[PATCH] gdb/dwarf: turn some errors into asserts

[PATCH] gdb/dwarf: turn some errors into asserts

[Simon Marchi]

From: Simon Marchi <simon.marchi@polymtl.ca>

There are some nullptr checks on the return value of load_cu that I
think can't possibly ever be true.  load_cu returns nullptr only if the
CU is "dummy".  A "dummy" CU is one that has a header but no DIE.  The
body (the part after the header, included in the unit reported by the
header) is either empty or consists only of zeroes (null entries).

In the various spots I modified we are doing some work (e.g. evaluating an
expression) in the context of PER_CU, so PER_CU can't possibly be empty
/ dummy.

Change those conditions to asserts.  No behavior change expected, unless
my reasoning is incorrect.

Change-Id: Ic8b0614a15d82d0aaadb8182ee641662cc71dc54
---
 gdb/dwarf2/read.c | 27 +++++++++------------------
 1 file changed, 9 insertions(+), 18 deletions(-)

diff --git a/gdb/dwarf2/read.c b/gdb/dwarf2/read.c
index 036ec7e6337f..28a03ab1fbe8 100644
--- a/gdb/dwarf2/read.c
+++ b/gdb/dwarf2/read.c
@@ -17134,14 +17134,9 @@ dwarf2_fetch_die_loc_sect_off (sect_offset sect_off, dwarf2_per_cu *per_cu,
   if (cu == nullptr)
     cu = load_cu (per_cu, per_objfile, false);
 
-  if (cu == nullptr)
-    {
-      /* We shouldn't get here for a dummy CU, but don't crash on the user.
-	 Instead just throw an error, not much else we can do.  */
-      error (_(DWARF_ERROR_PREFIX
-	       "Dummy CU at %s referenced [in module %s]"),
-	     sect_offset_str (sect_off), objfile_name (objfile));
-    }
+  /* We know this can't be a dummy CU, since we're executing something from
+     it.  */
+  gdb_assert (cu != nullptr);
 
   die_info *die = follow_die_offset ({ &cu->section (), sect_off }, &cu);
   if (die == nullptr)
@@ -17274,14 +17269,9 @@ dwarf2_fetch_constant_bytes (sect_offset sect_off,
   if (cu == nullptr)
     cu = load_cu (per_cu, per_objfile, false);
 
-  if (cu == nullptr)
-    {
-      /* We shouldn't get here for a dummy CU, but don't crash on the user.
-	 Instead just throw an error, not much else we can do.  */
-      error (_(DWARF_ERROR_PREFIX
-	       "Dummy CU at %s referenced [in module %s]"),
-	     sect_offset_str (sect_off), objfile_name (objfile));
-    }
+  /* We know this can't be a dummy CU, since we're executing something from
+     it.  */
+  gdb_assert (cu != nullptr);
 
   die_info *die = follow_die_offset ({ &cu->section (), sect_off }, &cu);
   if (!die)
@@ -17377,8 +17367,9 @@ dwarf2_fetch_die_type_sect_off (sect_offset sect_off, dwarf2_per_cu *per_cu,
   if (cu == nullptr)
     cu = load_cu (per_cu, per_objfile, false);
 
-  if (cu == nullptr)
-    return nullptr;
+  /* We know this can't be a dummy CU, since we're executing something from
+     it.  */
+  gdb_assert (cu != nullptr);
 
   die_info *die = follow_die_offset ({ &cu->section (), sect_off }, &cu);
   if (!die)

base-commit: 6660ba84d5484a51e09985c741091152f9febe5c
-- 
2.52.0

Re: [PATCH] gdb/dwarf: turn some errors into asserts

[Kevin Buettner]

On Mon, 26 Jan 2026 12:01:17 -0500
Simon Marchi <simon.marchi@efficios.com> wrote:

> From: Simon Marchi <simon.marchi@efficios.com>
> To: gdb-patches@sourceware.org
> Cc: Simon Marchi <simon.marchi@polymtl.ca>
> Subject: [PATCH] gdb/dwarf: turn some errors into asserts
> Date: Mon, 26 Jan 2026 12:01:17 -0500
> 
> From: Simon Marchi <simon.marchi@polymtl.ca>
> 
> There are some nullptr checks on the return value of load_cu that I
> think can't possibly ever be true.  load_cu returns nullptr only if the
> CU is "dummy".  A "dummy" CU is one that has a header but no DIE.  The
> body (the part after the header, included in the unit reported by the
> header) is either empty or consists only of zeroes (null entries).
> 
> In the various spots I modified we are doing some work (e.g. evaluating an
> expression) in the context of PER_CU, so PER_CU can't possibly be empty
> / dummy.
> 
> Change those conditions to asserts.  No behavior change expected, unless
> my reasoning is incorrect.

I agree with your reasoning.

Approved-by: Kevin Buettner <kevinb@redhat.com>

Re: [PATCH] gdb/dwarf: turn some errors into asserts

[Simon Marchi]

On 2026-02-02 12:42, Kevin Buettner wrote:
> On Mon, 26 Jan 2026 12:01:17 -0500
> Simon Marchi <simon.marchi@efficios.com> wrote:
> 
>> From: Simon Marchi <simon.marchi@efficios.com>
>> To: gdb-patches@sourceware.org
>> Cc: Simon Marchi <simon.marchi@polymtl.ca>
>> Subject: [PATCH] gdb/dwarf: turn some errors into asserts
>> Date: Mon, 26 Jan 2026 12:01:17 -0500
>>
>> From: Simon Marchi <simon.marchi@polymtl.ca>
>>
>> There are some nullptr checks on the return value of load_cu that I
>> think can't possibly ever be true.  load_cu returns nullptr only if the
>> CU is "dummy".  A "dummy" CU is one that has a header but no DIE.  The
>> body (the part after the header, included in the unit reported by the
>> header) is either empty or consists only of zeroes (null entries).
>>
>> In the various spots I modified we are doing some work (e.g. evaluating an
>> expression) in the context of PER_CU, so PER_CU can't possibly be empty
>> / dummy.
>>
>> Change those conditions to asserts.  No behavior change expected, unless
>> my reasoning is incorrect.
> 
> I agree with your reasoning.
> 
> Approved-by: Kevin Buettner <kevinb@redhat.com>
> 

Pushed, thanks!

Simon