Tidy up symbol end fixup
Subsystems (2)
v2 cover Tidy up symbol end fixup
Fixing up more symbol ends as introduced in:
https://lore.kernel.org/lkml/[email protected]/
caused perf annotate to run into memory limits - every symbol holds all the disassembled code in the annotation, and so making symbols ends further away dramatically increased memory usage (40MB to
Quoted 1 line
1GB). Modify the symbol end logic so that special kernel cases aren't
applied in the common case.
v2. Drops a merged patch. Fixes a build issue with libbfd enabled.
Ian Rogers (4):
perf symbols: Always do architecture specific fixups perf symbols: Add is_kernel argument to fixup end perf symbol: By default only fix zero length symbols perf symbols: More specific architecture end fixing
tools/perf/arch/arm64/util/machine.c | 14 +++++++++----- tools/perf/arch/powerpc/util/machine.c | 10 +++++++--- tools/perf/arch/s390/util/machine.c | 12 ++++++++---- tools/perf/util/symbol-elf.c | 2 +- tools/perf/util/symbol.c | 16 +++++++++------- tools/perf/util/symbol.h | 4 ++-- 6 files changed, 36 insertions(+), 22 deletions(-)
Signature · 1 line
2.35.1.1178.g4f1659d476-goog
Ian Rogers
reviewerPATCH v2 1/4 perf symbols: Always do architecture specific fixups
The change:
https://lore.kernel.org/lkml/[email protected]/
modified the condition for architecture specific fixups motivated by a PowerPC case. So that architectures can independently modify their condition, move the if into the called architecture symbols__fixup_end function and always call it.
--- tools/perf/arch/arm64/util/machine.c | 14 ++++++++------ tools/perf/arch/powerpc/util/machine.c | 14 ++++++++------ tools/perf/arch/s390/util/machine.c | 14 ++++++++------ tools/perf/util/symbol.c | 6 +++--- 4 files changed, 27 insertions(+), 21 deletions(-)
tools/perf/arch/arm64/util/machine.c +8 -6
@@ -20,13 +20,15 @@void arch__symbols__fixup_end (struct symbol *p,struct symbol *c) { -if ((strchr (p->name,'[' ) &&strchr (c->name,'[' ) ==NULL ) || +if (p->end == p->start || p->end != c->start) { +if ((strchr (p->name,'[' ) &&strchr (c->name,'[' ) ==NULL ) || (strchr (p->name,'[' ) ==NULL &&strchr (c->name,'[' ))) -/* Limit range of last symbol in module and kernel */ - p->end +=SYMBOL_LIMIT ; -else - p->end = c->start; -pr_debug4 ("%s sym:%s end:%#" PRIx64"\n" , __func__, p->name, p->end); +/* Limit range of last symbol in module and kernel */ + p->end +=SYMBOL_LIMIT ; +else + p->end = c->start; +pr_debug4 ("%s sym:%s end:%#" PRIx64"\n" , __func__, p->name, p->end); + } }void arch__add_leaf_frame_record_opts (struct record_opts *opts)
tools/perf/arch/powerpc/util/machine.c +8 -6
@@ -16,10 +16,12 @@void arch__symbols__fixup_end (struct symbol *p,struct symbol *c) { -if (strchr (p->name,'[' ) ==NULL &&strchr (c->name,'[' )) -/* Limit the range of last kernel symbol */ - p->end += page_size; -else - p->end = c->start; -pr_debug4 ("%s sym:%s end:%#" PRIx64"\n" , __func__, p->name, p->end); +if (p->end == p->start || p->end != c->start) { +if (strchr (p->name,'[' ) ==NULL &&strchr (c->name,'[' )) +/* Limit the range of last kernel symbol */ + p->end += page_size; +else + p->end = c->start; +pr_debug4 ("%s sym:%s end:%#" PRIx64"\n" , __func__, p->name, p->end); + } }
tools/perf/arch/s390/util/machine.c +8 -6
