[PATCH 0/2] Kbuild: kallsyms: update suffix ignore logic

[PATCH 0/2] Kbuild: kallsyms: update suffix ignore logic

[Ard Biesheuvel]

Two patches that update the kallsyms ignore logic.

Patch #1 adds logic to introduce symbols whose names end in (or equal)
"_veneer". This prevent kallsyms generation errors on large ARM kernels
where the emitted veneers will be different between the first and second
pass, due to the fact that the size of the kallsyms region itself pushes
the .text and .init.text sections further apart during the second pass.

Patch #2 removes the handling of the "gcc[0-9]_compiled." symbol. It was
only ever emitted by GCC v2.x which we don't support anymore anyway.

This series is a followup to ' [RFC PATCH 0/2] ARM large kernels'
(http://thread.gmane.org/gmane.linux.kbuild.devel/13065), but the
other issue addressed by patch #2 of that series will be deferred
for now.

Ard Biesheuvel (2):
  Kbuild: kallsyms: ignore veneers emitted by the ARM linker
  Kbuild: kallsyms: drop special handling of pre-3.0 GCC symbols

 scripts/kallsyms.c | 29 ++++++++++++++++++++---------
 1 file changed, 20 insertions(+), 9 deletions(-)

-- 
1.8.3.2

[PATCH 0/2] Kbuild: kallsyms: update suffix ignore logic

[Ard Biesheuvel]

Two patches that update the kallsyms ignore logic.

Patch #1 adds logic to introduce symbols whose names end in (or equal)
"_veneer". This prevent kallsyms generation errors on large ARM kernels
where the emitted veneers will be different between the first and second
pass, due to the fact that the size of the kallsyms region itself pushes
the .text and .init.text sections further apart during the second pass.

Patch #2 removes the handling of the "gcc[0-9]_compiled." symbol. It was
only ever emitted by GCC v2.x which we don't support anymore anyway.

This series is a followup to ' [RFC PATCH 0/2] ARM large kernels'
(http://thread.gmane.org/gmane.linux.kbuild.devel/13065), but the
other issue addressed by patch #2 of that series will be deferred
for now.

Ard Biesheuvel (2):
  Kbuild: kallsyms: ignore veneers emitted by the ARM linker
  Kbuild: kallsyms: drop special handling of pre-3.0 GCC symbols

 scripts/kallsyms.c | 29 ++++++++++++++++++++---------
 1 file changed, 20 insertions(+), 9 deletions(-)

-- 
1.8.3.2

[PATCH 1/2] Kbuild: kallsyms: ignore veneers emitted by the ARM linker

[Ard Biesheuvel]

When linking large kernels on ARM, the linker will insert veneers
(i.e., PLT like stubs) when function symbols are out of reach for
the ordinary relative branch/branch-and-link instructions.

However, due to the fact that the kallsyms region sits in .rodata,
which is between .text and .init.text, additional veneers may be
emitted in the second pass due to the fact that the size of the
kallsyms region itself has pushed the .init.text section further
apart, requiring even more veneers.

So ignore the veneers when generating the symbol table. Veneers
have no corresponding source code, and they will not turn up in
backtraces anyway.

This patch also lightly refactors the symbol_valid() function
to use a local 'sym_name' rather than the obfuscated 'sym + 1'
and 'sym + offset'

Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
---
 scripts/kallsyms.c | 30 +++++++++++++++++++++---------
 1 file changed, 21 insertions(+), 9 deletions(-)

diff --git a/scripts/kallsyms.c b/scripts/kallsyms.c
index c6d33bd15b04..f4b016782f0d 100644
--- a/scripts/kallsyms.c
+++ b/scripts/kallsyms.c
@@ -212,15 +212,23 @@ static int symbol_valid(struct sym_entry *s)
 		"_SDA_BASE_",		/* ppc */
 		"_SDA2_BASE_",		/* ppc */
 		NULL };
+
+	static char *special_suffixes[] = {
+		"_compiled.",		/* gcc < 3.0: "gcc[0-9]_compiled." */
+		"_veneer",		/* arm */
+		NULL };
+
 	int i;
-	int offset = 1;
+	char *sym_name = (char *)s->sym + 1;
+
 
 	if (s->addr < kernel_start_addr)
 		return 0;
 
