| Message ID | ed94ad588dd91733178cf505a49b82f4cf031268.1726746877.git.oleksii.kurochko@gmail.com (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
| Series | Move percpu code to common | expand |
Hi, On 19/09/2024 17:59, Oleksii Kurochko wrote: > Introduce PERCPU_SECTION macro which manages: > * Alignment of the section start > * Insertion of per-CPU data sections > * Alignment and start/end markers for per-CPU data > This change simplifies the linker script maintenance and ensures a unified > approach for per-CPU sections across different architectures. > > Refactor the linker scripts for Arm, PPC, and x86 architectures by using > the common macro PERCPU_SECTION defined in xen/xen.lds.h to handle > per-CPU data sections. > > No functional changes. > > Signed-off-by: Oleksii Kurochko <oleksii.kurochko@gmail.com> Acked-by: Julien Grall <jgrall@amazon.com> Cheers, > --- > xen/arch/arm/xen.lds.S | 9 +-------- > xen/arch/ppc/xen.lds.S | 9 +-------- > xen/arch/x86/xen.lds.S | 9 +-------- > xen/include/xen/xen.lds.h | 10 ++++++++++ > 4 files changed, 13 insertions(+), 24 deletions(-) > > diff --git a/xen/arch/arm/xen.lds.S b/xen/arch/arm/xen.lds.S > index bd884664ad..669a882455 100644 > --- a/xen/arch/arm/xen.lds.S > +++ b/xen/arch/arm/xen.lds.S > @@ -198,14 +198,7 @@ SECTIONS > __bss_start = .; > *(.bss.stack_aligned) > *(.bss.page_aligned) > - . = ALIGN(PAGE_SIZE); > - __per_cpu_start = .; > - *(.bss.percpu.page_aligned) > - *(.bss.percpu) > - . = ALIGN(SMP_CACHE_BYTES); > - *(.bss.percpu.read_mostly) > - . = ALIGN(SMP_CACHE_BYTES); > - __per_cpu_data_end = .; > + PERCPU_SECTION > *(.bss .bss.*) > . = ALIGN(POINTER_ALIGN); > __bss_end = .; > diff --git a/xen/arch/ppc/xen.lds.S b/xen/arch/ppc/xen.lds.S > index 38cd857187..0833d80479 100644 > --- a/xen/arch/ppc/xen.lds.S > +++ b/xen/arch/ppc/xen.lds.S > @@ -148,14 +148,7 @@ SECTIONS > __bss_start = .; > *(.bss.stack_aligned) > *(.bss.page_aligned) > - . = ALIGN(PAGE_SIZE); > - __per_cpu_start = .; > - *(.bss.percpu.page_aligned) > - *(.bss.percpu) > - . = ALIGN(SMP_CACHE_BYTES); > - *(.bss.percpu.read_mostly) > - . = ALIGN(SMP_CACHE_BYTES); > - __per_cpu_data_end = .; > + PERCPU_SECTION > *(.bss .bss.*) > . = ALIGN(POINTER_ALIGN); > __bss_end = .; > diff --git a/xen/arch/x86/xen.lds.S b/xen/arch/x86/xen.lds.S > index d48de67cfd..eea8edc02b 100644 > --- a/xen/arch/x86/xen.lds.S > +++ b/xen/arch/x86/xen.lds.S > @@ -321,14 +321,7 @@ SECTIONS > DECL_SECTION(.bss) { > __bss_start = .; > *(.bss.page_aligned*) > - . = ALIGN(PAGE_SIZE); > - __per_cpu_start = .; > - *(.bss.percpu.page_aligned) > - *(.bss.percpu) > - . = ALIGN(SMP_CACHE_BYTES); > - *(.bss.percpu.read_mostly) > - . = ALIGN(SMP_CACHE_BYTES); > - __per_cpu_data_end = .; > + PERCPU_SECTION > *(.bss .bss.*) > . = ALIGN(POINTER_ALIGN); > __bss_end = .; > diff --git a/xen/include/xen/xen.lds.h b/xen/include/xen/xen.lds.h > index a17810bb28..f043c7b6c0 100644 > --- a/xen/include/xen/xen.lds.h > +++ b/xen/include/xen/xen.lds.h > @@ -151,6 +151,16 @@ > #define LOCK_PROFILE_DATA > #endif > > +#define PERCPU_SECTION \ > + . = ALIGN(PAGE_SIZE); \ > + __per_cpu_start = .; \ > + *(.bss.percpu.page_aligned) \ > + *(.bss.percpu) \ > + . = ALIGN(SMP_CACHE_BYTES); \ > + *(.bss.percpu.read_mostly) \ > + . = ALIGN(SMP_CACHE_BYTES); \ > + __per_cpu_data_end = .; \ > + > #ifdef CONFIG_HAS_VPCI > #define VPCI_ARRAY \ > . = ALIGN(POINTER_ALIGN); \