@@ -44,10 +44,12 @@ int arch__fix_module_text_start(u64 *start, u64 *size, const char *name)*/ void arch__symbols__fixup_end (struct symbol *p,struct symbol *c) { -if (strchr (p->name,'[' ) ==NULL &&strchr (c->name,'[' )) -/* Last kernel symbol mapped to end of page */ - p->end =roundup (p->end, page_size); -else - p->end = c->start; -pr_debug4 ("%s sym:%s end:%#" PRIx64"\n" , __func__, p->name, p->end); +if (p->end == p->start || p->end != c->start) { +if (strchr (p->name,'[' ) ==NULL &&strchr (c->name,'[' )) +/* Last kernel symbol mapped to end of page */ + p->end =roundup (p->end, page_size); +else + p->end = c->start; +pr_debug4 ("%s sym:%s end:%#" PRIx64"\n" , __func__, p->name, p->end); + } }
tools/perf/util/symbol.c +3 -3
@@ -103,7 +103,8 @@ static int prefix_underscores_count(const char *str)void __weak arch__symbols__fixup_end (struct symbol *p,struct symbol *c) { - p->end = c->start; +if (p->end == p->start || p->end != c->start) + p->end = c->start; }const char *__weak arch__normalize_symbol_name (const char *name)
@@ -231,8 +232,7 @@ void symbols__fixup_end(struct rb_root_cached *symbols) prev = curr; curr = rb_entry(nd, struct symbol, rb_node); -if (prev->end == prev->start || prev->end != curr->start) -arch__symbols__fixup_end (prev, curr); +arch__symbols__fixup_end (prev, curr); }/* Last entry */
Signature · 1 line
2.35.1.1178.g4f1659d476-goog
Ian Rogers
reviewerPATCH v2 2/4 perf symbols: Add is_kernel argument to fixup end
In reply to Ian RogersKernel fixups may be special so add a flag to quickly detect if they are necessary.
--- tools/perf/arch/arm64/util/machine.c | 3 ++- tools/perf/arch/powerpc/util/machine.c | 4 +++- tools/perf/arch/s390/util/machine.c | 3 ++- tools/perf/util/symbol-elf.c | 2 +- tools/perf/util/symbol.c | 11 ++++++----- tools/perf/util/symbol.h | 4 ++-- 6 files changed, 16 insertions(+), 11 deletions(-)
tools/perf/arch/arm64/util/machine.c +2 -1
@@ -18,7 +18,8 @@#define SYMBOL_LIMIT (1 << 12)/* 4K */ -void arch__symbols__fixup_end (struct symbol *p,struct symbol *c) +void arch__symbols__fixup_end (struct symbol *p,struct symbol *c, +bool is_kernel__maybe_unused ) {if (p->end == p->start || p->end != c->start) { if ((strchr(p->name, '[') && strchr(c->name, '[') == NULL) ||
tools/perf/arch/powerpc/util/machine.c +3 -1
@@ -14,7 +14,9 @@* Therefore do not fill this gap and do not assign it to the kernel dso map. */ -void arch__symbols__fixup_end (struct symbol *p,struct symbol *c) +void arch__symbols__fixup_end (struct symbol *p,struct symbol *c, +bool is_kernel__maybe_unused ) +) {if (p->end == p->start || p->end != c->start) {if (strchr (p->name,'[' ) ==NULL &&strchr (c->name,'[' ))
tools/perf/arch/s390/util/machine.c +2 -1
@@ -42,7 +42,8 @@ int arch__fix_module_text_start(u64 *start, u64 *size, const char *name)* and beginning of first module's text segment is very big. * Therefore do not fill this gap and do not assign it to the kernel dso map. */ -void arch__symbols__fixup_end (struct symbol *p,struct symbol *c) +void arch__symbols__fixup_end (struct symbol *p,struct symbol *c, +bool is_kernel__maybe_unused ) {if (p->end == p->start || p->end != c->start) {if (strchr (p->name,'[' ) ==NULL &&strchr (c->name,'[' ))
tools/perf/util/symbol-elf.c +1 -1
@@ -1290,7 +1290,7 @@ dso__load_sym_internal(struct dso *dso, struct map *map, struct symsrc *syms_ss,* For misannotated, zeroed, ASM function sizes. */ if (nr >0 ) { -symbols__fixup_end (&dso->symbols); +symbols__fixup_end (&dso->symbols, dso->kernel !=DSO_SPACE__USER );symbols__fixup_duplicate (&dso->symbols);if (kmap) {/*
tools/perf/util/symbol.c +6 -5
@@ -101,7 +101,8 @@ static int prefix_underscores_count(const char *str)return tail - str; } -void __weak arch__symbols__fixup_end (struct symbol *p,struct symbol *c) +void __weak arch__symbols__fixup_end (struct symbol *p,struct symbol *c, +bool is_kernel__maybe_unused ) {if (p->end == p->start || p->end != c->start) p->end = c->start;
