From: Andrew Burgess <aburgess@redhat.com>
To: Carl Love <cel@linux.ibm.com>,
Ulrich Weigand <Ulrich.Weigand@de.ibm.com>,
"gdb-patches@sourceware.org" <gdb-patches@sourceware.org>,
Keith Seitz <keiths@redhat.com>
Cc: cel@us.ibm.com
Subject: Re: [PATCH 1/2, ver3] PowerPC, Fix-test-gdb.base-store.exp
Date: Wed, 08 Nov 2023 10:54:31 +0000 [thread overview]
Message-ID: <874jhwtt08.fsf@redhat.com> (raw)
In-Reply-To: <77710bdab4d396d3a4d001bfb2217d00e686c823.camel@linux.ibm.com>
Carl Love <cel@linux.ibm.com> writes:
> Andrew, Ulrich:
>
> Per the discussions, the commit message has been updated to make it
> clear that the patch fixes the DWARF register mapping which fixes two
> of the test failures in step.exp. The fix exposed a dormant issue with
> the signal handling on PowerPC. The underlying issue which is then
> exposed causes the signal handling test sigstep.exp to fail. The test
> sigstep.exp tests stepping into a signal handler. The patch also fixes
> the underlying signal handler issue.
>
> This new verion of the patch only updates the commit log to make it
> clear the patch is fixing multiple issues.
>
> Please let me know if the commit description is clear and acceptable.
>
> Thanks.
>
> Carl
> ------------------------------------------------------------
> rs6000, Fix Linux DWARF register mapping
>
> Overview of issues fixed by the patch.
>
> The primary issue this patch fixes is the DWARF register mapping for
> Linux. The changes in ppc-linux-tdep.c fix the DWARF register mapping
> issues. The register mapping issue is responsible for two of the
> five regression bugs seen in gdb.base/store.exp.
>
> Once the register mapping is fixed, an underlying issue with the unwinding
> of the signal trampoline in common-code in ifrun.c was found. This
> underlying bug is best described by Ulrich in the following description.
>
> The unwinder bug shows up on platforms where the kernel uses a trampoline
> to dispatch "calls to" the signal handler (not just *returns from* the
> signal handler). Many platforms use a trampoline for signal return, and
> that is working fine, but the only platform I'm (Ulrich) aware of that
> uses a trampoline for signal handler calls is (recent kernels for)
> PowerPC. I believe the rationale for using a trampoline here
> is to improve performance by avoiding unbalancing of the
> branch predictor's call/return stack.
>
> However, on PowerPC the bug is dormant as well as it is hidden
> by *another* bug that prevents correct unwinding out of the
> signal trampoline. This is because the custom CFI for the
> trampoline uses a register number (VSCR) that is not ever used
> by compiler-generated CFI, and that particular register is
> mapped to an invalid number by the current PowerPC DWARF mapper.
>
> The underlying unwinder bug is exposed by the "new" regression failures
> in gdb.base/sigstep.exp. These failures were previously masked by
> the fact that GDB was not seeing a valid frame when it tried to unwind
> the frames. The sigstep.exp test is specifically testing stepping into
> a signal handler. With the correct DWARF register mapping in place,
> specifically the VSCR mapping, the signal trampoline code now unwinds to a
> valid frame exposing the pre-existing bug in how the signal handler on
> PowerPC works. The one line change infrun.c fixes the exiting bug in
> the common-code for platforms that use a trampoline to dispatch calls
> to the signal handler by not stopping in the SIGTRAMP_FRAME.
>
> Detailed description of the DWARF register mapping fix.
>
> The PowerPC DWARF register mapping is the same for the .eh_frame and
> .debug_frame on Linux. PowerPC uses different mapping for .eh_frame and
> .debug_frame on other operating systems. The current GDB support for
> mapping the DWARF registers in rs6000_linux_dwarf2_reg_to_regnum and
> rs6000_adjust_frame_regnum file gdb/rs6000-tdep.c is not correct for Linux.
> The files have some legacy mappings for spe_acc, spefscr, EV which was
> removed from GCC in 2017.
>
> This patch adds a two new functions rs6000_linux_dwarf2_reg_to_regnum,
> and rs6000_linux_adjust_frame_regnum in file gdb/ppc-linux-tdep.c to handle
> the DWARF register mappings on Linux. Function
> rs6000_linux_dwarf2_reg_to_regnum is installed for both gdb_dwarf_to_regnum
> and gdbarch_stab_reg_to_regnum since the mappings are the same.
>
> The ppc_linux_init_abi function in gdb/ppc-linux-tdep.c is updated to
> call set_gdbarch_dwarf2_reg_to_regnum map the new function
> rs6000_linux_dwarf2_reg_to_regnum for the architecture. Similarly,
> dwarf2_frame_set_adjust_regnum is called to map
> rs6000_linux_adjust_frame_regnum into the architecture.
