Re: [PATCH 1/5] glibc: Perform rseq(2) registration at C startup and thread creation (v10)

[Mathieu Desnoyers <mathieu dot desnoyers at efficios dot com>]

----- On Jun 14, 2019, at 3:24 PM, Florian Weimer fweimer@redhat.com wrote:

> * Mathieu Desnoyers:
> 
>> ----- On Jun 14, 2019, at 3:09 PM, Florian Weimer fweimer@redhat.com wrote:
>>
>>> * Mathieu Desnoyers:
>>> 
>>>> But my original issue remains: if I define a variable called __rseq_handled
>>>> within either the main executable or the preloaded library, it overshadows
>>>> the libc one:
>>>>
>>>> efficios@compudjdev:~/test/libc-sym$ ./a
>>>> __rseq_handled main: 0 0x56135fd5102c
>>>> __rseq_abi.cpu_id main: 29 0x7fcbeca6d5a0
>>>> efficios@compudjdev:~/test/libc-sym$ LD_PRELOAD=./s.so ./a
>>>> __rseq_handled s.so: 0 0x558f70aeb02c
>>>> __rseq_abi.cpu_id s.so: -1 0x7fdca78b7760
>>>> __rseq_handled main: 0 0x558f70aeb02c
>>>> __rseq_abi.cpu_id main: 27 0x7fdca78b7760
>>>>
>>>> Which is unexpected.
>>> 
>>> Why is this unexpected?  It has to be this way if the main program uses
>>> a copy relocation of __rseq_handled.  As long as there is just one
>>> address across the entire program and ld.so initializes the copy of the
>>> variable that is actually used, everything will be fine.
>>
>> Here is a printout of the __rseq_handled address observed by ld.so, it
>> does not match:
>>
>> LD_PRELOAD=./s.so ./a
>> elf: __rseq_handled addr: 7f501c98a140
>> __rseq_handled s.so: 0 0x55817a88d02c
>> __rseq_abi.cpu_id s.so: -1 0x7f501c983760
>> __rseq_handled main: 0 0x55817a88d02c
>> __rseq_abi.cpu_id main: 27 0x7f501c983760
> 
> Where do you print the address?  Before or after the self-relocation of
> the dynamic loader?  The address is only correct after self-relocation.

I printed the address within rseq_init (), which happened to be invoked
by the linker startup waaaay too early. I followed your advice and moved
the rseq_init () invocation after linker re-relocation:

diff --git a/elf/rtld.c b/elf/rtld.c
index f29f284a7c..66b0894f9d 100644
--- a/elf/rtld.c
+++ b/elf/rtld.c
@@ -1410,9 +1410,6 @@ ERROR: '%s': cannot process note segment.\n", _dl_argv[0]);
     /* Assign a module ID.  Do this before loading any audit modules.  */
     GL(dl_rtld_map).l_tls_modid = _dl_next_tls_modid ();
 
-  /* Publicize rseq registration ownership.  */
-  rseq_init ();
-
   /* If we have auditing DSOs to load, do it now.  */
   bool need_security_init = true;
   if (__glibc_unlikely (audit_list != NULL)
@@ -2284,6 +2281,11 @@ ERROR: ld.so: object '%s' cannot be loaded as audit interface: %s; ignored.\n",
       HP_TIMING_ACCUM_NT (relocate_time, add);
     }
 
+  /* Publicize rseq registration ownership.  This must be performed
+     after rtld re-relocation, before invoking constructors of
+     preloaded libraries.  */
+  rseq_init ();
+
   /* Do any necessary cleanups for the startup OS interface code.
      We do these now so that no calls are made after rtld re-relocation
      which might be resolved to different functions than we expect.

It works fine now!

LD_PRELOAD=./s.so ./a
elf: __rseq_handled addr: 56300f0a402c
__rseq_handled s.so: 1 0x56300f0a402c
__rseq_abi.cpu_id s.so: -1 0x7fad2ff58760
__rseq_handled main: 1 0x56300f0a402c
__rseq_abi.cpu_id main: 27 0x7fad2ff58760

Thanks!

Mathieu

-- 
Mathieu Desnoyers
EfficiOS Inc.
http://www.efficios.com