@@ -218,7 +219,7 @@ void symbols__fixup_duplicate(struct rb_root_cached *symbols) } } -void symbols__fixup_end (struct rb_root_cached *symbols) +void symbols__fixup_end (struct rb_root_cached *symbols,bool is_kernel) {struct rb_node *nd, *prevnd =rb_first_cached (symbols);struct symbol *curr, *prev;
@@ -232,7 +233,7 @@ void symbols__fixup_end(struct rb_root_cached *symbols) prev = curr; curr = rb_entry(nd, struct symbol, rb_node); -arch__symbols__fixup_end (prev, curr); +arch__symbols__fixup_end (prev, curr, is_kernel); }/* Last entry */
@@ -1467,7 +1468,7 @@ int __dso__load_kallsyms(struct dso *dso, const char *filename,if (kallsyms__delta (kmap, filename, &delta))return -1 ; -symbols__fixup_end (&dso->symbols); +symbols__fixup_end (&dso->symbols,/*is_kernel=*/ true);symbols__fixup_duplicate (&dso->symbols);if (dso->kernel ==DSO_SPACE__KERNEL_GUEST )
@@ -1659,7 +1660,7 @@ int dso__load_bfd_symbols(struct dso *dso, const char *debugfile)#undef bfd_asymbol_section#endif -symbols__fixup_end (&dso->symbols); +symbols__fixup_end (&dso->symbols, dso->kernel !=DSO_SPACE__USER );symbols__fixup_duplicate (&dso->symbols); dso->adjust_symbols =1 ;
tools/perf/util/symbol.h +2 -2
@@ -203,7 +203,7 @@ void __symbols__insert(struct rb_root_cached *symbols, struct symbol *sym, bool kernel);void symbols__insert (struct rb_root_cached *symbols,struct symbol *sym);void symbols__fixup_duplicate (struct rb_root_cached *symbols); -void symbols__fixup_end (struct rb_root_cached *symbols); +void symbols__fixup_end (struct rb_root_cached *symbols,bool is_kernel);void maps__fixup_end (struct maps *maps);typedef int (*mapfn_t )(u64 start,u64 len,u64 pgoff,void *data);
@@ -241,7 +241,7 @@ const char *arch__normalize_symbol_name(const char *name);#define SYMBOL_A 0#define SYMBOL_B 1 -void arch__symbols__fixup_end (struct symbol *p,struct symbol *c); +void arch__symbols__fixup_end (struct symbol *p,struct symbol *c,bool is_kernel);int arch__compare_symbol_names (const char *namea,const char *nameb);int arch__compare_symbol_names_n (const char *namea,const char *nameb,unsigned int n);
Signature · 1 line
2.35.1.1178.g4f1659d476-goog
Ian Rogers
reviewerPATCH v2 3/4 perf symbol: By default only fix zero length symbols
In reply to Ian RogersFor architectures without a specific end fixup (ie not arm64, powerpc, s390) only fix up the end of zero length symbols. This reverts the behavior introduced by:
https://lore.kernel.org/lkml/[email protected]/
where non-zero length symbols were expanded to the start of the current symbol.
--- tools/perf/util/symbol.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
tools/perf/util/symbol.c +2 -1
@@ -104,7 +104,8 @@ static int prefix_underscores_count(const char *str)void __weak arch__symbols__fixup_end (struct symbol *p,struct symbol *c,bool is_kernel__maybe_unused ) { -if (p->end == p->start || p->end != c->start) +/* If the previous symbol is zero length, make its end the start of the current symbol. */ +if (p->end == p->start) p->end = c->start; }
Signature · 1 line
2.35.1.1178.g4f1659d476-goog
Ian Rogers
reviewerPATCH v2 4/4 perf symbols: More specific architecture end fixing
In reply to Ian RogersMake the conditions used for symbol fixup closer to the comments. The logic aims to combine the current conditions as last modified in:
https://lore.kernel.org/lkml/[email protected]/
--- tools/perf/arch/arm64/util/machine.c | 19 ++++++++++--------- tools/perf/arch/powerpc/util/machine.c | 18 +++++++++--------- tools/perf/arch/s390/util/machine.c | 17 +++++++++-------- 3 files changed, 28 insertions(+), 26 deletions(-)
tools/perf/arch/arm64/util/machine.c +10 -9
