[PATCH] Revert "perf tools: Allow to specify custom linker command"

[PATCH] Revert "perf tools: Allow to specify custom linker command"

[Alexey Brodkin]

This reverts commit 5ef7bbb09f7b
("perf tools: Allow to specify custom linker command").

LD is a pre-defined variable in GNU Make. I.e. it is always defined.
Which means there's no point to check "LD ?= ..." because it will never
succeed. And so LD will be either that explicitly passed to make like
this:
 ------->8-------
 make LD=path_to_my_ld ...
 ------->8-------
 or default value, which is host's "ld".

Latter leads to failure of cross-linkage because instead of cross linker
"$(CROSS_COMPILE)ld" host's "ld" is used.

As for commit which is reverted here:
 [1] Usually for selection of non-default flavour of CPU core/options
     linker flags are used like "-mtune=xxx" or "-mMyCPUType" etc.

 [2] Still to implement ability to use "ld" that differs from
    "$(CROSS_COMPILE)ld" one will need to add new makefile variable like
    TARGET_LD and then check if $(TARGET_LD) is not specified on make
    invocation then use "$(CROSS_COMPILE)ld".

But for now to fix cross-building of perf this revert is enough.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Vineet Gupta <vgupta@synopsys.com>
Cc: Aaro Koskinen <aaro.koskinen@nokia.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: linux-kernel@vger.kernel.org
---
 tools/perf/Makefile.perf | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/perf/Makefile.perf b/tools/perf/Makefile.perf
index 7a4b549..0e0938a 100644
--- a/tools/perf/Makefile.perf
+++ b/tools/perf/Makefile.perf
@@ -110,7 +110,7 @@ $(OUTPUT)PERF-VERSION-FILE: ../../.git/HEAD
 	$(Q)touch $(OUTPUT)PERF-VERSION-FILE
 
 CC = $(CROSS_COMPILE)gcc
-LD ?= $(CROSS_COMPILE)ld
+LD = $(CROSS_COMPILE)ld
 AR = $(CROSS_COMPILE)ar
 PKG_CONFIG = $(CROSS_COMPILE)pkg-config
 
-- 
2.4.3

Re: [PATCH] Revert "perf tools: Allow to specify custom linker command"

[Aaro Koskinen]

Hi,

On Mon, Jul 13, 2015 at 02:10:53PM +0300, Alexey Brodkin wrote:
> This reverts commit 5ef7bbb09f7b
> ("perf tools: Allow to specify custom linker command").
> 
> LD is a pre-defined variable in GNU Make. I.e. it is always defined.
> Which means there's no point to check "LD ?= ..." because it will never
> succeed. And so LD will be either that explicitly passed to make like
> this:
>  ------->8-------
>  make LD=path_to_my_ld ...
>  ------->8-------
>  or default value, which is host's "ld".
> 
> Latter leads to failure of cross-linkage because instead of cross linker
> "$(CROSS_COMPILE)ld" host's "ld" is used.
> 
> As for commit which is reverted here:
>  [1] Usually for selection of non-default flavour of CPU core/options
>      linker flags are used like "-mtune=xxx" or "-mMyCPUType" etc.
> 
>  [2] Still to implement ability to use "ld" that differs from
>     "$(CROSS_COMPILE)ld" one will need to add new makefile variable like
>     TARGET_LD and then check if $(TARGET_LD) is not specified on make
>     invocation then use "$(CROSS_COMPILE)ld".
> 
> But for now to fix cross-building of perf this revert is enough.

Hmm. You are probably right, my build system always exports LD with
correct linker for cross builds so perhaps that's why I this "worked"
for me when testing. Sorry.

I guess the correct fix would be [1], i.e. there should be some new
variable to pass flags to ld command.

Or maybe [2], could we use make "origin" function? If LD is "default",
then use "$(CROSS_COMPILE)ld", otherwise use what the user passed?

A.

