diff mbox series

[1/1] woff2: fix build on m68k

Message ID 20180510101101.22094-1-fontaine.fabrice@gmail.com
State Accepted
Headers show
Series [1/1] woff2: fix build on m68k | expand

Commit Message

Fabrice Fontaine May 10, 2018, 10:11 a.m. UTC
Same issue on woff2 than in some other packages such as lcdapi: see
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=79509

Fixes:
 - http://autobuild.buildroot.net/results/6c97a08e6ca2e914d54fbb525b3f0bb90023ef07

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
---
 package/woff2/woff2.mk | 6 ++++++
 1 file changed, 6 insertions(+)

Comments

Yann E. MORIN May 13, 2018, 3:14 p.m. UTC | #1
Fabrice, All,

On 2018-05-10 12:11 +0200, Fabrice Fontaine spake thusly:
> Same issue on woff2 than in some other packages such as lcdapi: see
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=79509

Since we now have at least 5 packages affected by this bug, can't we
just include this into the CFLAGS uncoditionally, i.e.:

    diff --git a/package/Makefile.in b/package/Makefile.in
    index 4325f7b3a9..584c7fafb1 100644
    --- a/package/Makefile.in
    +++ b/package/Makefile.in
    @@ -173,6 +173,12 @@ TARGET_CFLAGS = $(TARGET_CPPFLAGS) $(TARGET_ABI) $(TARGET_OPTIMIZATION) $(TARGET
     TARGET_CXXFLAGS = $(TARGET_CFLAGS)
     TARGET_FCFLAGS = $(TARGET_ABI) $(TARGET_OPTIMIZATION) $(TARGET_DEBUGGING)
     
    +# https://gcc.gnu.org/bugzilla/show_bug.cgi?id=79509
    +ifeq ($(BR2_m68k_cf),y)
    +TARGET_CFLAGS += -fno-dwarf2-cfi-asm
    +TARGET_CXXFLAGS += -fno-dwarf2-cfi-asm
    +endif
    +
     ifeq ($(BR2_BINFMT_FLAT),y)
     TARGET_CFLAGS += $(if $($(PKG)_FLAT_STACKSIZE),-Wl$(comma)-elf2flt=-s$($(PKG)_FLAT_STACKSIZE),\
        -Wl$(comma)-elf2flt)

...with the workaround removed from the 4 affected packages: berkeleydb,
kmsxx, lcdapi, and libasplib?

Regards,
Yann E. MORIN.

> Fixes:
>  - http://autobuild.buildroot.net/results/6c97a08e6ca2e914d54fbb525b3f0bb90023ef07
> 
> Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
> ---
>  package/woff2/woff2.mk | 6 ++++++
>  1 file changed, 6 insertions(+)
> 
> diff --git a/package/woff2/woff2.mk b/package/woff2/woff2.mk
> index b2ff33fe98..2f0206c82b 100644
> --- a/package/woff2/woff2.mk
> +++ b/package/woff2/woff2.mk
> @@ -21,4 +21,10 @@ ifneq ($(BR2_SHARED_LIBS),y)
>  WOFF2_CONF_OPTS += -DCMAKE_SKIP_RPATH=ON
>  endif
>  
> +# Internal error, aborting at dw2gencfi.c:215 in emit_expr_encoded
> +# https://gcc.gnu.org/bugzilla/show_bug.cgi?id=79509
> +ifeq ($(BR2_m68k_cf),y)
> +WOFF2_CONF_OPTS += -DCMAKE_CXX_FLAGS="$(TARGET_CXXFLAGS) -fno-dwarf2-cfi-asm"
> +endif
> +
>  $(eval $(cmake-package))
> -- 
> 2.14.1
> 
> _______________________________________________
> buildroot mailing list
> buildroot@busybox.net
> http://lists.busybox.net/mailman/listinfo/buildroot
Thomas Petazzoni May 13, 2018, 8:08 p.m. UTC | #2
Hello,

+Waldemar in Cc, since he reported this issue to gcc.

On Thu, 10 May 2018 12:11:01 +0200, Fabrice Fontaine wrote:
> Same issue on woff2 than in some other packages such as lcdapi: see
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=79509
> 
> Fixes:
>  - http://autobuild.buildroot.net/results/6c97a08e6ca2e914d54fbb525b3f0bb90023ef07
> 
> Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
> ---
>  package/woff2/woff2.mk | 6 ++++++
>  1 file changed, 6 insertions(+)

I've applied to master, but as Yann pointed out, it starts to be a
problem affecting a number of packages.

