public inbox for gcc-bugs@sourceware.org help / color / mirror / Atom feed
From: "hp at gcc dot gnu dot org" <gcc-bugzilla@gcc.gnu.org> To: gcc-bugs@gcc.gnu.org Subject: [Bug target/14346] New: With -fpic/-fPIC, thunks jump through PLT, not directly to thunked function Date: Sun, 29 Feb 2004 16:12:00 -0000 [thread overview] Message-ID: <20040229161232.14346.hp@gcc.gnu.org> (raw) As in summary. Because thunks do not set up the PIC register, going through the PLT causes a SEGV in best case. Another PIC construct than for normal calls must be used, but unfortunately the only (currently) available PC-relative PIC reloc for cris-axis-linux-gnu requests a PLT. The thunked function is global. Using a local symbol makes the linker omit the redirection through the PLT and the jump goes directly to the right place. On trunk and 3.4 branch, the thunk uses a local symbol (equated to the thunked function, generated by generic code), so TRT happens. See <URL:http://gcc.gnu.org/ml/gcc-patches/2003-06/msg02603.html> for the first of the patch where the generic local-thunk was implemented (needed follow-up patches for correctness). I'll go with a target-local patch. Testcase: -------------- th1.h class B { int bb1; int bb2; public: virtual void b(); }; extern class B* mka(); -------------- th1.cc (compile and link with -fpic -shared -o libth1.so) #include "th1.h" class A { int aa; public: virtual void a() = 0; }; class C : public A, public B { int cc; public: virtual void b(); virtual void a(); }; extern "C" void abort (void); extern "C" void exit (int); class B* mka (void) { return new C; } void C::b(void) { exit (0); } void B::b(void) { abort (); } void C::a(void) { abort (); } -------------- m.cc (compile and link with -L. -lth1, expect normal exit) #include "th1.h" extern "C" void abort (void); int ii[10]; int main(void) { register int picreg #if defined (__CRIS__) && !defined (__PIC__) && !defined (__pic__) __asm__ ("r0") #endif = -1; class B *bb = mka(); bb->b(); // Make sure picreg is alive and used after bb->b() above. asm ("" : : "r" (picreg)); abort (); } -- Summary: With -fpic/-fPIC, thunks jump through PLT, not directly to thunked function Product: gcc Version: unknown Status: UNCONFIRMED Keywords: wrong-code Severity: normal Priority: P1 Component: target AssignedTo: hp at gcc dot gnu dot org ReportedBy: hp at gcc dot gnu dot org CC: gcc-bugs at gcc dot gnu dot org GCC target triplet: cris-axis-linux-gnu http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14346
next reply other threads:[~2004-02-29 16:12 UTC|newest] Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top 2004-02-29 16:12 hp at gcc dot gnu dot org [this message] 2004-02-29 16:14 ` [Bug target/14346] " hp at gcc dot gnu dot org 2004-02-29 16:31 ` [Bug target/14346] [3.3 only] " hp at gcc dot gnu dot org 2004-02-29 21:46 ` cvs-commit at gcc dot gnu dot org 2004-02-29 21:52 ` hp at gcc dot gnu dot org
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=20040229161232.14346.hp@gcc.gnu.org \ --to=gcc-bugzilla@gcc.gnu.org \ --cc=gcc-bugs@gcc.gnu.org \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).