| Message ID | 1526020226-13904-1-git-send-email-vladimir.mezentsev@oracle.com |
|---|---|
| State | New |
| Headers | show |
| Series | PR gcc/84923 - gcc.dg/attr-weakref-1.c failed on aarch64 | expand |
Ping. -Vladimir On 05/10/2018 11:30 PM, vladimir.mezentsev@oracle.com wrote: > From: Vladimir Mezentsev <vladimir.mezentsev@oracle.com> > > When weakref_targets is not empty a target cannot be removed from the weak list. > A small example is below when 'wv12' is removed from the weak list on aarch64: > static vtype Wv12 __attribute__((weakref ("wv12"))); > extern vtype wv12 __attribute__((weak)); > > Bootstrapped on aarch64-unknown-linux-gnu including (c,c++ and go). > Tested on aarch64-linux-gnu. > No regression. The attr-weakref-1.c test passed. > > ChangeLog: > 2018-05-10 Vladimir Mezentsev <vladimir.mezentsev@oracle.com> > > PR gcc/84923 > * varasm.c (weak_finish): clean up weak_decls > --- > gcc/varasm.c | 3 ++- > 1 file changed, 2 insertions(+), 1 deletion(-) > > diff --git a/gcc/varasm.c b/gcc/varasm.c > index 85296b4..8cf6e1e 100644 > --- a/gcc/varasm.c > +++ b/gcc/varasm.c > @@ -5652,7 +5652,8 @@ weak_finish (void) > tree alias_decl = TREE_PURPOSE (t); > tree target = ultimate_transparent_alias_target (&TREE_VALUE (t)); > > - if (! TREE_SYMBOL_REFERENCED (DECL_ASSEMBLER_NAME (alias_decl))) > + if (! TREE_SYMBOL_REFERENCED (DECL_ASSEMBLER_NAME (alias_decl)) > + || TREE_SYMBOL_REFERENCED (target)) > /* Remove alias_decl from the weak list, but leave entries for > the target alone. */ > target = NULL_TREE;
Hi, Given this is a midend change it's a good idea to CC some of the maintainers of that area. I've copied richi and Honza. Thanks, Kyrill On 17/05/18 05:35, vladimir.mezentsev@oracle.com wrote: > Ping. > > -Vladimir > > > On 05/10/2018 11:30 PM, vladimir.mezentsev@oracle.com wrote: > > From: Vladimir Mezentsev <vladimir.mezentsev@oracle.com> > > > > When weakref_targets is not empty a target cannot be removed from the weak list. > > A small example is below when 'wv12' is removed from the weak list on aarch64: > > static vtype Wv12 __attribute__((weakref ("wv12"))); > > extern vtype wv12 __attribute__((weak)); > > > > Bootstrapped on aarch64-unknown-linux-gnu including (c,c++ and go). > > Tested on aarch64-linux-gnu. > > No regression. The attr-weakref-1.c test passed. > > > > ChangeLog: > > 2018-05-10 Vladimir Mezentsev <vladimir.mezentsev@oracle.com> > > > > PR gcc/84923 > > * varasm.c (weak_finish): clean up weak_decls > > --- > > gcc/varasm.c | 3 ++- > > 1 file changed, 2 insertions(+), 1 deletion(-) > > > > diff --git a/gcc/varasm.c b/gcc/varasm.c > > index 85296b4..8cf6e1e 100644 > > --- a/gcc/varasm.c > > +++ b/gcc/varasm.c > > @@ -5652,7 +5652,8 @@ weak_finish (void) > > tree alias_decl = TREE_PURPOSE (t); > > tree target = ultimate_transparent_alias_target (&TREE_VALUE (t)); > > > > - if (! TREE_SYMBOL_REFERENCED (DECL_ASSEMBLER_NAME (alias_decl))) > > + if (! TREE_SYMBOL_REFERENCED (DECL_ASSEMBLER_NAME (alias_decl)) > > + || TREE_SYMBOL_REFERENCED (target)) > > /* Remove alias_decl from the weak list, but leave entries for > > the target alone. */ > > target = NULL_TREE; >
