* [PATCH] elf: Don't call fatal_error in _dl_signal_error
@ 2024-04-03 12:39 H.J. Lu
2024-04-03 19:54 ` Adhemerval Zanella Netto
2024-04-08 15:28 ` Florian Weimer
0 siblings, 2 replies; 5+ messages in thread
From: H.J. Lu @ 2024-04-03 12:39 UTC (permalink / raw)
To: libc-alpha
Don't call fatal_error in _dl_signal_error since _dl_signal_error should
only be called from _dl_catch_exception and get_catch should never return
NULL. Otherwise, fatal_error should be called directly instead.
---
elf/dl-catch.c | 13 ++++---------
1 file changed, 4 insertions(+), 9 deletions(-)
diff --git a/elf/dl-catch.c b/elf/dl-catch.c
index 2109516dba..a023555165 100644
--- a/elf/dl-catch.c
+++ b/elf/dl-catch.c
@@ -117,16 +117,11 @@ _dl_signal_error (int errcode, const char *objname, const char *occasion,
if (! errstring)
errstring = N_("DYNAMIC LINKER BUG!!!");
- if (lcatch != NULL)
- {
- _dl_exception_create (lcatch->exception, objname, errstring);
- *lcatch->errcode = errcode;
+ _dl_exception_create (lcatch->exception, objname, errstring);
+ *lcatch->errcode = errcode;
- /* We do not restore the signal mask because none was saved. */
- __longjmp (lcatch->env[0].__jmpbuf, 1);
- }
- else
- fatal_error (errcode, objname, occasion, errstring);
+ /* We do not restore the signal mask because none was saved. */
+ __longjmp (lcatch->env[0].__jmpbuf, 1);
}
rtld_hidden_def (_dl_signal_error)
--
2.44.0
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] elf: Don't call fatal_error in _dl_signal_error
2024-04-03 12:39 [PATCH] elf: Don't call fatal_error in _dl_signal_error H.J. Lu
@ 2024-04-03 19:54 ` Adhemerval Zanella Netto
2024-04-03 20:00 ` H.J. Lu
2024-04-08 15:28 ` Florian Weimer
1 sibling, 1 reply; 5+ messages in thread
From: Adhemerval Zanella Netto @ 2024-04-03 19:54 UTC (permalink / raw)
To: libc-alpha, H.J. Lu
On 03/04/24 09:39, H.J. Lu wrote:
> Don't call fatal_error in _dl_signal_error since _dl_signal_error should
> only be called from _dl_catch_exception and get_catch should never return
> NULL. Otherwise, fatal_error should be called directly instead.
Is this related to BZ#31596 or did you catch it by code review? I think
this is required for lazy binding being fatal error on initializer
functions (elf/dl-open.c:829).
> ---
> elf/dl-catch.c | 13 ++++---------
> 1 file changed, 4 insertions(+), 9 deletions(-)
>
> diff --git a/elf/dl-catch.c b/elf/dl-catch.c
> index 2109516dba..a023555165 100644
> --- a/elf/dl-catch.c
> +++ b/elf/dl-catch.c
> @@ -117,16 +117,11 @@ _dl_signal_error (int errcode, const char *objname, const char *occasion,
> if (! errstring)
> errstring = N_("DYNAMIC LINKER BUG!!!");
>
> - if (lcatch != NULL)
> - {
> - _dl_exception_create (lcatch->exception, objname, errstring);
> - *lcatch->errcode = errcode;
> + _dl_exception_create (lcatch->exception, objname, errstring);
> + *lcatch->errcode = errcode;
>
> - /* We do not restore the signal mask because none was saved. */
> - __longjmp (lcatch->env[0].__jmpbuf, 1);
> - }
> - else
> - fatal_error (errcode, objname, occasion, errstring);
> + /* We do not restore the signal mask because none was saved. */
> + __longjmp (lcatch->env[0].__jmpbuf, 1);
> }
> rtld_hidden_def (_dl_signal_error)
>
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] elf: Don't call fatal_error in _dl_signal_error
2024-04-03 19:54 ` Adhemerval Zanella Netto
@ 2024-04-03 20:00 ` H.J. Lu
2024-04-04 16:56 ` Adhemerval Zanella Netto
0 siblings, 1 reply; 5+ messages in thread
From: H.J. Lu @ 2024-04-03 20:00 UTC (permalink / raw)
To: Adhemerval Zanella Netto; +Cc: libc-alpha
On Wed, Apr 3, 2024 at 12:54 PM Adhemerval Zanella Netto
<adhemerval.zanella@linaro.org> wrote:
>
>
>
> On 03/04/24 09:39, H.J. Lu wrote:
> > Don't call fatal_error in _dl_signal_error since _dl_signal_error should
> > only be called from _dl_catch_exception and get_catch should never return
> > NULL. Otherwise, fatal_error should be called directly instead.
