Mirror of the gdb-patches mailing list
 help / color / mirror / Atom feed
From: Simon Marchi via Gdb-patches <gdb-patches@sourceware.org>
To: Zoran Zaric <Zoran.Zaric@amd.com>, gdb-patches@sourceware.org
Subject: Re: [PATCH 03/43] Move frame context info to dwarf_expr_context
Date: Wed, 28 Apr 2021 10:14:43 -0400	[thread overview]
Message-ID: <5a2c4d75-2059-cc93-6581-59cde0d6ef40@polymtl.ca> (raw)
In-Reply-To: <0edee4c1-9a04-9bf5-7d44-0973c675462b@amd.com>

>> On an unrelated topic, things like ensure_have_frame always worry me
>> because I'm scared I'll forget to call it where necessary.  An
>> alternative could be to make sure we get the frame through a getter that
>> throws if frame == nullptr.
>>
>> Basically, a tiny class (could be internal to dwarf_expr_context):
>>
>> struct frame_safe
>> {
>>    frame_safe (frame_info *frame)
>>      : m_frame (frame)
>>    {}
>>
>>    frame_info *frame (const char *op)
>>    {
>>      if (m_frame == nullptr)
>>        ... throw ...
>>
>>      return m_frame;
>>    }
>>
>> private:
>>    frame_info *m_frame;
>> };
>>
>> dwarf_expr_context would have a `frame_safe m_frame` field and a
>> `frame (const char *op)` getter that just does
>> `return m_frame->frame (op)`.
>>
>> This way, it's impossible to get the frame and forgetting an
>> ensure_have_frame call.
> 
> True, but then you end up with even more "switch" statements in different parts of the gdb that one needs to update every time they add a new operation and this is one of the problems with the current design as is.

I don't understand what you mean, why it would mean adding more switch
statements.  In practice, it would just mean changing `this->frame` to
`this->frame ()`.

> This would force us to add two new places (one for frame and one for compilation unit) and there will be more in the future when we add a concept of a thread to the context.
> 
> At least with the current implementation the implementer of the case statement that process that operation needs to know which part of the context has to be present for that operation, which is clearly defined in the DWARF standard as part of that operations definition (or at least it will be after our extensions).
> 
> Another problem with the new class approach is that it makes the information about what each operation needs convoluted and hidden which in my mind adds even bigger chance of someone forgetting to add their operation to the query calls.

What do you mean by "add their operation to the query calls"?

I don't think see how it helps to make it explicit which operation needs
which part of the context.  If an operation needs a frame, it will call
`this->frame ()`  which will throw if the context does not have a
frame.

Simon

  reply	other threads:[~2021-04-28 14:14 UTC|newest]

