 August 2007 10:21:46 Herbert Xu wrote:
> > Paul E. McKenney <[EMAIL PROTECTED]> wrote:
> > >
> > > The compiler is within its rights to read a 32-bit quantity 16 bits at
> > > at time, even on a 32-bit machine. I would be g
On Friday 10 August 2007 10:21:46 Herbert Xu wrote:
> Paul E. McKenney <[EMAIL PROTECTED]> wrote:
> >
> > The compiler is within its rights to read a 32-bit quantity 16 bits at
> > at time, even on a 32-bit machine. I would be glad to help pummel any
> > compiler writer that pulls such a dirty tri
Paul E. McKenney <[EMAIL PROTECTED]> wrote:
>
> The compiler is within its rights to read a 32-bit quantity 16 bits at
> at time, even on a 32-bit machine. I would be glad to help pummel any
> compiler writer that pulls such a dirty trick, but the C standard really
> does permit this.
Code all ov
On Thu, Aug 09, 2007 at 03:24:40PM -0400, Chris Snook wrote:
> Paul E. McKenney wrote:
> >On Thu, Aug 09, 2007 at 02:13:52PM -0400, Chris Snook wrote:
> >>Paul E. McKenney wrote:
> >>>On Thu, Aug 09, 2007 at 01:14:35PM -0400, Chris Snook wrote:
> If you're dependi
On Thu, 9 Aug 2007, Chris Snook wrote:
> Segher Boessenkool wrote:
> > > > The only safe way to get atomic accesses is to write
> > > > assembler code. Are there any downsides to that? I don't
> > > > see any.
> > >
> > > The assumption that aligned word reads and writes are atomic, and that
> >
So, why not use the well-defined alternative?
Because we don't need to, and it hurts performance.
It hurts performance by implementing 32-bit atomic reads in assembler?
No, I misunderstood the question. Implementing 32-bit atomic reads in
assembler is redundant, because any sane compiler, *p
Geert Uytterhoeven wrote:
On Thu, 9 Aug 2007, Chris Snook wrote:
Segher Boessenkool wrote:
The only safe way to get atomic accesses is to write
assembler code. Are there any downsides to that? I don't
see any.
The assumption that aligned word reads and writes are atomic, and that
words are a
Segher Boessenkool wrote:
The compiler is within its rights to read a 32-bit quantity 16 bits at
at time, even on a 32-bit machine. I would be glad to help pummel any
compiler writer that pulls such a dirty trick, but the C standard really
does permit this.
Yes, but we don't write code for the
Paul E. McKenney wrote:
On Thu, Aug 09, 2007 at 02:13:52PM -0400, Chris Snook wrote:
Paul E. McKenney wrote:
On Thu, Aug 09, 2007 at 01:14:35PM -0400, Chris Snook wrote:
If you're depending on volatile writes
being visible to other CPUs, you're screwed either way
If you need to guarantee that the value is written to memory at a
particular time in your execution sequence, you either have to read it
from memory to force the compiler to store it first
That isn't enough. The CPU will happily read the datum back from
its own store queue before it ever hit m
The only safe way to get atomic accesses is to write
assembler code. Are there any downsides to that? I don't
see any.
The assumption that aligned word reads and writes are atomic, and
that words are aligned unless explicitly packed otherwise, is
endemic in the kernel. No sane compiler viol
Segher Boessenkool wrote:
The only safe way to get atomic accesses is to write
assembler code. Are there any downsides to that? I don't
see any.
The assumption that aligned word reads and writes are atomic, and that
words are aligned unless explicitly packed otherwise, is endemic in
the ker
The compiler is within its rights to read a 32-bit quantity 16 bits at
at time, even on a 32-bit machine. I would be glad to help pummel any
compiler writer that pulls such a dirty trick, but the C standard
really
does permit this.
Yes, but we don't write code for these compilers. There are
On Thu, Aug 09, 2007 at 02:13:52PM -0400, Chris Snook wrote:
> Paul E. McKenney wrote:
> >On Thu, Aug 09, 2007 at 01:14:35PM -0400, Chris Snook wrote:
> >>If you're depending on volatile writes
> >>being visible to other CPUs, you're screwed either way, because the
The only safe way to get atomic accesses is to write
assembler code. Are there any downsides to that? I don't
see any.