@@ -18,16 +18,17 @@#define SYMBOL_LIMIT (1 << 12)/* 4K */ -void arch__symbols__fixup_end (struct symbol *p,struct symbol *c, -bool is_kernel__maybe_unused ) +void arch__symbols__fixup_end (struct symbol *p,struct symbol *c,bool is_kernel) { -if (p->end == p->start || p->end != c->start) { -if ((strchr (p->name,'[' ) &&strchr (c->name,'[' ) ==NULL ) || - (strchr (p->name,'[' ) ==NULL &&strchr (c->name,'[' ))) -/* Limit range of last symbol in module and kernel */ - p->end +=SYMBOL_LIMIT ; -else - p->end = c->start; +if (is_kernel && (p->end == p->start || p->end != c->start) && + ((strchr (p->name,'[' ) &&strchr (c->name,'[' ) ==NULL ) || + (strchr (p->name,'[' ) ==NULL &&strchr (c->name,'[' )))) { +/* Limit range of last symbol in module and kernel */ + p->end +=SYMBOL_LIMIT ; +pr_debug4 ("%s sym:%s end:%#" PRIx64"\n" , __func__, p->name, p->end); + }else if (p->end == p->start) { +/* Expand empty symbols. */ + p->end = c->start;pr_debug4 ("%s sym:%s end:%#" PRIx64"\n" , __func__, p->name, p->end); } }
tools/perf/arch/powerpc/util/machine.c +9 -9
@@ -14,16 +14,16 @@* Therefore do not fill this gap and do not assign it to the kernel dso map. */ -void arch__symbols__fixup_end (struct symbol *p,struct symbol *c, -bool is_kernel__maybe_unused ) +void arch__symbols__fixup_end (struct symbol *p,struct symbol *c,bool is_kernel) ) { -if (p->end == p->start || p->end != c->start) { -if (strchr (p->name,'[' ) ==NULL &&strchr (c->name,'[' )) -/* Limit the range of last kernel symbol */ - p->end += page_size; -else - p->end = c->start; -pr_debug4 ("%s sym:%s end:%#" PRIx64"\n" , __func__, p->name, p->end); +if (is_kernel && (p->end == p->start || p->end != c->start) && +strchr (p->name,'[' ) ==NULL &&strchr (c->name,'[' )) { +/* Limit the range of last kernel symbol */ + p->end += page_size; + }else if (p->end == p->start) { +/* Expand empty symbols. */ + p->end = c->start; } +pr_debug4 ("%s sym:%s end:%#" PRIx64"\n" , __func__, p->name, p->end); }
tools/perf/arch/s390/util/machine.c +9 -8
@@ -42,15 +42,16 @@ int arch__fix_module_text_start(u64 *start, u64 *size, const char *name)* and beginning of first module's text segment is very big. * Therefore do not fill this gap and do not assign it to the kernel dso map. */ -void arch__symbols__fixup_end (struct symbol *p,struct symbol *c, -bool is_kernel__maybe_unused ) +void arch__symbols__fixup_end (struct symbol *p,struct symbol *c,bool is_kernel) { -if (p->end == p->start || p->end != c->start) { -if (strchr (p->name,'[' ) ==NULL &&strchr (c->name,'[' )) -/* Last kernel symbol mapped to end of page */ - p->end =roundup (p->end, page_size); -else - p->end = c->start; +if (is_kernel && (p->end == p->start || p->end != c->start) && +strchr (p->name,'[' ) ==NULL &&strchr (c->name,'[' )) { +/* Last kernel symbol mapped to end of page */ + p->end =roundup (p->end, page_size); +pr_debug4 ("%s sym:%s end:%#" PRIx64"\n" , __func__, p->name, p->end); + }else if (p->end == p->start) { +/* Expand empty symbols. */ + p->end = c->start;pr_debug4 ("%s sym:%s end:%#" PRIx64"\n" , __func__, p->name, p->end); } }
Signature · 1 line
2.35.1.1178.g4f1659d476-goog
Ian Rogers
reviewerQuoted 17 lines · Ian Rogers
Fixing up more symbol ends as introduced in:
https://lore.kernel.org/lkml/[email protected]/
caused perf annotate to run into memory limits - every symbol holds all the disassembled code in the annotation, and so making symbols ends further away dramatically increased memory usage (40MB toQuoted 1 line
1GB). Modify the symbol end logic so that special kernel cases aren't
applied in the common case.
v2. Drops a merged patch. Fixes a build issue with libbfd enabled.