On Thu, 17 May 2018, Kyrill Tkachov wrote: > Hi, > > Given this is a midend change it's a good idea to CC some of the maintainers > of that area. > I've copied richi and Honza. The patch is ok for trunk (it's actually mine...) and for the branch after a while. Thanks, Richard. > Thanks, > Kyrill > > On 17/05/18 05:35, vladimir.mezentsev@oracle.com wrote: > > Ping. > > > > -Vladimir > > > > > > On 05/10/2018 11:30 PM, vladimir.mezentsev@oracle.com wrote: > > > From: Vladimir Mezentsev <vladimir.mezentsev@oracle.com> > > > > > > When weakref_targets is not empty a target cannot be removed from the weak > > list. > > > A small example is below when 'wv12' is removed from the weak list on > > aarch64: > > > static vtype Wv12 __attribute__((weakref ("wv12"))); > > > extern vtype wv12 __attribute__((weak)); > > > > > > Bootstrapped on aarch64-unknown-linux-gnu including (c,c++ and go). > > > Tested on aarch64-linux-gnu. > > > No regression. The attr-weakref-1.c test passed. > > > > > > ChangeLog: > > > 2018-05-10 Vladimir Mezentsev <vladimir.mezentsev@oracle.com> > > > > > > PR gcc/84923 > > > * varasm.c (weak_finish): clean up weak_decls > > > --- > > > gcc/varasm.c | 3 ++- > > > 1 file changed, 2 insertions(+), 1 deletion(-) > > > > > > diff --git a/gcc/varasm.c b/gcc/varasm.c > > > index 85296b4..8cf6e1e 100644 > > > --- a/gcc/varasm.c > > > +++ b/gcc/varasm.c > > > @@ -5652,7 +5652,8 @@ weak_finish (void) > > > tree alias_decl = TREE_PURPOSE (t); > > > tree target = ultimate_transparent_alias_target (&TREE_VALUE (t)); > > > > > > - if (! TREE_SYMBOL_REFERENCED (DECL_ASSEMBLER_NAME (alias_decl))) > > > + if (! TREE_SYMBOL_REFERENCED (DECL_ASSEMBLER_NAME (alias_decl)) > > > + || TREE_SYMBOL_REFERENCED (target)) > > > /* Remove alias_decl from the weak list, but leave entries for > > > the target alone. */ > > > target = NULL_TREE; > > > >
On 05/17/2018 02:42 AM, Richard Biener wrote: > On Thu, 17 May 2018, Kyrill Tkachov wrote: > >> Hi, >> >> Given this is a midend change it's a good idea to CC some of the maintainers >> of that area. >> I've copied richi and Honza. > > The patch is ok for trunk (it's actually mine...) and for the branch > after a while. Installed. Vladimir -- do you have write access? ISTM it would make sense for you to have access if you plan to continue submitting changes. Jeff
On 05/21/2018 01:30 PM, Jeff Law wrote: > On 05/17/2018 02:42 AM, Richard Biener wrote: >> On Thu, 17 May 2018, Kyrill Tkachov wrote: >> >>> Hi, >>> >>> Given this is a midend change it's a good idea to CC some of the maintainers >>> of that area. >>> I've copied richi and Honza. >> The patch is ok for trunk (it's actually mine...) and for the branch >> after a while. > Installed. > > Vladimir -- do you have write access? No. I don't. > ISTM it would make sense for you > to have access if you plan to continue submitting changes. yes. I plan to continue submitting changes. -Vladimir > > Jeff
diff --git a/gcc/varasm.c b/gcc/varasm.c index 85296b4..8cf6e1e 100644 --- a/gcc/varasm.c +++ b/gcc/varasm.c @@ -5652,7 +5652,8 @@ weak_finish (void) tree alias_decl = TREE_PURPOSE (t); tree target = ultimate_transparent_alias_target (&TREE_VALUE (t)); - if (! TREE_SYMBOL_REFERENCED (DECL_ASSEMBLER_NAME (alias_decl))) + if (! TREE_SYMBOL_REFERENCED (DECL_ASSEMBLER_NAME (alias_decl)) + || TREE_SYMBOL_REFERENCED (target)) /* Remove alias_decl from the weak list, but leave entries for the target alone. */ target = NULL_TREE;
From: Vladimir Mezentsev <vladimir.mezentsev@oracle.com> When weakref_targets is not empty a target cannot be removed from the weak list. A small example is below when 'wv12' is removed from the weak list on aarch64: static vtype Wv12 __attribute__((weakref ("wv12"))); extern vtype wv12 __attribute__((weak)); Bootstrapped on aarch64-unknown-linux-gnu including (c,c++ and go). Tested on aarch64-linux-gnu. No regression. The attr-weakref-1.c test passed. ChangeLog: 2018-05-10 Vladimir Mezentsev <vladimir.mezentsev@oracle.com> PR gcc/84923 * varasm.c (weak_finish): clean up weak_decls --- gcc/varasm.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)