On 19.09.2024 17:59, Oleksii Kurochko wrote: > --- a/xen/arch/x86/xen.lds.S > +++ b/xen/arch/x86/xen.lds.S > @@ -321,14 +321,7 @@ SECTIONS > DECL_SECTION(.bss) { > __bss_start = .; > *(.bss.page_aligned*) > - . = ALIGN(PAGE_SIZE); > - __per_cpu_start = .; > - *(.bss.percpu.page_aligned) > - *(.bss.percpu) > - . = ALIGN(SMP_CACHE_BYTES); > - *(.bss.percpu.read_mostly) > - . = ALIGN(SMP_CACHE_BYTES); > - __per_cpu_data_end = .; > + PERCPU_SECTION > *(.bss .bss.*) > . = ALIGN(POINTER_ALIGN); > __bss_end = .; Like the _SEC in the other patch I question _SECTION here, albeit for a different reason: This is no separate output section, and it's more than one kind of input ones. Perhaps PERCPU_DATA? With that Acked-by: Jan Beulich <jbeulich@suse.com> Jan
On Mon, 2024-09-23 at 12:56 +0200, Jan Beulich wrote: > On 19.09.2024 17:59, Oleksii Kurochko wrote: > > --- a/xen/arch/x86/xen.lds.S > > +++ b/xen/arch/x86/xen.lds.S > > @@ -321,14 +321,7 @@ SECTIONS > > DECL_SECTION(.bss) { > > __bss_start = .; > > *(.bss.page_aligned*) > > - . = ALIGN(PAGE_SIZE); > > - __per_cpu_start = .; > > - *(.bss.percpu.page_aligned) > > - *(.bss.percpu) > > - . = ALIGN(SMP_CACHE_BYTES); > > - *(.bss.percpu.read_mostly) > > - . = ALIGN(SMP_CACHE_BYTES); > > - __per_cpu_data_end = .; > > + PERCPU_SECTION > > *(.bss .bss.*) > > . = ALIGN(POINTER_ALIGN); > > __bss_end = .; > > Like the _SEC in the other patch I question _SECTION here, albeit for > a different > reason: This is no separate output section, and it's more than one > kind of input > ones. Perhaps PERCPU_DATA? With that > Acked-by: Jan Beulich <jbeulich@suse.com> Sure, I will drop _SECTION here too. Thanks. ~ Oleksii
On 23/09/2024 12:48 pm, oleksii.kurochko@gmail.com wrote: > On Mon, 2024-09-23 at 12:56 +0200, Jan Beulich wrote: >> On 19.09.2024 17:59, Oleksii Kurochko wrote: >>> --- a/xen/arch/x86/xen.lds.S >>> +++ b/xen/arch/x86/xen.lds.S >>> @@ -321,14 +321,7 @@ SECTIONS >>> DECL_SECTION(.bss) { >>> __bss_start = .; >>> *(.bss.page_aligned*) >>> - . = ALIGN(PAGE_SIZE); >>> - __per_cpu_start = .; >>> - *(.bss.percpu.page_aligned) >>> - *(.bss.percpu) >>> - . = ALIGN(SMP_CACHE_BYTES); >>> - *(.bss.percpu.read_mostly) >>> - . = ALIGN(SMP_CACHE_BYTES); >>> - __per_cpu_data_end = .; >>> + PERCPU_SECTION >>> *(.bss .bss.*) >>> . = ALIGN(POINTER_ALIGN); >>> __bss_end = .; >> Like the _SEC in the other patch I question _SECTION here, albeit for >> a different >> reason: This is no separate output section, and it's more than one >> kind of input >> ones. Perhaps PERCPU_DATA? With that >> Acked-by: Jan Beulich <jbeulich@suse.com> > Sure, I will drop _SECTION here too. Thanks. Can we call it PERCPU_BSS? Just to highlight the fact that it really is BSS, and not initialised DATA. ~Andrew
On 23.09.2024 13:53, Andrew Cooper wrote: > On 23/09/2024 12:48 pm, oleksii.kurochko@gmail.com wrote: >> On Mon, 2024-09-23 at 12:56 +0200, Jan Beulich wrote: >>> On 19.09.2024 17:59, Oleksii Kurochko wrote: >>>> --- a/xen/arch/x86/xen.lds.S >>>> +++ b/xen/arch/x86/xen.lds.S >>>> @@ -321,14 +321,7 @@ SECTIONS >>>> DECL_SECTION(.bss) { >>>> __bss_start = .; >>>> *(.bss.page_aligned*) >>>> - . = ALIGN(PAGE_SIZE); >>>> - __per_cpu_start = .; >>>> - *(.bss.percpu.page_aligned) >>>> - *(.bss.percpu) >>>> - . = ALIGN(SMP_CACHE_BYTES); >>>> - *(.bss.percpu.read_mostly) >>>> - . = ALIGN(SMP_CACHE_BYTES); >>>> - __per_cpu_data_end = .; >>>> + PERCPU_SECTION >>>> *(.bss .bss.