 	/* skip prefix char */
-	if (symbol_prefix_char && *(s->sym + 1) == symbol_prefix_char)
-		offset++;
+	if (symbol_prefix_char && *sym_name == symbol_prefix_char)
+		sym_name++;
+
 
 	/* if --all-symbols is not specified, then symbols outside the text
 	 * and inittext sections are discarded */
@@ -235,22 +243,26 @@ static int symbol_valid(struct sym_entry *s)
 		 * rules.
 		 */
 		if ((s->addr == text_range_text->end &&
-				strcmp((char *)s->sym + offset,
+				strcmp(sym_name,
 				       text_range_text->end_sym)) ||
 		    (s->addr == text_range_inittext->end &&
-				strcmp((char *)s->sym + offset,
+				strcmp(sym_name,
 				       text_range_inittext->end_sym)))
 			return 0;
 	}
 
 	/* Exclude symbols which vary between passes. */
-	if (strstr((char *)s->sym + offset, "_compiled."))
-		return 0;
-
 	for (i = 0; special_symbols[i]; i++)
-		if( strcmp((char *)s->sym + offset, special_symbols[i]) == 0 )
+		if (strcmp(sym_name, special_symbols[i]) == 0)
 			return 0;
 
+	for (i = 0; special_suffixes[i]; i++) {
+		int l = strlen(sym_name) - strlen(special_suffixes[i]);
+
+		if (l >= 0 && strcmp(sym_name + l, special_suffixes[i]) == 0)
+			return 0;
+	}
+
 	return 1;
 }
 
-- 
1.8.3.2

[PATCH 1/2] Kbuild: kallsyms: ignore veneers emitted by the ARM linker

[Ard Biesheuvel]

When linking large kernels on ARM, the linker will insert veneers
(i.e., PLT like stubs) when function symbols are out of reach for
the ordinary relative branch/branch-and-link instructions.

However, due to the fact that the kallsyms region sits in .rodata,
which is between .text and .init.text, additional veneers may be
emitted in the second pass due to the fact that the size of the
kallsyms region itself has pushed the .init.text section further
apart, requiring even more veneers.

So ignore the veneers when generating the symbol table. Veneers
have no corresponding source code, and they will not turn up in
backtraces anyway.

This patch also lightly refactors the symbol_valid() function
to use a local 'sym_name' rather than the obfuscated 'sym + 1'
and 'sym + offset'

Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
---
 scripts/kallsyms.c | 30 +++++++++++++++++++++---------
 1 file changed, 21 insertions(+), 9 deletions(-)

diff --git a/scripts/kallsyms.c b/scripts/kallsyms.c
index c6d33bd15b04..f4b016782f0d 100644
--- a/scripts/kallsyms.c
+++ b/scripts/kallsyms.c
@@ -212,15 +212,23 @@ static int symbol_valid(struct sym_entry *s)
 		"_SDA_BASE_",		/* ppc */
 		"_SDA2_BASE_",		/* ppc */
 		NULL };
+
+	static char *special_suffixes[] = {
+		"_compiled.",		/* gcc < 3.0: "gcc[0-9]_compiled." */
+		"_veneer",		/* arm */
+		NULL };
+
 	int i;
-	int offset = 1;
+	char *sym_name = (char *)s->sym + 1;
+
 
 	if (s->addr < kernel_start_addr)
 		return 0;
 
 	/* skip prefix char */
-	if (symbol_prefix_char && *(s->sym + 1) == symbol_prefix_char)
-		offset++;
+	if (symbol_prefix_char && *sym_name == symbol_prefix_char)
+		sym_name++;
+
 
