public inbox for glibc-bugs@sourceware.org
help / color / mirror / Atom feed
From: "palmer at gcc dot gnu.org" <sourceware-bugzilla@sourceware.org>
To: glibc-bugs@sourceware.org
Subject: [Bug dynamic-link/31317] [RISCV] static PIE crashes during self relocation
Date: Thu, 22 Feb 2024 22:53:11 +0000	[thread overview]
Message-ID: <bug-31317-131-XQp39h97ai@http.sourceware.org/bugzilla/> (raw)
In-Reply-To: <bug-31317-131@http.sourceware.org/bugzilla/>

https://sourceware.org/bugzilla/show_bug.cgi?id=31317

Palmer Dabbelt <palmer at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |adhemerval.zanella at linaro dot o
                   |                            |rg

--- Comment #3 from Palmer Dabbelt <palmer at gcc dot gnu.org> ---
(In reply to Andreas Schwab from comment #2)
> l_scope is never null, this condition needs to be removed again.  It's
> simply the wrong check.

Sorry, I think I'm a bit lost here.  IIUC we don't actually need GP set at this
point for static-PIE, as it's done in _start.  So I maybe something like this

diff --git a/sysdeps/riscv/dl-machine.h b/sysdeps/riscv/dl-machine.h
index 0cbb476c05..ae38df4bc1 100644
--- a/sysdeps/riscv/dl-machine.h
+++ b/sysdeps/riscv/dl-machine.h
@@ -348,7 +348,7 @@ elf_machine_runtime_setup (struct link_map *l, struct
r_scope_elem *scope[],
       gotplt[1] = (ElfW(Addr)) l;
     }

-  if (l->l_type == lt_executable && l->l_scope != NULL)
+  if (l->l_type == lt_executable && scope != NULL)
     {
       /* The __global_pointer$ may not be defined by the linker if the
         $gp register does not be used to access the global variable
@@ -360,7 +360,7 @@ elf_machine_runtime_setup (struct link_map *l, struct
r_scope_elem *scope[],

       const ElfW(Sym) *ref = &gp_sym;
       _dl_lookup_symbol_x ("__global_pointer$", l, &ref,
-                          l->l_scope, NULL, 0, 0, NULL);
+                          scope, NULL, 0, 0, NULL);
       if (ref)
         asm (
           "mv gp, %0\n"

should do it?  This relies on the caller to obtain scope, the static-PIE call
sets it to NULL and the others set it to l_scope.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

  parent reply	other threads:[~2024-02-22 22:53 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-01-30 13:25 [Bug dynamic-link/31317] New: " schwab@linux-m68k.org
2024-02-22 18:44 ` [Bug dynamic-link/31317] " palmer at gcc dot gnu.org
2024-02-22 21:47 ` schwab@linux-m68k.org
2024-02-22 22:53 ` palmer at gcc dot gnu.org [this message]
2024-02-22 23:01 ` vineet.gupta at linux dot dev
2024-02-22 23:16 ` schwab@linux-m68k.org
2024-02-22 23:26 ` palmer at gcc dot gnu.org
2024-02-23  3:06 ` yanzhang.wang at intel dot com
2024-03-06 14:44 ` rjones at redhat dot com
2024-03-25 14:20 ` cvs-commit at gcc dot gnu.org
2024-03-26 13:47 ` schwab@linux-m68k.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=bug-31317-131-XQp39h97ai@http.sourceware.org/bugzilla/ \
    --to=sourceware-bugzilla@sourceware.org \
    --cc=glibc-bugs@sourceware.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: link
Be 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).