From: Guinevere Larsen <blarsen@redhat.com>
To: Guinevere Larsen <blarsen@redhat.com>,
"gdb-patches@sourceware.org" <gdb-patches@sourceware.org>
Subject: Re: [PATCH v2 3/5] gdb: Migrate frame unwinders to use C++ classes
Date: Fri, 12 Apr 2024 11:40:31 -0300 [thread overview]
Message-ID: <89a78bd0-504b-4fd5-a085-066c2efe492b@redhat.com> (raw)
In-Reply-To: <20240408201915.1482831-4-blarsen@redhat.com>
[-- Attachment #1: Type: text/plain, Size: 3770 bytes --]
I mentioned as a reply to the cover letter that this patch series had a
crash in arm32 bit. This patch had the problem and a quick fix I inlined.
Trimming the diff for readability.
On 4/8/24 17:19, Guinevere Larsen wrote:
> diff --git a/gdb/tramp-frame.c b/gdb/tramp-frame.c
> index 2fcc3814e59..7fee4d5c71d 100644
> --- a/gdb/tramp-frame.c
> +++ b/gdb/tramp-frame.c
> @@ -123,7 +123,7 @@ tramp_frame_sniffer (const struct frame_unwind *self,
> const frame_info_ptr &this_frame,
> void **this_cache)
> {
> - const struct tramp_frame *tramp = self->unwind_data->tramp_frame;
> + const struct tramp_frame *tramp = self->unwind_data ()->tramp_frame;
> CORE_ADDR pc = get_frame_pc (this_frame);
> CORE_ADDR func;
> struct tramp_frame_cache *tramp_cache;
> @@ -142,6 +142,46 @@ tramp_frame_sniffer (const struct frame_unwind *self,
> return 1;
> }
>
> +class frame_unwind_trampoline : public frame_unwind
> +{
> +private:
> + frame_prev_arch_ftype *prev_arch_p;
> +public:
> + frame_unwind_trampoline (enum frame_type t, const struct frame_data *d,
> + frame_prev_arch_ftype *pa)
> + : frame_unwind ("trampoline", t, FRAME_UNWIND_GDB, d), prev_arch_p (pa)
> + { }
> +
> + int sniffer(const frame_unwind *self, const frame_info_ptr &this_frame,
> + void **this_prologue_cache) const override
> + {
> + return tramp_frame_sniffer (self, this_frame, this_prologue_cache);
> + }
> + void this_id (const frame_info_ptr &this_frame, void **this_prologue_cache,
> + struct frame_id *id) const override
> + {
> + tramp_frame_this_id (this_frame, this_prologue_cache, id);
> + }
> + struct value *prev_register (const frame_info_ptr &this_frame,
> + void **this_prologue_cache,
> + int regnum) const override
> + {
> + return tramp_frame_prev_register (this_frame, this_prologue_cache, regnum);
> + }
> +
> + struct gdbarch *prev_arch (const frame_info_ptr &this_frame,
> + void **this_prologue_cache) const override
> + {
This place should have:
if (prev_arch_p == nullptr)
error (_("No prev_arch callback installed"));
I've fixed this locally, but since this patch is over 100Kb, I would
prefer to not spam the list :)
--
Cheers,
Guinevere Larsen
She/Her/Hers
> + return prev_arch_p (this_frame, this_prologue_cache);
> + }
> +
> + /* FIXME: This should have a proper algorithm to deallocate the cache,
> + otherwise memory is leaked. This method is empty here just so the
> + migration to c++ classes doesn't add regressions. */
> + void dealloc_cache (frame_info *self, void *this_cache) const override
> + { }
> +};
> +
> void
> tramp_frame_prepend_unwinder (struct gdbarch *gdbarch,
> const struct tramp_frame *tramp_frame)
> @@ -160,16 +200,11 @@ tramp_frame_prepend_unwinder (struct gdbarch *gdbarch,
> gdb_assert (tramp_frame->insn_size <= sizeof (tramp_frame->insn[0].bytes));
>
> data = GDBARCH_OBSTACK_ZALLOC (gdbarch, struct frame_data);
> - unwinder = GDBARCH_OBSTACK_ZALLOC (gdbarch, struct frame_unwind);
> -
> data->tramp_frame = tramp_frame;
> - unwinder->type = tramp_frame->frame_type;
> - unwinder->unwind_data = data;
> - unwinder->unwinder_class = FRAME_UNWIND_GDB;
> - unwinder->sniffer = tramp_frame_sniffer;
> - unwinder->stop_reason = default_frame_unwind_stop_reason;
> - unwinder->this_id = tramp_frame_this_id;
> - unwinder->prev_register = tramp_frame_prev_register;
> - unwinder->prev_arch = tramp_frame->prev_arch;
> +
> + unwinder = obstack_new <frame_unwind_trampoline> (gdbarch_obstack (gdbarch),
> + tramp_frame->frame_type,
> + data,
> + tramp_frame->prev_arch);
> frame_unwind_prepend_unwinder (gdbarch, unwinder);
> }
next prev parent reply other threads:[~2024-04-12 14:40 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-04-08 20:19 [PATCH v2 0/5] Modernize frame unwinders and add disable feature Guinevere Larsen
2024-04-08 20:19 ` [PATCH v2 1/5] gdb: make gdbarch store a vector of frame unwinders Guinevere Larsen
2024-04-08 20:19 ` [PATCH v2 2/5] gdb: add "unwinder class" to " Guinevere Larsen
2024-04-08 20:19 ` [PATCH v2 3/5] gdb: Migrate frame unwinders to use C++ classes Guinevere Larsen
2024-04-12 14:40 ` Guinevere Larsen [this message]
2024-04-08 20:19 ` [PATCH v2 4/5] gdb: introduce ability to disable frame unwinders Guinevere Larsen
2024-04-09 2:47 ` Eli Zaretskii
2024-04-09 13:33 ` Guinevere Larsen
2024-04-09 14:08 ` Eli Zaretskii
2024-04-08 20:19 ` [PATCH v2 5/5] gdb/testsuite: Test for a backtrace through object without debuginfo Guinevere Larsen
2024-04-09 17:23 ` [PATCH v2 0/5] Modernize frame unwinders and add disable feature Guinevere Larsen
2024-04-24 12:47 ` [PING][PATCH " Guinevere Larsen
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=89a78bd0-504b-4fd5-a085-066c2efe492b@redhat.com \
--to=blarsen@redhat.com \
--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;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).