Thread overview: 86+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-03-01 14:45 [PATCH 00/43 V2] Allow location description on the DWARF stack Zoran Zaric via Gdb-patches
2021-03-01 14:45 ` [PATCH 01/43] Replace the symbol needs evaluator with a parser Zoran Zaric via Gdb-patches
2021-04-27  1:20   ` Simon Marchi via Gdb-patches
2021-04-28 10:17     ` Zoran Zaric via Gdb-patches
2021-04-28 14:08       ` Simon Marchi via Gdb-patches
2021-04-28 15:02         ` Zoran Zaric via Gdb-patches
2021-04-28 15:31         ` Zoran Zaric via Gdb-patches
2021-03-01 14:45 ` [PATCH 02/43] Cleanup of the dwarf_expr_context constructor Zoran Zaric via Gdb-patches
2021-04-27  1:23   ` Simon Marchi via Gdb-patches
2021-04-28 10:19     ` Zoran Zaric via Gdb-patches
2021-03-01 14:45 ` [PATCH 03/43] Move frame context info to dwarf_expr_context Zoran Zaric via Gdb-patches
2021-04-27  2:19   ` Simon Marchi via Gdb-patches
2021-04-28 10:51     ` Zoran Zaric via Gdb-patches
2021-04-28 14:14       ` Simon Marchi via Gdb-patches [this message]
2021-04-28 15:55         ` Zoran Zaric via Gdb-patches
2021-03-01 14:45 ` [PATCH 04/43] Remove get_frame_cfa from dwarf_expr_context Zoran Zaric via Gdb-patches
2021-03-01 14:45 ` [PATCH 05/43] Move compilation unit info to dwarf_expr_context Zoran Zaric via Gdb-patches
2021-04-27  2:58   ` Simon Marchi via Gdb-patches
2021-04-28 11:28     ` Zoran Zaric via Gdb-patches
2021-03-01 14:45 ` [PATCH 06/43] Move dwarf_call " Zoran Zaric via Gdb-patches
2021-03-01 14:45 ` [PATCH 07/43] Move get_object_address " Zoran Zaric via Gdb-patches
2021-04-27  3:12   ` Simon Marchi via Gdb-patches
2021-04-28 11:34     ` Zoran Zaric via Gdb-patches
2021-03-01 14:45 ` [PATCH 08/43] Move read_mem " Zoran Zaric via Gdb-patches
2021-03-01 14:45 ` [PATCH 09/43] Move push_dwarf_reg_entry_value to expr.c Zoran Zaric via Gdb-patches
2021-04-27  3:56   ` Simon Marchi via Gdb-patches
2021-04-28 11:36     ` Zoran Zaric via Gdb-patches
2021-03-01 14:45 ` [PATCH 10/43] Inline get_reg_value method of dwarf_expr_context Zoran Zaric via Gdb-patches
2021-03-01 14:45 ` [PATCH 11/43] Remove empty frame and full evaluators Zoran Zaric via Gdb-patches
2021-03-01 14:45 ` [PATCH 12/43] Merge evaluate_for_locexpr_baton evaluator Zoran Zaric via Gdb-patches
2021-04-28  1:33   ` Simon Marchi via Gdb-patches
2021-04-28 11:39     ` Zoran Zaric via Gdb-patches
2021-03-01 14:45 ` [PATCH 13/43] Move piece_closure and its support to expr.c Zoran Zaric via Gdb-patches
2021-04-28  1:56   ` Simon Marchi via Gdb-patches
2021-04-28 11:40     ` Zoran Zaric via Gdb-patches
2021-03-01 14:45 ` [PATCH 14/43] Make value_copy also copy the stack data member Zoran Zaric via Gdb-patches
2021-04-28  2:01   ` Simon Marchi via Gdb-patches
2021-04-28 11:43     ` Zoran Zaric via Gdb-patches
2021-03-01 14:45 ` [PATCH 15/43] Make DWARF evaluator return a single struct value Zoran Zaric via Gdb-patches
2021-04-28  2:21   ` Simon Marchi via Gdb-patches
2021-04-28 11:47     ` Zoran Zaric via Gdb-patches
2021-04-28 14:24       ` Simon Marchi via Gdb-patches
2021-03-01 14:45 ` [PATCH 16/43] Simplify dwarf_expr_context class interface Zoran Zaric via Gdb-patches
2021-04-28  2:45   ` Simon Marchi via Gdb-patches
2021-04-28 13:15     ` Zoran Zaric via Gdb-patches
2021-04-28 14:41       ` Simon Marchi via Gdb-patches
2021-04-28 15:39         ` Zoran Zaric via Gdb-patches
2021-04-28 19:19           ` Simon Marchi via Gdb-patches
2021-04-29 15:49       ` Simon Marchi via Gdb-patches
