Mirror of the gdb-patches mailing list
 help / color / mirror / Atom feed
From: Ronald Hecht <ronald.hecht@gmx.de>
To: gdb-patches@sourceware.org
Cc: Ronald Hecht <ronald.hecht@gmx.de>
Subject: [PATCH v3] gdb: z80: Fix endless loop in frame unwinder and validate saved register types
Date: Mon, 15 Jun 2026 09:04:01 +0200	[thread overview]
Message-ID: <20260615070401.8671-1-ronald.hecht@gmx.de> (raw)
In-Reply-To: <20260614150230.2fd33d6e@f44-mesa-1>

In z80_frame_unwind_cache, the loop scanning for the frame base pointer
(for (;; ++sp)) could run into an endless loop or scan too far if the
termination condition (sp < this_base) was not met due to corrupted
or unexpected stack layouts. This patch introduces a loop_count to
limit the scan to a maximum of 2 * addr_len iterations.

Additionally, when iterating through saved registers to adjust their
offsets into concrete addresses, the code now explicitly checks if
the register actually holds an address using is_addr() before calling
addr(). This prevents potential assertions or undefined behavior for
registers that do not contain valid address data.

gdb/ChangeLog:

	* z80-tdep.c (z80_frame_unwind_cache): Limit stack scanning
	to 2 * addr_len iterations to prevent endless loops. Check
	is_addr() before adjusting saved register addresses.
---
 gdb/z80-tdep.c | 29 ++++++++++++++++++++++-------
 1 file changed, 22 insertions(+), 7 deletions(-)

diff --git a/gdb/z80-tdep.c b/gdb/z80-tdep.c
index 3a0d5f7e..6828894f 100644
--- a/gdb/z80-tdep.c
+++ b/gdb/z80-tdep.c
@@ -595,17 +595,27 @@ z80_frame_unwind_cache (const frame_info_ptr &this_frame,
     {
       CORE_ADDR addr;
       CORE_ADDR sp;
-      CORE_ADDR sp_mask = (1 << gdbarch_ptr_bit(gdbarch)) - 1;
+      CORE_ADDR addr_space_max = (1 << gdbarch_ptr_bit(gdbarch)) - 1;
       enum bfd_endian byte_order = gdbarch_byte_order (gdbarch);
+      int loop_count = 0;
       /* Assume that the FP is this frame's SP but with that pushed
 	 stack space added back.  */
       this_base = get_frame_register_unsigned (this_frame, Z80_SP_REGNUM);
       sp = this_base + info->size;
       for (;; ++sp)
 	{
-	  sp &= sp_mask;
-	  if (sp < this_base)
-	    { /* overflow, looks like end of stack */
+	  /* Limit the scan to 2 * addr_len iterations.  If the unwinder's
+	     frame size calculation is slightly off (e.g. due to unpopped
+	     arguments or temporary pushes), the return address might be
+	     hidden a few bytes deeper.  Scanning up to 2 * addr_len bytes
+	     comfortably covers two misplaced pointer-sized values.
+	     Scanning further drastically increases the risk of false
+	     positives: we might wander into the caller's local variables,
+	     hit a random CALL or conditional CALL opcode, and generate a
+	     corrupted backtrace.  It also prevents massive slow-downs on
+	     remote serial targets if the stack is severely corrupted.  */
+	  if (++loop_count > 2 * addr_len || sp > addr_space_max)
+	    { /* Limit reached or end of address space, assume end of stack.  */
 	      sp = this_base + info->size;
 	      break;
 	    }
@@ -644,9 +654,14 @@ z80_frame_unwind_cache (const frame_info_ptr &this_frame,
   /* Adjust all the saved registers so that they contain addresses and not
      offsets.  */
   for (i = 0; i < gdbarch_num_regs (gdbarch) - 1; i++)
-    if (info->saved_regs[i].addr () > 0)
-      info->saved_regs[i].set_addr
-	(info->prev_sp - info->saved_regs[i].addr () * addr_len);
+    {
+      if (info->saved_regs[i].is_addr ())
+	{
+	  if (info->saved_regs[i].addr () > 0)
+	    info->saved_regs[i].set_addr
+	      (info->prev_sp - info->saved_regs[i].addr () * addr_len);
+	}
+    }
 
   /* Except for the startup code, the return PC is always saved on
      the stack and is at the base of the frame.  */
-- 
2.43.0


  parent reply	other threads:[~2026-06-15  7:10 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-06-05 18:56 [PATCH] gdb: z80: Fix endless backtrace loop and assertion crashes Ronald Hecht
2026-06-06 19:40 ` Kevin Buettner
2026-06-07  7:36   ` Ronald Hecht
2026-06-07  7:48   ` [PATCH v2] " Ronald Hecht
2026-06-14 22:06     ` Kevin Buettner
2026-06-15  6:49       ` Ronald Hecht
2026-06-15  7:04       ` Ronald Hecht [this message]
2026-06-20 20:11         ` [PATCH v3] gdb: z80: Fix endless loop in frame unwinder and validate saved register types Kevin Buettner
2026-06-23  7:16           ` Ronald Hecht
2026-07-09  3:52             ` Kevin Buettner

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=20260615070401.8671-1-ronald.hecht@gmx.de \
    --to=ronald.hecht@gmx.de \
    --cc=gdb-patches@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