 	/* if --all-symbols is not specified, then symbols outside the text
 	 * and inittext sections are discarded */
@@ -235,22 +243,26 @@ static int symbol_valid(struct sym_entry *s)
 		 * rules.
 		 */
 		if ((s->addr == text_range_text->end &&
-				strcmp((char *)s->sym + offset,
+				strcmp(sym_name,
 				       text_range_text->end_sym)) ||
 		    (s->addr == text_range_inittext->end &&
-				strcmp((char *)s->sym + offset,
+				strcmp(sym_name,
 				       text_range_inittext->end_sym)))
 			return 0;
 	}
 
 	/* Exclude symbols which vary between passes. */
-	if (strstr((char *)s->sym + offset, "_compiled."))
-		return 0;
-
 	for (i = 0; special_symbols[i]; i++)
-		if( strcmp((char *)s->sym + offset, special_symbols[i]) == 0 )
+		if (strcmp(sym_name, special_symbols[i]) == 0)
 			return 0;
 
+	for (i = 0; special_suffixes[i]; i++) {
+		int l = strlen(sym_name) - strlen(special_suffixes[i]);
+
+		if (l >= 0 && strcmp(sym_name + l, special_suffixes[i]) == 0)
+			return 0;
+	}
+
 	return 1;
 }
 
-- 
1.8.3.2

[PATCH 2/2] Kbuild: kallsyms: drop special handling of pre-3.0 GCC symbols

[Ard Biesheuvel]

Since we have required at least GCC v3.2 for some time now, we
can drop the special handling of the 'gcc[0-9]_compiled.' label
which is not emitted anymore since GCC v3.0.

Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
---
 scripts/kallsyms.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/scripts/kallsyms.c b/scripts/kallsyms.c
index f4b016782f0d..8fa81e84e295 100644
--- a/scripts/kallsyms.c
+++ b/scripts/kallsyms.c
@@ -214,7 +214,6 @@ static int symbol_valid(struct sym_entry *s)
 		NULL };
 
 	static char *special_suffixes[] = {
-		"_compiled.",		/* gcc < 3.0: "gcc[0-9]_compiled." */
 		"_veneer",		/* arm */
 		NULL };
 
-- 
1.8.3.2

[PATCH 2/2] Kbuild: kallsyms: drop special handling of pre-3.0 GCC symbols

[Ard Biesheuvel]

Since we have required at least GCC v3.2 for some time now, we
can drop the special handling of the 'gcc[0-9]_compiled.' label
which is not emitted anymore since GCC v3.0.

Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
---
 scripts/kallsyms.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/scripts/kallsyms.c b/scripts/kallsyms.c
index f4b016782f0d..8fa81e84e295 100644
--- a/scripts/kallsyms.c
+++ b/scripts/kallsyms.c
@@ -214,7 +214,6 @@ static int symbol_valid(struct sym_entry *s)
 		NULL };
 
 	static char *special_suffixes[] = {
-		"_compiled.",		/* gcc < 3.0: "gcc[0-9]_compiled." */
 		"_veneer",		/* arm */
 		NULL };
 
-- 
1.8.3.2

Re: [PATCH 0/2] Kbuild: kallsyms: update suffix ignore logic

[Ard Biesheuvel]

On 30 March 2015 at 15:20, Ard Biesheuvel <ard.biesheuvel@linaro.org> wrote:
> Two patches that update the kallsyms ignore logic.
>
> Patch #1 adds logic to introduce symbols whose names end in (or equal)
> "_veneer". This prevent kallsyms generation errors on large ARM kernels
> where the emitted veneers will be different between the first and second
> pass, due to the fact that the size of the kallsyms region itself pushes
> the .text and .init.text sections further apart during the second pass.
>
> Patch #2 removes the handling of the "gcc[0-9]_compiled." symbol. It was
> only ever emitted by GCC v2.x which we don't support anymore anyway.
>
> This series is a followup to ' [RFC PATCH 0/2] ARM large kernels'
> (http://thread.gmane.org/gmane.linux.kbuild.devel/13065), but the
> other issue addressed by patch #2 of that series will be deferred
> for now.
>
> Ard Biesheuvel (2):
>   Kbuild: kallsyms: ignore veneers emitted by the ARM linker
>   Kbuild: kallsyms: drop special handling of pre-3.0 GCC symbols
>
>  scripts/kallsyms.c | 29 ++++++++++++++++++++---------
>  1 file changed, 20 insertions(+), 9 deletions(-)
>

Ping?