Re: [PATCH] Revert "perf tools: Allow to specify custom linker command"

[Aaro Koskinen]

Hi,

On Mon, Jul 13, 2015 at 02:10:53PM +0300, Alexey Brodkin wrote:
> This reverts commit 5ef7bbb09f7b
> ("perf tools: Allow to specify custom linker command").
> 
> LD is a pre-defined variable in GNU Make. I.e. it is always defined.
> Which means there's no point to check "LD ?= ..." because it will never
> succeed. And so LD will be either that explicitly passed to make like
> this:
>  ------->8-------
>  make LD=path_to_my_ld ...
>  ------->8-------
>  or default value, which is host's "ld".
> 
> Latter leads to failure of cross-linkage because instead of cross linker
> "$(CROSS_COMPILE)ld" host's "ld" is used.
> 
> As for commit which is reverted here:
>  [1] Usually for selection of non-default flavour of CPU core/options
>      linker flags are used like "-mtune=xxx" or "-mMyCPUType" etc.
> 
>  [2] Still to implement ability to use "ld" that differs from
>     "$(CROSS_COMPILE)ld" one will need to add new makefile variable like
>     TARGET_LD and then check if $(TARGET_LD) is not specified on make
>     invocation then use "$(CROSS_COMPILE)ld".
> 
> But for now to fix cross-building of perf this revert is enough.

Hmm. You are probably right, my build system always exports LD with
correct linker for cross builds so perhaps that's why I this "worked"
for me when testing. Sorry.

I guess the correct fix would be [1], i.e. there should be some new
variable to pass flags to ld command.

Or maybe [2], could we use make "origin" function? If LD is "default",
then use "$(CROSS_COMPILE)ld", otherwise use what the user passed?

A.

Re: [PATCH] Revert "perf tools: Allow to specify custom linker command"

[Jiri Olsa]

On Mon, Jul 13, 2015 at 11:24:06PM +0300, Aaro Koskinen wrote:
> Hi,
> 
> On Mon, Jul 13, 2015 at 02:10:53PM +0300, Alexey Brodkin wrote:
> > This reverts commit 5ef7bbb09f7b
> > ("perf tools: Allow to specify custom linker command").
> > 
> > LD is a pre-defined variable in GNU Make. I.e. it is always defined.
> > Which means there's no point to check "LD ?= ..." because it will never
> > succeed. And so LD will be either that explicitly passed to make like
> > this:
> >  ------->8-------
> >  make LD=path_to_my_ld ...
> >  ------->8-------
> >  or default value, which is host's "ld".
> > 
> > Latter leads to failure of cross-linkage because instead of cross linker
> > "$(CROSS_COMPILE)ld" host's "ld" is used.
> > 
> > As for commit which is reverted here:
> >  [1] Usually for selection of non-default flavour of CPU core/options
> >      linker flags are used like "-mtune=xxx" or "-mMyCPUType" etc.
> > 
> >  [2] Still to implement ability to use "ld" that differs from
> >     "$(CROSS_COMPILE)ld" one will need to add new makefile variable like
> >     TARGET_LD and then check if $(TARGET_LD) is not specified on make
> >     invocation then use "$(CROSS_COMPILE)ld".
> > 
> > But for now to fix cross-building of perf this revert is enough.
> 
> Hmm. You are probably right, my build system always exports LD with
> correct linker for cross builds so perhaps that's why I this "worked"
> for me when testing. Sorry.
> 
> I guess the correct fix would be [1], i.e. there should be some new
> variable to pass flags to ld command.
> 
> Or maybe [2], could we use make "origin" function? If LD is "default",
> then use "$(CROSS_COMPILE)ld", otherwise use what the user passed?

thats what Steven did in traceevent/Makefile:

---
# Makefiles suck: This macro sets a default value of $(2) for the
# variable named by $(1), unless the variable has been set by
# environment or command line. This is necessary for CC and AR
# because make sets default values, so the simpler ?= approach
# won't work as expected.
define allow-override
  $(if $(or $(findstring environment,$(origin $(1))),\
            $(findstring command line,$(origin $(1)))),,\
    $(eval $(1) = $(2)))
endef

# Allow setting CC and AR, or setting CROSS_COMPILE as a prefix.
$(call allow-override,CC,$(CROSS_COMPILE)gcc)
$(call allow-override,AR,$(CROSS_COMPILE)ar)
$(call allow-override,NM,$(CROSS_COMPILE)nm)
---

jirka

[PATCH] perf tools: Really allow to specify custom CC, AR or LD

[Alexey Brodkin]

Commit 5ef7bbb09f7b ("perf tools: Allow to specify custom linker command")
was meant to enable usage non $(CROSS_COMPILE)ld linker during perf
building. But implementation didn't take in account fact that LD is
a pre-defined variable in GNU Make. I.e. it is always defined.
Which means there's no point to check "LD ?= ..." because it will never
succeed. And so LD will be either that explicitly passed to make like
this:
 ------->8-------
 make LD=path_to_my_ld ...
 ------->8-------
 or default value, which is host's "ld".