>
> Is this related to BZ#31596 or did you catch it by code review? I think
Yes, it is related. I noticed by code review.
> this is required for lazy binding being fatal error on initializer
> functions (elf/dl-open.c:829).
elf/dl-open.c:829 calls _dl_catch_exception.
> > ---
> > elf/dl-catch.c | 13 ++++---------
> > 1 file changed, 4 insertions(+), 9 deletions(-)
> >
> > diff --git a/elf/dl-catch.c b/elf/dl-catch.c
> > index 2109516dba..a023555165 100644
> > --- a/elf/dl-catch.c
> > +++ b/elf/dl-catch.c
> > @@ -117,16 +117,11 @@ _dl_signal_error (int errcode, const char *objname, const char *occasion,
> > if (! errstring)
> > errstring = N_("DYNAMIC LINKER BUG!!!");
> >
> > - if (lcatch != NULL)
> > - {
> > - _dl_exception_create (lcatch->exception, objname, errstring);
> > - *lcatch->errcode = errcode;
> > + _dl_exception_create (lcatch->exception, objname, errstring);
> > + *lcatch->errcode = errcode;
> >
> > - /* We do not restore the signal mask because none was saved. */
> > - __longjmp (lcatch->env[0].__jmpbuf, 1);
> > - }
> > - else
> > - fatal_error (errcode, objname, occasion, errstring);
> > + /* We do not restore the signal mask because none was saved. */
> > + __longjmp (lcatch->env[0].__jmpbuf, 1);
> > }
> > rtld_hidden_def (_dl_signal_error)
> >
--
H.J.
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] elf: Don't call fatal_error in _dl_signal_error
2024-04-03 20:00 ` H.J. Lu
@ 2024-04-04 16:56 ` Adhemerval Zanella Netto
0 siblings, 0 replies; 5+ messages in thread
From: Adhemerval Zanella Netto @ 2024-04-04 16:56 UTC (permalink / raw)
To: H.J. Lu; +Cc: libc-alpha
On 03/04/24 17:00, H.J. Lu wrote:
> On Wed, Apr 3, 2024 at 12:54 PM Adhemerval Zanella Netto
> <adhemerval.zanella@linaro.org> wrote:
>>
>>
>>
>> On 03/04/24 09:39, H.J. Lu wrote:
>>> Don't call fatal_error in _dl_signal_error since _dl_signal_error should
>>> only be called from _dl_catch_exception and get_catch should never return
>>> NULL. Otherwise, fatal_error should be called directly instead.
>>
>> Is this related to BZ#31596 or did you catch it by code review? I think
>
> Yes, it is related. I noticed by code review.
>
>> this is required for lazy binding being fatal error on initializer
>> functions (elf/dl-open.c:829).
>
> elf/dl-open.c:829 calls _dl_catch_exception.
I think it because lazy binding errors from _dl_lookup_symbol_x calls
_dl_exception_create_format/_dl_signal_cexception instead of _dl_signal_error
mainly because it can not translate the message at the time.
From _dl_catch_exception comment, it aims to support first argument being
NULL so all exceptions are fatal. This change change this assumption,
meaning _dl_signal_error can not be called in this way (since now it always
expects to return the exception to caller).