Re: [PATCH 0/2] Kbuild: kallsyms: update suffix ignore logic

[Ard Biesheuvel]

On 30 March 2015 at 15:20, Ard Biesheuvel <ard.biesheuvel@linaro.org> wrote:
> Two patches that update the kallsyms ignore logic.
>
> Patch #1 adds logic to introduce symbols whose names end in (or equal)
> "_veneer". This prevent kallsyms generation errors on large ARM kernels
> where the emitted veneers will be different between the first and second
> pass, due to the fact that the size of the kallsyms region itself pushes
> the .text and .init.text sections further apart during the second pass.
>
> Patch #2 removes the handling of the "gcc[0-9]_compiled." symbol. It was
> only ever emitted by GCC v2.x which we don't support anymore anyway.
>
> This series is a followup to ' [RFC PATCH 0/2] ARM large kernels'
> (http://thread.gmane.org/gmane.linux.kbuild.devel/13065), but the
> other issue addressed by patch #2 of that series will be deferred
> for now.
>
> Ard Biesheuvel (2):
>   Kbuild: kallsyms: ignore veneers emitted by the ARM linker
>   Kbuild: kallsyms: drop special handling of pre-3.0 GCC symbols
>
>  scripts/kallsyms.c | 29 ++++++++++++++++++++---------
>  1 file changed, 20 insertions(+), 9 deletions(-)
>

Ping?

[PATCH 0/2] Kbuild: kallsyms: update suffix ignore logic

[Ard Biesheuvel]

On 30 March 2015 at 15:20, Ard Biesheuvel <ard.biesheuvel@linaro.org> wrote:
> Two patches that update the kallsyms ignore logic.
>
> Patch #1 adds logic to introduce symbols whose names end in (or equal)
> "_veneer". This prevent kallsyms generation errors on large ARM kernels
> where the emitted veneers will be different between the first and second
> pass, due to the fact that the size of the kallsyms region itself pushes
> the .text and .init.text sections further apart during the second pass.
>
> Patch #2 removes the handling of the "gcc[0-9]_compiled." symbol. It was
> only ever emitted by GCC v2.x which we don't support anymore anyway.
>
> This series is a followup to ' [RFC PATCH 0/2] ARM large kernels'
> (http://thread.gmane.org/gmane.linux.kbuild.devel/13065), but the
> other issue addressed by patch #2 of that series will be deferred
> for now.
>
> Ard Biesheuvel (2):
>   Kbuild: kallsyms: ignore veneers emitted by the ARM linker
>   Kbuild: kallsyms: drop special handling of pre-3.0 GCC symbols
>
>  scripts/kallsyms.c | 29 ++++++++++++++++++++---------
>  1 file changed, 20 insertions(+), 9 deletions(-)
>

Ping?

Re: [PATCH 0/2] Kbuild: kallsyms: update suffix ignore logic

[Michal Marek]

