| Message ID | 1465823125-171114-1-git-send-email-arnd@arndb.de (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Mon, Jun 13, 2016 at 3:04 PM, Arnd Bergmann <arnd@arndb.de> wrote: > The genksyms helper in the kernel cannot parse a type definition > like "typeof(((type *)0)->keyfld)" that is used in the DEFINE_RB_FUNCS > helper, causing the following EXPORT_SYMBOL() statement to be ignored > when computing the crcs, and triggering a warning about this: > > WARNING: "ceph_monc_do_statfs" [fs/ceph/ceph.ko] has no CRC > > To work around the problem, we can rewrite the type to reference > an undefined 'extern' symbol instead of a NULL pointer. This is > evidently ok for genksyms, and it no longer complains about the > line when calling it with 'genksyms -w'. > > I've looked briefly into extending genksyms instead, but it seems > really hard to do. Jan Beulich introduced basic support for 'typeof' > a while ago in dc53324060f3 ("genksyms: fix typeof() handling"), > but that is not sufficient for the expression we have here. > > Signed-off-by: Arnd Bergmann <arnd@arndb.de> > Fixes: fcd00b68bbe2 ("libceph: DEFINE_RB_FUNCS macro") > Cc: Jan Beulich <jbeulich@suse.com> > Cc: Michal Marek <mmarek@suse.cz> > --- > include/linux/ceph/libceph.h | 3 ++- > 1 file changed, 2 insertions(+), 1 deletion(-) > > diff --git a/include/linux/ceph/libceph.h b/include/linux/ceph/libceph.h > index 690985daad1c..6b79a6ba39ca 100644 > --- a/include/linux/ceph/libceph.h > +++ b/include/linux/ceph/libceph.h > @@ -214,8 +214,9 @@ static void erase_##name(struct rb_root *root, type *t) \ > } > > #define DEFINE_RB_LOOKUP_FUNC(name, type, keyfld, nodefld) \ > +extern type __lookup_##name##_key; \ > static type *lookup_##name(struct rb_root *root, \ > - typeof(((type *)0)->keyfld) key) \ > + typeof(__lookup_##name##_key.keyfld) key) \ > { \ > struct rb_node *n = root->rb_node; \ > \ Out of curiosity, did you figure out why is it that only ceph_monc_do_statfs() is affected and not the other exports in net/ceph/osd_client.c? Thanks, Ilya -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Monday, June 13, 2016 3:19:12 PM CEST Ilya Dryomov wrote: > > diff --git a/include/linux/ceph/libceph.h b/include/linux/ceph/libceph.h > > index 690985daad1c..6b79a6ba39ca 100644 > > --- a/include/linux/ceph/libceph.h > > +++ b/include/linux/ceph/libceph.h > > @@ -214,8 +214,9 @@ static void erase_##name(struct rb_root *root, type *t) \ > > } > > > > #define DEFINE_RB_LOOKUP_FUNC(name, type, keyfld, nodefld) \ > > +extern type __lookup_##name##_key; \ > > static type *lookup_##name(struct rb_root *root, \ > > - typeof(((type *)0)->keyfld) key) \ > > + typeof(__lookup_##name##_key.keyfld) key) \ > > { \ > > struct rb_node *n = root->rb_node; \ > > \ > > Out of curiosity, did you figure out why is it that only > ceph_monc_do_statfs() is affected and not the other exports in > net/ceph/osd_client.c? No, I did not investigate it further at that point. My guess is that something resets the state of the genksyms parser when it encounters some tokens (like the semicolon I added earlier, or an EXPORT_SYMBOL statement), and that this happens in one file but not the other. Arnd -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 2016-06-13 15:04, Arnd Bergmann wrote: > The genksyms helper in the kernel cannot parse a type definition > like "typeof(((type *)0)->keyfld)" that is used in the DEFINE_RB_FUNCS > helper, causing the following EXPORT_SYMBOL() statement to be ignored > when computing the crcs, and triggering a warning about this: > > WARNING: "ceph_monc_do_statfs" [fs/ceph/ceph.ko] has no CRC FYI, Nick fixed the genksyms state machine to cope with this syntax. See commits 5c6f3225d00d kbuild: Regenerate genksyms lexer 4fab91605a6b kbuild: genksyms fix for typeof handling in today's linux-next. So the workaround can eventually be reverted. Michal -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/include/linux/ceph/libceph.h b/include/linux/ceph/libceph.h index 690985daad1c..6b79a6ba39ca 100644 --- a/include/linux/ceph/libceph.h +++ b/include/linux/ceph/libceph.h @@ -214,8 +214,9 @@ static void erase_##name(struct rb_root *root, type *t) \ } #define DEFINE_RB_LOOKUP_FUNC(name, type, keyfld, nodefld) \ +extern type __lookup_##name##_key; \ static type *lookup_##name(struct rb_root *root, \ - typeof(((type *)0)->keyfld) key) \ + typeof(__lookup_##name##_key.keyfld) key) \ { \ struct rb_node *n = root->rb_node; \ \
The genksyms helper in the kernel cannot parse a type definition like "typeof(((type *)0)->keyfld)" that is used in the DEFINE_RB_FUNCS helper, causing the following EXPORT_SYMBOL() statement to be ignored when computing the crcs, and triggering a warning about this: WARNING: "ceph_monc_do_statfs" [fs/ceph/ceph.ko] has no CRC To work around the problem, we can rewrite the type to reference an undefined 'extern' symbol instead of a NULL pointer. This is evidently ok for genksyms, and it no longer complains about the line when calling it with 'genksyms -w'. I've looked briefly into extending genksyms instead, but it seems really hard to do. Jan Beulich introduced basic support for 'typeof' a while ago in dc53324060f3 ("genksyms: fix typeof() handling"), but that is not sufficient for the expression we have here. Signed-off-by: Arnd Bergmann <arnd@arndb.de> Fixes: fcd00b68bbe2 ("libceph: DEFINE_RB_FUNCS macro") Cc: Jan Beulich <jbeulich@suse.com> Cc: Michal Marek <mmarek@suse.cz> --- include/linux/ceph/libceph.h | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)