about summary refs log tree commit diff
diff options
context:
space:
mode:
authorRich Felker <dalias@aerifal.cx>2021-04-16 10:20:46 -0400
committerRich Felker <dalias@aerifal.cx>2021-04-16 10:20:46 -0400
commitaad50fcd791e009961621ddfbe3d4c245fd689a3 (patch)
tree84d9494a2123186c7ceeb32a5a90c48bad97c45a
parent0ea78a6421322cab24d448670006ee2f99af3ac9 (diff)
downloadmusl-aad50fcd791e009961621ddfbe3d4c245fd689a3.tar.gz
musl-aad50fcd791e009961621ddfbe3d4c245fd689a3.tar.xz
musl-aad50fcd791e009961621ddfbe3d4c245fd689a3.zip
fix regression in dl_iterate_phdr reporting of modules with no TLS
__tls_get_addr should not be called with an invalid TLS module id of
0. in practice it probably "works", returning the DTV length as if it
were a pointer, and the callback should probably not inspect
dlpi_tls_data in this case, but it's likely that some real-world
callbacks use a check on dlpi_tls_data being non-null, rather than on
dlpi_tls_modid being nonzero, to conclude that the module has TLS.
-rw-r--r--ldso/dynlink.c3
1 files changed, 2 insertions, 1 deletions
diff --git a/ldso/dynlink.c b/ldso/dynlink.c
index b66ad537..8b67ef59 100644
--- a/ldso/dynlink.c
+++ b/ldso/dynlink.c
@@ -2331,7 +2331,8 @@ int dl_iterate_phdr(int(*callback)(struct dl_phdr_info *info, size_t size, void
 		info.dlpi_adds      = gencnt;
 		info.dlpi_subs      = 0;
 		info.dlpi_tls_modid = current->tls_id;
-		info.dlpi_tls_data = __tls_get_addr((tls_mod_off_t[]){current->tls_id,0});
+		info.dlpi_tls_data = !current->tls_id ? 0 :
+			__tls_get_addr((tls_mod_off_t[]){current->tls_id,0});
 
 		ret = (callback)(&info, sizeof (info), data);