Dne 7.4.2015 v 11:02 Ard Biesheuvel napsal(a):
> On 30 March 2015 at 15:20, Ard Biesheuvel <ard.biesheuvel@linaro.org> wrote:
>> Two patches that update the kallsyms ignore logic.
>>
>> Patch #1 adds logic to introduce symbols whose names end in (or equal)
>> "_veneer". This prevent kallsyms generation errors on large ARM kernels
>> where the emitted veneers will be different between the first and second
>> pass, due to the fact that the size of the kallsyms region itself pushes
>> the .text and .init.text sections further apart during the second pass.
>>
>> Patch #2 removes the handling of the "gcc[0-9]_compiled." symbol. It was
>> only ever emitted by GCC v2.x which we don't support anymore anyway.
>>
>> This series is a followup to ' [RFC PATCH 0/2] ARM large kernels'
>> (http://thread.gmane.org/gmane.linux.kbuild.devel/13065), but the
>> other issue addressed by patch #2 of that series will be deferred
>> for now.
>>
>> Ard Biesheuvel (2):
>>   Kbuild: kallsyms: ignore veneers emitted by the ARM linker
>>   Kbuild: kallsyms: drop special handling of pre-3.0 GCC symbols
>>
>>  scripts/kallsyms.c | 29 ++++++++++++++++++++---------
>>  1 file changed, 20 insertions(+), 9 deletions(-)
>>
> 
> Ping?

Applied to kbuild.git#kbuild.

Michal

Re: [PATCH 0/2] Kbuild: kallsyms: update suffix ignore logic

[Michal Marek]

Dne 7.4.2015 v 11:02 Ard Biesheuvel napsal(a):
> On 30 March 2015 at 15:20, Ard Biesheuvel <ard.biesheuvel@linaro.org> wrote:
>> Two patches that update the kallsyms ignore logic.
>>
>> Patch #1 adds logic to introduce symbols whose names end in (or equal)
>> "_veneer". This prevent kallsyms generation errors on large ARM kernels
>> where the emitted veneers will be different between the first and second
>> pass, due to the fact that the size of the kallsyms region itself pushes
>> the .text and .init.text sections further apart during the second pass.
>>
>> Patch #2 removes the handling of the "gcc[0-9]_compiled." symbol. It was
>> only ever emitted by GCC v2.x which we don't support anymore anyway.
>>
>> This series is a followup to ' [RFC PATCH 0/2] ARM large kernels'
>> (http://thread.gmane.org/gmane.linux.kbuild.devel/13065), but the
>> other issue addressed by patch #2 of that series will be deferred
>> for now.
>>
>> Ard Biesheuvel (2):
>>   Kbuild: kallsyms: ignore veneers emitted by the ARM linker
>>   Kbuild: kallsyms: drop special handling of pre-3.0 GCC symbols
>>
>>  scripts/kallsyms.c | 29 ++++++++++++++++++++---------
>>  1 file changed, 20 insertions(+), 9 deletions(-)
>>
> 
> Ping?

Applied to kbuild.git#kbuild.

Michal

[PATCH 0/2] Kbuild: kallsyms: update suffix ignore logic

[Michal Marek]

Dne 7.4.2015 v 11:02 Ard Biesheuvel napsal(a):
> On 30 March 2015 at 15:20, Ard Biesheuvel <ard.biesheuvel@linaro.org> wrote:
>> Two patches that update the kallsyms ignore logic.
>>
>> Patch #1 adds logic to introduce symbols whose names end in (or equal)
>> "_veneer". This prevent kallsyms generation errors on large ARM kernels
>> where the emitted veneers will be different between the first and second
>> pass, due to the fact that the size of the kallsyms region itself pushes
>> the .text and .init.text sections further apart during the second pass.
>>
>> Patch #2 removes the handling of the "gcc[0-9]_compiled." symbol. It was
>> only ever emitted by GCC v2.x which we don't support anymore anyway.
>>
>> This series is a followup to ' [RFC PATCH 0/2] ARM large kernels'
>> (http://thread.gmane.org/gmane.linux.kbuild.devel/13065), but the
>> other issue addressed by patch #2 of that series will be deferred
>> for now.
>>
>> Ard Biesheuvel (2):
>>   Kbuild: kallsyms: ignore veneers emitted by the ARM linker
>>   Kbuild: kallsyms: drop special handling of pre-3.0 GCC symbols
>>
>>  scripts/kallsyms.c | 29 ++++++++++++++++++++---------
>>  1 file changed, 20 insertions(+), 9 deletions(-)
>>
> 
> Ping?

Applied to kbuild.git#kbuild.

Michal