The assumption that aligned word reads and writes are atomic, and that
words are aligned unless explicitly packed otherwise, is endemic in
the kernel. No sane compiler viol
Paul E. McKenney wrote:
On Thu, Aug 09, 2007 at 01:14:35PM -0400, Chris Snook wrote:
If you're depending on volatile writes
being visible to other CPUs, you're screwed either way, because the CPU can
hold that data in cache as long as it wants before it writes it
On Thu, Aug 09, 2007 at 01:14:35PM -0400, Chris Snook wrote:
> Paul E. McKenney wrote:
> >On Thu, Aug 09, 2007 at 12:36:17PM -0400, Chris Snook wrote:
> >>Paul E. McKenney wrote:
> >>>The compiler is within its rights to read a 32-bit quantity 16 bits at
> >>>at time, even on a 32-bit machine. I w
Paul E. McKenney wrote:
On Thu, Aug 09, 2007 at 12:36:17PM -0400, Chris Snook wrote:
Paul E. McKenney wrote:
The compiler is within its rights to read a 32-bit quantity 16 bits at
at time, even on a 32-bit machine. I would be glad to help pummel any
compiler writer that pulls such a dirty tric
On Thu, Aug 09, 2007 at 12:36:17PM -0400, Chris Snook wrote:
> Paul E. McKenney wrote:
> >The compiler is within its rights to read a 32-bit quantity 16 bits at
> >at time, even on a 32-bit machine. I would be glad to help pummel any
> >compiler writer that pulls such a dirty trick, but the C stan
Paul E. McKenney wrote:
The compiler is within its rights to read a 32-bit quantity 16 bits at
at time, even on a 32-bit machine. I would be glad to help pummel any
compiler writer that pulls such a dirty trick, but the C standard really
does permit this.
Yes, but we don't write code for these
Segher Boessenkool wrote:
We can't have split stores because we don't use atomic64_t on 32-bit
architectures.
That's not true; the compiler is free to split all stores
(and reads) from memory however it wants. It is debatable
whether "volatile" would prevent this as well, certainly
it is unsaf
On Thu, Aug 09, 2007 at 11:24:22AM -0400, Chris Snook wrote:
> Paul E. McKenney wrote:
> >On Thu, Aug 09, 2007 at 10:53:14AM -0400, Chris Snook wrote:
> >>Paul E. McKenney wrote:
> >>>Why not the same access-once semantics for atomic_set() as
> >>>for atomic_read()? As this patch stands, it might
We can't have split stores because we don't use atomic64_t on 32-bit
architectures.
That's not true; the compiler is free to split all stores
(and reads) from memory however it wants. It is debatable
whether "volatile" would prevent this as well, certainly
it is unsafe if you want to be portabl
Paul E. McKenney wrote:
On Thu, Aug 09, 2007 at 10:53:14AM -0400, Chris Snook wrote:
Paul E. McKenney wrote:
Why not the same access-once semantics for atomic_set() as
for atomic_read()? As this patch stands, it might introduce
architecture-specific compiler-induced bugs due to the fact that
a
On Thu, Aug 09, 2007 at 10:53:14AM -0400, Chris Snook wrote:
> Paul E. McKenney wrote:
> >Why not the same access-once semantics for atomic_set() as
> >for atomic_read()? As this patch stands, it might introduce
> >architecture-specific compiler-induced bugs due to the fact that
> >atomic_set() us
Paul E. McKenney wrote:
Why not the same access-once semantics for atomic_set() as
for atomic_read()? As this patch stands, it might introduce
architecture-specific compiler-induced bugs due to the fact that
atomic_set() used to imply volatile behavior but no longer does.
When we make the vola
On Thu, Aug 09, 2007 at 09:24:42AM -0400, Chris Snook wrote:
> From: Chris Snook <[EMAIL PROTECTED]>
>
> Purify volatile use for atomic[64]_t on alpha.
Why not the same access-once semantics for atomic_set() as
for atomic_read()? As this patch stands, it might introduce
architecture-specific com
From: Chris Snook <[EMAIL PROTECTED]>
Purify volatile use for atomic[64]_t on alpha.
Signed-off-by: Chris Snook <[EMAIL PROTECTED]>
--- linux-2.6.23-rc2-orig/include/asm-alpha/atomic.h2007-07-08
19:32:17.0 -0400
+++ linux-2.6.23-rc2/include/asm-alpha/atomic.h 2007-08-09 09:19:00.000
36 matches
Mail list logo