Ian Rogers (4):
perf symbols: Always do architecture specific fixups perf symbols: Addis_kernelargument to fixup end perf symbol: By default only fix zero length symbols perf symbols: More specific architecture end fixing
Quoted 7 lines
tools/perf/arch/arm64/util/machine.c | 14 +++++++++----- tools/perf/arch/powerpc/util/machine.c | 10 +++++++--- tools/perf/arch/s390/util/machine.c | 12 ++++++++---- tools/perf/util/symbol-elf.c | 2 +- tools/perf/util/symbol.c | 16 +++++++++------- tools/perf/util/symbol.h | 4 ++-- 6 files changed, 36 insertions(+), 22 deletions(-)
Missed the:
Thanks,
Ian
Quoted 3 lines
--
2.35.1.1178.g4f1659d476-goog
Namhyung Kim
maintainerHi Ian,
Quoted 9 lines · Ian Rogers
Fixing up more symbol ends as introduced in:
https://lore.kernel.org/lkml/[email protected]/
caused perf annotate to run into memory limits - every symbol holds all the disassembled code in the annotation, and so making symbols ends further away dramatically increased memory usage (40MB toQuoted 1 line
1GB). Modify the symbol end logic so that special kernel cases aren't
applied in the common case.
v2. Drops a merged patch. Fixes a build issue with libbfd enabled.
How about just like this? We can get rid of arch functions as they mostly do the same thing (kernel vs module boundary check).
Not tested. ;-)
Thanks,
Namhyung
--------------8<-------------
tools/perf/util/symbol.c +11 -2
@@ -35,6 +35,7 @@#include "path.h" #include <linux/ctype.h> #include <linux/zalloc.h> +#include <internal/lib.h> // page_size #include <elf.h> #include <limits.h>
@@ -231,8 +226,16 @@ void symbols__fixup_end(struct rb_root_cached *symbols) prev = curr; curr = rb_entry(nd, struct symbol, rb_node); -if (prev->end == prev->start || prev->end != curr->start) -arch__symbols__fixup_end (prev, curr); +if (prev->end == prev->start) { +/* Last kernel symbol mapped to end of page */ +if (!strchr (prev->name,'[' ) != !strchr (curr->name,'[' )) + prev->end =roundup (prev->end +1 , page_size); +else + prev->end = curr->start; + +pr_debug4 ("%s sym:%s end:%#" PRIx64"\n" , + __func__, prev->name, prev->end); + } }/* Last entry */
_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
Ian Rogers
reviewerQuoted 37 lines · Namhyung Kim
Hi Ian,
Quoted 9 lines · Ian Rogers
Fixing up more symbol ends as introduced in:
https://lore.kernel.org/lkml/[email protected]/
caused perf annotate to run into memory limits - every symbol holds all the disassembled code in the annotation, and so making symbols ends further away dramatically increased memory usage (40MB toQuoted 1 line
1GB). Modify the symbol end logic so that special kernel cases aren't
applied in the common case.
v2. Drops a merged patch. Fixes a build issue with libbfd enabled.
How about just like this? We can get rid of arch functions as they mostly do the same thing (kernel vs module boundary check).
Not tested. ;-)
Thanks,
Namhyung--------------8<-------------
tools/perf/util/symbol.c +1
-0@@ -35,6 +35,7 @@#include "path.h" #include <linux/ctype.h> #include <linux/zalloc.h> +#include <internal/lib.h> // page_size #include <elf.h> #include <limits.h>
tools/perf/util/symbol.c void symbols__fixup_end(struct rb_root_cached *symbols) · line 230
Quoted 7 lines
tools/perf/util/symbol.c +2
-2@@ -231,8 +226,16 @@ void symbols__fixup_end(struct rb_root_cached *symbols) prev = curr; curr = rb_entry(nd, struct symbol, rb_node); -if (prev->end == prev->start || prev->end != curr->start) -arch__symbols__fixup_end (prev, curr); +if (prev->end == prev->start) { +/* Last kernel symbol mapped to end of page */
I like the simpler logic but don't like applying this in symbol-elf given the comment says it is kernel specific - so we could keep the is_kernel change.
tools/perf/util/symbol.c void symbols__fixup_end(struct rb_root_cached *symbols) · line 231
Quoted 1 line
tools/perf/util/symbol.c +1
-0+if (!strchr (prev->name,'[' ) != !strchr (curr->name,'[' ))
I find this condition not to be intention revealing. On ARM there is also an || for the condition reversed. When this is in an is_kernel block then I think it is clear this is kernel hack, so I think it is good to comment on what the condition is for.
tools/perf/util/symbol.c void symbols__fixup_end(struct rb_root_cached *symbols) · line 232
Quoted 1 line
tools/perf/util/symbol.c +1
-0+ prev->end =roundup (prev->end +1 , page_size);
Currently the roundup varies per architecture, but it is not clear to me that it matters.