>
>>> ---
>>> elf/dl-catch.c | 13 ++++---------
>>> 1 file changed, 4 insertions(+), 9 deletions(-)
>>>
>>> diff --git a/elf/dl-catch.c b/elf/dl-catch.c
>>> index 2109516dba..a023555165 100644
>>> --- a/elf/dl-catch.c
>>> +++ b/elf/dl-catch.c
>>> @@ -117,16 +117,11 @@ _dl_signal_error (int errcode, const char *objname, const char *occasion,
>>> if (! errstring)
>>> errstring = N_("DYNAMIC LINKER BUG!!!");
>>>
>>> - if (lcatch != NULL)
>>> - {
>>> - _dl_exception_create (lcatch->exception, objname, errstring);
>>> - *lcatch->errcode = errcode;
>>> + _dl_exception_create (lcatch->exception, objname, errstring);
>>> + *lcatch->errcode = errcode;
>>>
>>> - /* We do not restore the signal mask because none was saved. */
>>> - __longjmp (lcatch->env[0].__jmpbuf, 1);
>>> - }
>>> - else
>>> - fatal_error (errcode, objname, occasion, errstring);
>>> + /* We do not restore the signal mask because none was saved. */
>>> + __longjmp (lcatch->env[0].__jmpbuf, 1);
>>> }
>>> rtld_hidden_def (_dl_signal_error)
>>>
>
>
>
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] elf: Don't call fatal_error in _dl_signal_error
2024-04-03 12:39 [PATCH] elf: Don't call fatal_error in _dl_signal_error H.J. Lu
2024-04-03 19:54 ` Adhemerval Zanella Netto
@ 2024-04-08 15:28 ` Florian Weimer
1 sibling, 0 replies; 5+ messages in thread
From: Florian Weimer @ 2024-04-08 15:28 UTC (permalink / raw)
To: H.J. Lu; +Cc: libc-alpha
* H. J. Lu:
> Don't call fatal_error in _dl_signal_error since _dl_signal_error should
> only be called from _dl_catch_exception and get_catch should never return
> NULL. Otherwise, fatal_error should be called directly instead.
> ---
> elf/dl-catch.c | 13 ++++---------
> 1 file changed, 4 insertions(+), 9 deletions(-)
>
> diff --git a/elf/dl-catch.c b/elf/dl-catch.c
> index 2109516dba..a023555165 100644
> --- a/elf/dl-catch.c
> +++ b/elf/dl-catch.c
> @@ -117,16 +117,11 @@ _dl_signal_error (int errcode, const char *objname, const char *occasion,
> if (! errstring)
> errstring = N_("DYNAMIC LINKER BUG!!!");
>
> - if (lcatch != NULL)
> - {
> - _dl_exception_create (lcatch->exception, objname, errstring);
> - *lcatch->errcode = errcode;
> + _dl_exception_create (lcatch->exception, objname, errstring);
> + *lcatch->errcode = errcode;
>
> - /* We do not restore the signal mask because none was saved. */
> - __longjmp (lcatch->env[0].__jmpbuf, 1);
> - }
> - else
> - fatal_error (errcode, objname, occasion, errstring);
> + /* We do not restore the signal mask because none was saved. */
> + __longjmp (lcatch->env[0].__jmpbuf, 1);
> }
> rtld_hidden_def (_dl_signal_error)
elf/ld.so program-does-not-exist triggers are null pointer dereference
with this patch.
I'll post a proper test case for this, looks like we have a gap here.
Thanks,
Florian
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2024-04-08 15:28 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2024-04-03 12:39 [PATCH] elf: Don't call fatal_error in _dl_signal_error H.J. Lu
2024-04-03 19:54 ` Adhemerval Zanella Netto
2024-04-03 20:00 ` H.J. Lu
2024-04-04 16:56 ` Adhemerval Zanella Netto
2024-04-08 15:28 ` Florian Weimer
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).