Latter leads to failure of cross-linkage because instead of cross linker
"$(CROSS_COMPILE)ld" host's "ld" is used.

Fortunately there's a way to do correct substitution of $(CROSS_COMPILE)ld
with user defined LD on command-line.

As a reference was used implementation in "tools/lib/traceevent/Makefile".

Build tested for x86_64 and ARC.

Thanks Jiri for this hint.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Vineet Gupta <vgupta@synopsys.com>
Cc: Aaro Koskinen <aaro.koskinen@nokia.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: linux-kernel@vger.kernel.org
---
 tools/perf/Makefile.perf | 19 ++++++++++++++++---
 1 file changed, 16 insertions(+), 3 deletions(-)

diff --git a/tools/perf/Makefile.perf b/tools/perf/Makefile.perf
index 7a4b549..bba3463 100644
--- a/tools/perf/Makefile.perf
+++ b/tools/perf/Makefile.perf
@@ -109,9 +109,22 @@ $(OUTPUT)PERF-VERSION-FILE: ../../.git/HEAD
 	$(Q)$(SHELL_PATH) util/PERF-VERSION-GEN $(OUTPUT)
 	$(Q)touch $(OUTPUT)PERF-VERSION-FILE
 
-CC = $(CROSS_COMPILE)gcc
-LD ?= $(CROSS_COMPILE)ld
-AR = $(CROSS_COMPILE)ar
+# Makefiles suck: This macro sets a default value of $(2) for the
+# variable named by $(1), unless the variable has been set by
+# environment or command line. This is necessary for CC and AR
+# because make sets default values, so the simpler ?= approach
+# won't work as expected.
+define allow-override
+  $(if $(or $(findstring environment,$(origin $(1))),\
+            $(findstring command line,$(origin $(1)))),,\
+    $(eval $(1) = $(2)))
+endef
+
+# Allow setting CC and AR and LD, or setting CROSS_COMPILE as a prefix.
+$(call allow-override,CC,$(CROSS_COMPILE)gcc)
+$(call allow-override,AR,$(CROSS_COMPILE)ar)
+$(call allow-override,LD,$(CROSS_COMPILE)ld)
+
 PKG_CONFIG = $(CROSS_COMPILE)pkg-config
 
 RM      = rm -f
-- 
2.4.3

Re: [PATCH] Revert "perf tools: Allow to specify custom linker command"

[Arnaldo Carvalho de Melo]

Em Mon, Jul 13, 2015 at 11:24:06PM +0300, Aaro Koskinen escreveu:
> Hi,
> 
> On Mon, Jul 13, 2015 at 02:10:53PM +0300, Alexey Brodkin wrote:
> > This reverts commit 5ef7bbb09f7b
> > ("perf tools: Allow to specify custom linker command").
> > 
> > LD is a pre-defined variable in GNU Make. I.e. it is always defined.
> > Which means there's no point to check "LD ?= ..." because it will never
> > succeed. And so LD will be either that explicitly passed to make like
> > this:
> >  ------->8-------
> >  make LD=path_to_my_ld ...
> >  ------->8-------
> >  or default value, which is host's "ld".
> > 
> > Latter leads to failure of cross-linkage because instead of cross linker
> > "$(CROSS_COMPILE)ld" host's "ld" is used.
> > 
> > As for commit which is reverted here:
> >  [1] Usually for selection of non-default flavour of CPU core/options
> >      linker flags are used like "-mtune=xxx" or "-mMyCPUType" etc.
> > 
> >  [2] Still to implement ability to use "ld" that differs from
> >     "$(CROSS_COMPILE)ld" one will need to add new makefile variable like
> >     TARGET_LD and then check if $(TARGET_LD) is not specified on make
> >     invocation then use "$(CROSS_COMPILE)ld".
> > 
> > But for now to fix cross-building of perf this revert is enough.
> 
> Hmm. You are probably right, my build system always exports LD with
> correct linker for cross builds so perhaps that's why I this "worked"
> for me when testing. Sorry.

So, I think this is an "Acked-by: Aaro Koskinen <aaro.koskinen@iki.fi>"
for the revert, right?
 