tools/perf/util/symbol.c void symbols__fixup_end(struct rb_root_cached *symbols) · line 233
Quoted 1 line
tools/perf/util/symbol.c +1
-0+else
I think we should comment here that we're extending zero sized symbols to the start of the next symbol.
tools/perf/util/symbol.c void symbols__fixup_end(struct rb_root_cached *symbols) · line 238
Quoted 5 lines
tools/perf/util/symbol.c +5
-0+ prev->end = curr->start; + +pr_debug4 ("%s sym:%s end:%#" PRIx64"\n" , + __func__, prev->name, prev->end); + }
Thanks,
Ian
Quoted 3 lines
tools/perf/util/symbol.c +0
-0}/* Last entry */
Namhyung Kim
maintainerQuoted 42 lines · Ian Rogers
Quoted 40 lines · Namhyung Kim
Hi Ian,
Quoted 9 lines · Ian Rogers
Fixing up more symbol ends as introduced in:
https://lore.kernel.org/lkml/[email protected]/
caused perf annotate to run into memory limits - every symbol holds all the disassembled code in the annotation, and so making symbols ends further away dramatically increased memory usage (40MB toQuoted 1 line
1GB). Modify the symbol end logic so that special kernel cases aren't
applied in the common case.
v2. Drops a merged patch. Fixes a build issue with libbfd enabled.
How about just like this? We can get rid of arch functions as they mostly do the same thing (kernel vs module boundary check).
Not tested. ;-)
Thanks,
Namhyung--------------8<-------------
tools/perf/util/symbol.c +1
-0@@ -35,6 +35,7 @@#include "path.h" #include <linux/ctype.h> #include <linux/zalloc.h> +#include <internal/lib.h> // page_size #include <elf.h> #include <limits.h> @@ -231,8 +226,16 @@ void symbols__fixup_end(struct rb_root_cached *symbols) prev = curr; curr = rb_entry(nd, struct symbol, rb_node);
Quoted 8 lines
tools/perf/util/symbol.c void symbols__fixup_end(struct rb_root_cached *symbols) · line 230
Quoted 4 lines
tools/perf/util/symbol.c +2
-2-if (prev->end == prev->start || prev->end != curr->start) -arch__symbols__fixup_end (prev, curr); +if (prev->end == prev->start) { +/* Last kernel symbol mapped to end of page */ I like the simpler logic but don't like applying this in symbol-elf given the comment says it is kernel specific - so we could keep the
is_kernelchange.
I'm fine with the change. :)
Quoted 7 lines
tools/perf/util/symbol.c void symbols__fixup_end(struct rb_root_cached *symbols) · line 231
Quoted 2 lines
tools/perf/util/symbol.c +1
-0+if (!strchr (prev->name,'[' ) != !strchr (curr->name,'[' ))I find this condition not to be intention revealing. On ARM there is also an || for the condition reversed. When this is in an
is_kernelblock then I think it is clear this is kernel hack, so I think it is good to comment on what the condition is for.
Yeah, usually modules are loaded after the kernel image but it seems ARM could load them before the kernel.
So I made the change not to call strchr() again.
But we might need to consider the special "[__builtin_kprobes]" symbols.
Quoted 5 lines
tools/perf/util/symbol.c void symbols__fixup_end(struct rb_root_cached *symbols) · line 232
Quoted 2 lines
tools/perf/util/symbol.c +1
-0+ prev->end =roundup (prev->end +1 , page_size);Currently the roundup varies per architecture, but it is not clear to me that it matters.
Yeah, it would be the same as the logic for the last entry to be more conservative.
Quoted 5 lines
tools/perf/util/symbol.c void symbols__fixup_end(struct rb_root_cached *symbols) · line 233
Quoted 2 lines
tools/perf/util/symbol.c +1
-0+else I think we should comment here that we're extending zero sized symbols to the start of the next symbol.
Sounds good.
Thanks,
Namhyung
Quoted 13 lines
tools/perf/util/symbol.c void symbols__fixup_end(struct rb_root_cached *symbols) · line 238
Quoted 6 lines
tools/perf/util/symbol.c +5
-0+ prev->end = curr->start; + +pr_debug4 ("%s sym:%s end:%#" PRIx64"\n" , + __func__, prev->name, prev->end); + }Thanks,
IanQuoted 3 lines
tools/perf/util/symbol.c +0
-0}/* Last entry */