The bug report you point out at
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=79509 has been closed,
because they say the issue is on binutils side. Has the issue been
reported to binutils ? Could you try to make this issue move forward ?

Thomas
Waldemar Brodkorb May 14, 2018, 5:19 a.m. UTC | #3
Hi,
Thomas Petazzoni wrote,

> Hello,
> 
> +Waldemar in Cc, since he reported this issue to gcc.
> 
> On Thu, 10 May 2018 12:11:01 +0200, Fabrice Fontaine wrote:
> > Same issue on woff2 than in some other packages such as lcdapi: see
> > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=79509
> > 
> > Fixes:
> >  - http://autobuild.buildroot.net/results/6c97a08e6ca2e914d54fbb525b3f0bb90023ef07
> > 
> > Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
> > ---
> >  package/woff2/woff2.mk | 6 ++++++
> >  1 file changed, 6 insertions(+)
> 
> I've applied to master, but as Yann pointed out, it starts to be a
> problem affecting a number of packages.
> 
> The bug report you point out at
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=79509 has been closed,
> because they say the issue is on binutils side. Has the issue been
> reported to binutils ? Could you try to make this issue move forward ?

I recently added the binutils report for this bug on my TODO list,
but if Fabrice have fun to do it, I have'nt started with it, yet.
For binutils/gcc bugs you need to be responsive if you ignite a
fire, otherwise an issue might take 4-6 years to be resolved.  This
is only valid for so popular architectures like m68k or bfin and
hackers not able to provide a patch ;)

best regards
 Waldemar
Fabrice Fontaine May 14, 2018, 8:43 p.m. UTC | #4
Dear all,

2018-05-14 7:19 GMT+02:00 Waldemar Brodkorb <wbx@openadk.org>:

> Hi,
> Thomas Petazzoni wrote,
>
> > Hello,
> >
> > +Waldemar in Cc, since he reported this issue to gcc.
> >
> > On Thu, 10 May 2018 12:11:01 +0200, Fabrice Fontaine wrote:
> > > Same issue on woff2 than in some other packages such as lcdapi: see
> > > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=79509
> > >
> > > Fixes:
> > >  - http://autobuild.buildroot.net/results/6c97a08e6ca2e914d54fb
> b525b3f0bb90023ef07
> > >
> > > Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
> > > ---
> > >  package/woff2/woff2.mk | 6 ++++++
> > >  1 file changed, 6 insertions(+)
> >
> > I've applied to master, but as Yann pointed out, it starts to be a
> > problem affecting a number of packages.
> >
> > The bug report you point out at
> > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=79509 has been closed,
> > because they say the issue is on binutils side. Has the issue been
> > reported to binutils ? Could you try to make this issue move forward ?
>
> I recently added the binutils report for this bug on my TODO list,
> but if Fabrice have fun to do it, I have'nt started with it, yet.
> For binutils/gcc bugs you need to be responsive if you ignite a
> fire, otherwise an issue might take 4-6 years to be resolved.  This
> is only valid for so popular architectures like m68k or bfin and
> hackers not able to provide a patch ;)

I can let you report the bug as you're the one who found the issue as well
as the workaround. I'm not used to report bugs to the binutils folks.
As a side note, the issue was also raised by log4cplus:
*http://autobuild.buildroot.net/results/035/035cc60c0fd44a5a44de64bfadfe3788ef3c6576/build-end.log
<http://autobuild.buildroot.net/results/035/035cc60c0fd44a5a44de64bfadfe3788ef3c6576/build-end.log>*.
However, after applying the patch that removed compilation on toolchains
with the gcc bug 64735, the issue was "hidden".