> I guess the correct fix would be [1], i.e. there should be some new
> variable to pass flags to ld command.
> 
> Or maybe [2], could we use make "origin" function? If LD is "default",
> then use "$(CROSS_COMPILE)ld", otherwise use what the user passed?

When you guys get to some conclusion, please submit a new patch, for now
I am taking the revert with the (implied) Acked-by Aaro, ok?

- Arnaldo

Re: [PATCH] Revert "perf tools: Allow to specify custom linker command"

[Alexey Brodkin]

Hi Arnaldo,

On Tue, 2015-07-14 at 10:46 -0300, Arnaldo Carvalho de Melo wrote:
> Em Mon, Jul 13, 2015 at 11:24:06PM +0300, Aaro Koskinen escreveu:
> > Hi,
> > 
> > On Mon, Jul 13, 2015 at 02:10:53PM +0300, Alexey Brodkin wrote:
> > > This reverts commit 5ef7bbb09f7b
> > > ("perf tools: Allow to specify custom linker command").
> > > 
> > > LD is a pre-defined variable in GNU Make. I.e. it is always defined.
> > > Which means there's no point to check "LD ?= ..." because it will never
> > > succeed. And so LD will be either that explicitly passed to make like
> > > this:
> > >  ------->8-------
> > >  make LD=path_to_my_ld ...
> > >  ------->8-------
> > >  or default value, which is host's "ld".
> > > 
> > > Latter leads to failure of cross-linkage because instead of cross linker
> > > "$(CROSS_COMPILE)ld" host's "ld" is used.
> > > 
> > > As for commit which is reverted here:
> > >  [1] Usually for selection of non-default flavour of CPU core/options
> > >      linker flags are used like "-mtune=xxx" or "-mMyCPUType" etc.
> > > 
> > >  [2] Still to implement ability to use "ld" that differs from
> > >     "$(CROSS_COMPILE)ld" one will need to add new makefile variable like
> > >     TARGET_LD and then check if $(TARGET_LD) is not specified on make
> > >     invocation then use "$(CROSS_COMPILE)ld".
> > > 
> > > But for now to fix cross-building of perf this revert is enough.
> > 
> > Hmm. You are probably right, my build system always exports LD with
> > correct linker for cross builds so perhaps that's why I this "worked"
> > for me when testing. Sorry.
> 
> So, I think this is an "Acked-by: Aaro Koskinen <aaro.koskinen@iki.fi>"
> for the revert, right?
>  
> > I guess the correct fix would be [1], i.e. there should be some new
> > variable to pass flags to ld command.
> > 
> > Or maybe [2], could we use make "origin" function? If LD is "default",
> > then use "$(CROSS_COMPILE)ld", otherwise use what the user passed?
> 
> When you guys get to some conclusion, please submit a new patch, for now
> I am taking the revert with the (implied) Acked-by Aaro, ok?

Please find my patch that both reverts initial Aaro's patch and implements
proper handling of CC, AR and LD substitution here - https://lkml.org/lkml/2015/7/14/149

-Alexey

Re: [PATCH] Revert "perf tools: Allow to specify custom linker command"

[acme]

Em Tue, Jul 14, 2015 at 02:00:05PM +0000, Alexey Brodkin escreveu:
> On Tue, 2015-07-14 at 10:46 -0300, Arnaldo Carvalho de Melo wrote:
> > Em Mon, Jul 13, 2015 at 11:24:06PM +0300, Aaro Koskinen escreveu:
> > > > But for now to fix cross-building of perf this revert is enough.

> > > Hmm. You are probably right, my build system always exports LD with
> > > correct linker for cross builds so perhaps that's why I this "worked"
> > > for me when testing. Sorry.

> > So, I think this is an "Acked-by: Aaro Koskinen <aaro.koskinen@iki.fi>"
> > for the revert, right?

> > > I guess the correct fix would be [1], i.e. there should be some new
> > > variable to pass flags to ld command.

> > > Or maybe [2], could we use make "origin" function? If LD is "default",
> > > then use "$(CROSS_COMPILE)ld", otherwise use what the user passed?

> > When you guys get to some conclusion, please submit a new patch, for now
> > I am taking the revert with the (implied) Acked-by Aaro, ok?

> Please find my patch that both reverts initial Aaro's patch and implements
> proper handling of CC, AR and LD substitution here - https://lkml.org/lkml/2015/7/14/149

I saw it, will do.

- Arnaldo