2021-04-29 15:55         ` Zoran Zaric via Gdb-patches
2021-03-01 14:45 ` [PATCH 17/43] Add as_lval argument to expression evaluator Zoran Zaric via Gdb-patches
2021-04-28  3:04   ` Simon Marchi via Gdb-patches
2021-04-28 13:16     ` Zoran Zaric via Gdb-patches
2021-04-28  3:30   ` Simon Marchi via Gdb-patches
2021-03-01 14:45 ` [PATCH 18/43] Add new register access interface to expr.c Zoran Zaric via Gdb-patches
2021-03-08 23:52   ` Lancelot SIX via Gdb-patches
2021-04-28  3:25   ` Simon Marchi via Gdb-patches
2021-04-28 13:29     ` Zoran Zaric via Gdb-patches
2021-04-28 14:48       ` Simon Marchi via Gdb-patches
2021-04-28 15:42         ` Zoran Zaric via Gdb-patches
2021-03-01 14:45 ` [PATCH 19/43] Add new memory " Zoran Zaric via Gdb-patches
2021-04-30 21:24   ` Simon Marchi via Gdb-patches
2021-03-01 14:45 ` [PATCH 20/43] Add new classes that model DWARF stack element Zoran Zaric via Gdb-patches
2021-03-01 14:45 ` [PATCH 21/43] Add to_location method to DWARF entry classes Zoran Zaric via Gdb-patches
2021-03-01 14:45 ` [PATCH 22/43] Add to_value " Zoran Zaric via Gdb-patches
2021-03-01 14:46 ` [PATCH 23/43] Add read method to location description classes Zoran Zaric via Gdb-patches
2021-03-01 14:46 ` [PATCH 24/43] Add write " Zoran Zaric via Gdb-patches
2021-03-01 14:46 ` [PATCH 25/43] Add deref " Zoran Zaric via Gdb-patches
2021-03-01 14:46 ` [PATCH 26/43] Add read_from_gdb_value method to dwarf_location Zoran Zaric via Gdb-patches
2021-03-01 14:46 ` [PATCH 27/43] Add write_to_gdb_value " Zoran Zaric via Gdb-patches
2021-03-01 14:46 ` [PATCH 28/43] Add is_implicit_ptr_at " Zoran Zaric via Gdb-patches
2021-03-01 14:46 ` [PATCH 29/43] Add indirect_implicit_ptr to dwarf_location class Zoran Zaric via Gdb-patches
2021-03-01 14:46 ` [PATCH 30/43] Add new computed struct value callback interface Zoran Zaric via Gdb-patches
2021-03-01 14:46 ` [PATCH 31/43] Add to_gdb_value method to DWARF entry class Zoran Zaric via Gdb-patches
2021-03-01 14:46 ` [PATCH 32/43] Change DWARF stack to use new dwarf_entry classes Zoran Zaric via Gdb-patches
2021-03-01 14:46 ` [PATCH 33/43] Remove old computed struct value callbacks Zoran Zaric via Gdb-patches
2021-03-01 14:46 ` [PATCH 34/43] Comments cleanup between expr.h and expr.c Zoran Zaric via Gdb-patches
2021-03-01 14:46 ` [PATCH 35/43] Remove dwarf_expr_context from expr.h interface Zoran Zaric via Gdb-patches
2021-03-01 14:46 ` [PATCH 36/43] Move read_addr_from_reg function to frame.c Zoran Zaric via Gdb-patches
2021-03-01 14:46 ` [PATCH 37/43] Add frame info check to DW_OP_reg operations Zoran Zaric via Gdb-patches
2021-03-01 14:46 ` [PATCH 38/43] Remove DWARF expression composition check Zoran Zaric via Gdb-patches
2021-03-01 14:46 ` [PATCH 39/43] Change back the symbol needs to use the evaluator Zoran Zaric via Gdb-patches
2021-03-01 14:46 ` [PATCH 40/43] Add support for any location description in CFI Zoran Zaric via Gdb-patches
2021-03-01 14:46 ` [PATCH 41/43] Add DWARF operations for byte and bit offset Zoran Zaric via Gdb-patches
2021-03-01 14:46 ` [PATCH 42/43] Add support for DW_OP_LLVM_undefined operation Zoran Zaric via Gdb-patches
2021-03-01 14:46 ` [PATCH 43/43] Add support for nested composite locations Zoran Zaric via Gdb-patches

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=5a2c4d75-2059-cc93-6581-59cde0d6ef40@polymtl.ca \
    --to=gdb-patches@sourceware.org \
    --cc=Zoran.Zaric@amd.com \
    --cc=simon.marchi@polymtl.ca \
    /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