*) >>>> . = ALIGN(POINTER_ALIGN); >>>> __bss_end = .; >>> Like the _SEC in the other patch I question _SECTION here, albeit for >>> a different >>> reason: This is no separate output section, and it's more than one >>> kind of input >>> ones. Perhaps PERCPU_DATA? With that >>> Acked-by: Jan Beulich <jbeulich@suse.com> >> Sure, I will drop _SECTION here too. Thanks. > > Can we call it PERCPU_BSS? Just to highlight the fact that it really is > BSS, and not initialised DATA. Fine with me. Jan
diff --git a/xen/arch/arm/xen.lds.S b/xen/arch/arm/xen.lds.S index bd884664ad..669a882455 100644 --- a/xen/arch/arm/xen.lds.S +++ b/xen/arch/arm/xen.lds.S @@ -198,14 +198,7 @@ SECTIONS __bss_start = .; *(.bss.stack_aligned) *(.bss.page_aligned) - . = ALIGN(PAGE_SIZE); - __per_cpu_start = .; - *(.bss.percpu.page_aligned) - *(.bss.percpu) - . = ALIGN(SMP_CACHE_BYTES); - *(.bss.percpu.read_mostly) - . = ALIGN(SMP_CACHE_BYTES); - __per_cpu_data_end = .; + PERCPU_SECTION *(.bss .bss.*) . = ALIGN(POINTER_ALIGN); __bss_end = .; diff --git a/xen/arch/ppc/xen.lds.S b/xen/arch/ppc/xen.lds.S index 38cd857187..0833d80479 100644 --- a/xen/arch/ppc/xen.lds.S +++ b/xen/arch/ppc/xen.lds.S @@ -148,14 +148,7 @@ SECTIONS __bss_start = .; *(.bss.stack_aligned) *(.bss.page_aligned) - . = ALIGN(PAGE_SIZE); - __per_cpu_start = .; - *(.bss.percpu.page_aligned) - *(.bss.percpu) - . = ALIGN(SMP_CACHE_BYTES); - *(.bss.percpu.read_mostly) - . = ALIGN(SMP_CACHE_BYTES); - __per_cpu_data_end = .; + PERCPU_SECTION *(.bss .bss.*) . = ALIGN(POINTER_ALIGN); __bss_end = .; diff --git a/xen/arch/x86/xen.lds.S b/xen/arch/x86/xen.lds.S index d48de67cfd..eea8edc02b 100644 --- a/xen/arch/x86/xen.lds.S +++ b/xen/arch/x86/xen.lds.S @@ -321,14 +321,7 @@ SECTIONS DECL_SECTION(.bss) { __bss_start = .; *(.bss.page_aligned*) - . = ALIGN(PAGE_SIZE); - __per_cpu_start = .; - *(.bss.percpu.page_aligned) - *(.bss.percpu) - . = ALIGN(SMP_CACHE_BYTES); - *(.bss.percpu.read_mostly) - . = ALIGN(SMP_CACHE_BYTES); - __per_cpu_data_end = .; + PERCPU_SECTION *(.bss .bss.*) . = ALIGN(POINTER_ALIGN); __bss_end = .; diff --git a/xen/include/xen/xen.lds.h b/xen/include/xen/xen.lds.h index a17810bb28..f043c7b6c0 100644 --- a/xen/include/xen/xen.lds.h +++ b/xen/include/xen/xen.lds.h @@ -151,6 +151,16 @@ #define LOCK_PROFILE_DATA #endif +#define PERCPU_SECTION \ + . = ALIGN(PAGE_SIZE); \ + __per_cpu_start = .; \ + *(.bss.percpu.page_aligned) \ + *(.bss.percpu) \ + . = ALIGN(SMP_CACHE_BYTES); \ + *(.bss.percpu.read_mostly) \ + . = ALIGN(SMP_CACHE_BYTES); \ + __per_cpu_data_end = .; \ + #ifdef CONFIG_HAS_VPCI #define VPCI_ARRAY \ . = ALIGN(POINTER_ALIGN); \
Introduce PERCPU_SECTION macro which manages: * Alignment of the section start * Insertion of per-CPU data sections * Alignment and start/end markers for per-CPU data This change simplifies the linker script maintenance and ensures a unified approach for per-CPU sections across different architectures. Refactor the linker scripts for Arm, PPC, and x86 architectures by using the common macro PERCPU_SECTION defined in xen/xen.lds.h to handle per-CPU data sections. No functional changes. Signed-off-by: Oleksii Kurochko <oleksii.kurochko@gmail.com> --- xen/arch/arm/xen.lds.S | 9 +-------- xen/arch/ppc/xen.lds.S | 9 +-------- xen/arch/x86/xen.lds.S | 9 +-------- xen/include/xen/xen.lds.h | 10 ++++++++++ 4 files changed, 13 insertions(+), 24 deletions(-)