From: Yichao Yu <yyc1992@gmail.com>
To: gdb@sourceware.org
Cc: Yichao Yu <yyc1992@gmail.com>
Subject: Re: [PATCH] Lookup the JIT descriptor symbol first to avoid finding the PLT entry of the breakpoint in the wrong object file.
Date: Fri, 24 Feb 2017 19:01:00 -0000 [thread overview]
Message-ID: <CAMvDr+Q3NWS9YK4Aefop_-eSPyM9A4btK5N+WkqnFsHJrJdZgQ@mail.gmail.com> (raw)
In-Reply-To: <CAMvDr+RnUgHwSAeYpfhAyhh4G_cF1fC=WG-+Nc0k1wJ85B8mWg@mail.gmail.com>
Ping again.
On Sun, Feb 5, 2017 at 7:46 PM, Yichao Yu <yyc1992@gmail.com> wrote:
> Ping. Can anyone please review this? Or is there a better place to post this?
>
> On Thu, Jan 26, 2017 at 3:31 PM, Yichao Yu <yyc1992@gmail.com> wrote:
>> On Thu, Jan 26, 2017 at 3:24 PM, Yichao Yu <yyc1992@gmail.com> wrote:
>>> Fix 20633
>>
>>
>> ..... I don't use git send-mail very often and missed the chance to
>> type more comments....
>>
>> The link to the bug report is
>> https://sourceware.org/bugzilla/show_bug.cgi?id=20633 . I think
>> ideally the interface should find all the global and local symbols and
>> breakpoint on all of them since there isn't an agreement on who will
>> provide it AFAICT (and for example,a process can load multiple copies
>> of statically linked LLVM each providing a jit debugging interface).
>> However, I'm not really sure what's the right function to use for that
>> and it'll probably be a much bigger change so I went for this smaller
>> change that fixes the original issue.
>>
>> Not sure how this should be tested....... =(
>>
>>> ---
>>> gdb/jit.c | 17 ++++++++++-------
>>> 1 file changed, 10 insertions(+), 7 deletions(-)
>>>
>>> diff --git a/gdb/jit.c b/gdb/jit.c
>>> index 158d6d8..d6eb800 100644
>>> --- a/gdb/jit.c
>>> +++ b/gdb/jit.c
>>> @@ -1051,18 +1051,21 @@ jit_breakpoint_re_set_internal (struct gdbarch *gdbarch,
>>> if (ps_data->objfile == NULL)
>>> {
>>> /* Lookup the registration symbol. If it is missing, then we
>>> - assume we are not attached to a JIT. */
>>> - reg_symbol = lookup_minimal_symbol_and_objfile (jit_break_name);
>>> - if (reg_symbol.minsym == NULL
>>> - || BMSYMBOL_VALUE_ADDRESS (reg_symbol) == 0)
>>> - return 1;
>>> + assume we are not attached to a JIT.
>>> + Lookup the descriptor first since looking up the breakpoint might
>>> + return a PLT entry in the wrong file. */
>>>
>>> - desc_symbol = lookup_minimal_symbol (jit_descriptor_name, NULL,
>>> - reg_symbol.objfile);
>>> + desc_symbol = lookup_minimal_symbol_and_objfile (jit_descriptor_name);
>>> if (desc_symbol.minsym == NULL
>>> || BMSYMBOL_VALUE_ADDRESS (desc_symbol) == 0)
>>> return 1;
>>>
>>> + reg_symbol = lookup_minimal_symbol (jit_break_name, NULL,
>>> + desc_symbol.objfile);
>>> + if (reg_symbol.minsym == NULL
>>> + || BMSYMBOL_VALUE_ADDRESS (reg_symbol) == 0)
>>> + return 1;
>>> +
>>> objf_data = get_jit_objfile_data (reg_symbol.objfile);
>>> objf_data->register_code = reg_symbol.minsym;
>>> objf_data->descriptor = desc_symbol.minsym;
>>> --
>>> 2.10.2
>>>
next prev parent reply other threads:[~2017-02-24 19:01 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-01-26 20:24 Yichao Yu
2017-01-26 20:31 ` Yichao Yu
2017-02-06 0:46 ` Yichao Yu
2017-02-24 19:01 ` Yichao Yu [this message]
2017-03-09 12:25 ` Yichao Yu
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=CAMvDr+Q3NWS9YK4Aefop_-eSPyM9A4btK5N+WkqnFsHJrJdZgQ@mail.gmail.com \
--to=yyc1992@gmail.com \
--cc=gdb@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