>


> best regards
>  Waldemar
>
Best Regards,

Fabrice
<div dir="ltr">Dear all,<div class="gmail_extra"><br><div class="gmail_quote">2018-05-14 7:19 GMT+02:00 Waldemar Brodkorb <span dir="ltr">&lt;<a href="mailto:wbx@openadk.org" target="_blank">wbx@openadk.org</a>&gt;</span>:<br><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">Hi,<br>
Thomas Petazzoni wrote,<br>
<span class="m_7433808820669306109gmail-m_-8582897048107033207gmail-"><br>
&gt; Hello,<br>
&gt; <br>
&gt; +Waldemar in Cc, since he reported this issue to gcc.<br>
&gt; <br>
&gt; On Thu, 10 May 2018 12:11:01 +0200, Fabrice Fontaine wrote:<br>
&gt; &gt; Same issue on woff2 than in some other packages such as lcdapi: see<br>
&gt; &gt; <a href="https://gcc.gnu.org/bugzilla/show_bug.cgi?id=79509" rel="noreferrer" target="_blank">https://gcc.gnu.org/bugzilla/s<wbr>how_bug.cgi?id=79509</a><br>
&gt; &gt; <br>
&gt; &gt; Fixes:<br>
&gt; &gt;  - <a href="http://autobuild.buildroot.net/results/6c97a08e6ca2e914d54fbb525b3f0bb90023ef07" rel="noreferrer" target="_blank">http://autobuild.buildroot.net<wbr>/results/6c97a08e6ca2e914d54fb<wbr>b525b3f0bb90023ef07</a><br>
&gt; &gt; <br>
&gt; &gt; Signed-off-by: Fabrice Fontaine &lt;<a href="mailto:fontaine.fabrice@gmail.com" target="_blank">fontaine.fabrice@gmail.com</a>&gt;<br>
&gt; &gt; ---<br>
&gt; &gt;  package/woff2/<a href="http://woff2.mk" rel="noreferrer" target="_blank">woff2.mk</a> | 6 ++++++<br>
&gt; &gt;  1 file changed, 6 insertions(+)<br>
&gt; <br>
&gt; I&#39;ve applied to master, but as Yann pointed out, it starts to be a<br>
&gt; problem affecting a number of packages.<br>
&gt; <br>
&gt; The bug report you point out at<br>
&gt; <a href="https://gcc.gnu.org/bugzilla/show_bug.cgi?id=79509" rel="noreferrer" target="_blank">https://gcc.gnu.org/bugzilla/s<wbr>how_bug.cgi?id=79509</a> has been closed,<br>
&gt; because they say the issue is on binutils side. Has the issue been<br>
&gt; reported to binutils ? Could you try to make this issue move forward ?<br>
<br>
</span>I recently added the binutils report for this bug on my TODO list,<br>
but if Fabrice have fun to do it, I have&#39;nt started with it, yet.<br>
For binutils/gcc bugs you need to be responsive if you ignite a<br>
fire, otherwise an issue might take 4-6 years to be resolved.  This<br>
is only valid for so popular architectures like m68k or bfin and<br>
hackers not able to provide a patch ;)</blockquote><div>I can let you report the bug as you&#39;re the one who found the issue as well as the workaround. I&#39;m not used to report bugs to the binutils folks.</div><div>As a side note, the issue was also raised by log4cplus: <font color="#1155cc"><u><a href="http://autobuild.buildroot.net/results/035/035cc60c0fd44a5a44de64bfadfe3788ef3c6576/build-end.log" target="_blank">http://autobuild.buildroot.<wbr>net/results/035/<wbr>035cc60c0fd44a5a44de64bfadfe37<wbr>88ef3c6576/build-end.log</a></u></font>. However, after applying the patch that removed compilation on toolchains with the gcc bug 64735, the issue was &quot;hidden&quot;.</div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"> </blockquote><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
<br>
best regards<br>
<span class="m_7433808820669306109gmail-m_-8582897048107033207gmail-HOEnZb"><font color="#888888"> Waldemar<br>
</font></span></blockquote></div>Best Regards,</div><div class="gmail_extra"><br></div><div class="gmail_extra">Fabrice</div></div>
diff mbox series

Patch

diff --git a/package/woff2/woff2.mk b/package/woff2/woff2.mk
index b2ff33fe98..2f0206c82b 100644
--- a/package/woff2/woff2.mk
+++ b/package/woff2/woff2.mk
@@ -21,4 +21,10 @@  ifneq ($(BR2_SHARED_LIBS),y)
 WOFF2_CONF_OPTS += -DCMAKE_SKIP_RPATH=ON
 endif
 
+# Internal error, aborting at dw2gencfi.c:215 in emit_expr_encoded
+# https://gcc.gnu.org/bugzilla/show_bug.cgi?id=79509
+ifeq ($(BR2_m68k_cf),y)
+WOFF2_CONF_OPTS += -DCMAKE_CXX_FLAGS="$(TARGET_CXXFLAGS) -fno-dwarf2-cfi-asm"
+endif
+
 $(eval $(cmake-package))