This is the mail archive of the
binutils@sourceware.org
mailing list for the binutils project.
Re: [PATCH] gas: allow labeling of CFI instructions
- From: "H.J. Lu" <hjl dot tools at gmail dot com>
- To: Jan Beulich <JBeulich at suse dot com>
- Cc: Binutils <binutils at sourceware dot org>
- Date: Fri, 19 Dec 2014 08:24:55 -0800
- Subject: Re: [PATCH] gas: allow labeling of CFI instructions
- Authentication-results: sourceware.org; auth=none
- References: <549443A90200007800051089 at mail dot emea dot novell dot com> <CAMe9rOqmE9NnzzGVLxOO=pJ2KJxKUpR5Ze07QVWGWxqcEjGvsA at mail dot gmail dot com> <549453E70200007800051111 at mail dot emea dot novell dot com> <CAMe9rOoWV-KL32gaQEwNH7j0wAhegu=7pN1P8TGaxZuyda77fw at mail dot gmail dot com> <54945E8E0200007800051160 at mail dot emea dot novell dot com>
On Fri, Dec 19, 2014 at 8:21 AM, Jan Beulich <JBeulich@suse.com> wrote:
>>>> On 19.12.14 at 16:47, <hjl.tools@gmail.com> wrote:
>> On Fri, Dec 19, 2014 at 7:35 AM, Jan Beulich <JBeulich@suse.com> wrote:
>>>>>> On 19.12.14 at 15:32, <hjl.tools@gmail.com> wrote:
>>>> On Fri, Dec 19, 2014 at 6:26 AM, Jan Beulich <JBeulich@suse.com> wrote:
>>>>> When runtime patching code (like e.g. done by the Linux kernel) there
>>>>> may be cases where the set of stack frame alterations differs between
>>>>> unpatched and patched code. Consequently the corresponding unwind data
>>>>> needs patching too. Locating the right places within an FDE, however,
>>>>> is rather cumbersome without a way to insert labels in the resulting
>>>>> section. Hence this patch introduces a new directive, .cfi_label. Note
>>>>> that with the way CFI data gets emitted currently (at the end of the
>>>>> assembly process) this can't support local FB- and dollar-labels.
>>>>>
>>>>> gas/
>>>>> 2014-12-19 Jan Beulich <jbeulich@suse.com>
>>>>>
>>>>> * gas/dw2gencfi.c (cfi_add_label, dot_cfi_label): New.
>>>>> (cfi_pseudo_table): Add "cfi_label".
>>>>> (output_cfi_insn): Handle CFI_label.
>>>>> (select_cie_for_fde): Als terminate CIE when encountering
>>>>> CFI_label.
>>>>> * dw2gencfi.h (cfi_add_label): Declare.
>>>>> (struct cfi_insn_data): New member "sym_name".
>>>>> (CFI_label): New.
>>>>> * read.c (read_symbol_name): Drop "static".
>>>>> * read.h (read_symbol_name): Declare.
>>>>>
>>>>
>>>> No testcases?
>>>
>>> Oh, I meant to say a word on the lack thereof: I can't see how to
>>> create a meaningful, yet architecture independent test case for
>>> this, despite having thought about possibly ways quite a bit. Any
>>> suggestions are welcome.
>>
>> Can you extract some testcases from x86/x86-64 kernel?
>
> The code to make use of this new directive has yet to be written
> (and is unlikely to go upstream due to Linus vetoing any such
> changes originating from me). And what help would x86-specific
> tests really be? Yes, they may be better than nothing, but
> they're in no way helping to avoid breakage (they'd only allow to
> maybe notice it earlier after some commit already happened).
>
At minimum, it will verify your changes do what they are supposed
to do and your changes won't be broken by accident.
--
H.J.