>
> Additional detail on the signal handling fix.
>
> The specific sequence of events for handling a signal on most
> architectures is as follows:
>
> 1) Some code is running when a signal arrives.
> 2) The kernel handles the signal and dispatches to the handler.
> ...
>
> However on PowerPC the sequence of events is:
>
> 1) Some code is running when a signal arrives.
> 2) The kernel handles the signal and dispatches to the trampoline.
> 3) The trampoline performs a normal function call to the handler.
> ...
>
> We want the "nexti" to step into, not over, signal handlers invoked by
> the kernel. This is the case for most platforms as the kernel puts a
> signal trampoline frame onto the stack to handle proper return after the
> handler. However, on some platforms such as PowerPC, the kernel actually
> uses a trampoline to handle *invocation* of the handler. We do not
> want GDB to stop in the SIGTRAMP_FRAME. The issue is fixed in function
> process_event_stop_test by adding a check that the frame is not a
> SIGTRAMP_FRAME to the if statement to stop in a subroutine call. This
> prevents GDB from erroneously detecting the trampoline invocation as a
> subroutine call.
>
> This patch fixes two regression test failures in gdb.base/store.exp.
>
> The patch then fixes an exposed, dormant, signal handling issue that
> is exposed in the signal handling test gdb.base/sigstep.exp.
>
> The patch has been tested on Power 8 LE/BE, Power 9 LE/BE, Power 10 with
> no new regressions. Note, only two of the five failures in store.exp
> are fixed. The remaining three failures are fixed in a following
> patch.
Carl,
Sorry for the slow delay, I had to find some time to play with this a
little. But thanks to the expanded explanation, it's now clear what's
going on. This all looks great.
Approved-By: Andrew Burgess <aburgess@redhat.com>
Thanks,
Andrew
> ---
> gdb/infrun.c | 13 +++++++++++
> gdb/ppc-linux-tdep.c | 53 ++++++++++++++++++++++++++++++++++++++++++++
> 2 files changed, 66 insertions(+)
>
> diff --git a/gdb/infrun.c b/gdb/infrun.c
> index 4730d290442..922d8a6913d 100644
> --- a/gdb/infrun.c
> +++ b/gdb/infrun.c
> @@ -7334,8 +7334,21 @@ process_event_stop_test (struct execution_control_state *ecs)
> initial outermost frame, before sp was valid, would
> have code_addr == &_start. See the comment in frame_id::operator==
> for more. */
> +
> + /* We want "nexti" to step into, not over, signal handlers invoked
> + by the kernel, therefore this subroutine check should not trigger
> + for a signal handler invocation. On most platforms, this is already
> + not the case, as the kernel puts a signal trampoline frame onto the
> + stack to handle proper return after the handler, and therefore at this
> + point, the current frame is a grandchild of the step frame, not a
> + child. However, on some platforms, the kernel actually uses a
> + trampoline to handle *invocation* of the handler. In that case,
> + when executing the first instruction of the trampoline, this check
> + would erroneously detect the trampoline invocation as a subroutine
> + call. Fix this by checking for SIGTRAMP_FRAME. */
> if ((get_stack_frame_id (frame)
> != ecs->event_thread->control.step_stack_frame_id)
> + && get_frame_type (frame) != SIGTRAMP_FRAME
> && ((frame_unwind_caller_id (get_current_frame ())
> == ecs->event_thread->control.step_stack_frame_id)
> && ((ecs->event_thread->control.step_stack_frame_id
> diff --git a/gdb/ppc-linux-tdep.c b/gdb/ppc-linux-tdep.c
> index 784dafa59db..8d975336fe5 100644
> --- a/gdb/ppc-linux-tdep.c
> +++ b/gdb/ppc-linux-tdep.c
> @@ -83,6 +83,7 @@
> #include "features/rs6000/powerpc-isa207-vsx64l.c"
> #include "features/rs6000/powerpc-isa207-htm-vsx64l.c"
> #include "features/rs6000/powerpc-e500l.c"
> +#include "dwarf2/frame.h"
>
> /* Shared library operations for PowerPC-Linux. */
> static struct target_so_ops powerpc_so_ops;
> @@ -2088,6 +2089,49 @@ ppc_linux_displaced_step_prepare (gdbarch *arch, thread_info *thread,
> return per_inferior->disp_step_buf->prepare (thread, displaced_pc);
> }
>
> +/* Convert a Dwarf 2 register number to a GDB register number for Linux. */
> +
> +static int
> +rs6000_linux_dwarf2_reg_to_regnum (struct gdbarch *gdbarch, int num)
> +{
> + ppc_gdbarch_tdep *tdep = gdbarch_tdep<ppc_gdbarch_tdep>(gdbarch);
> +
> + if (0 <= num && num <= 31)
> + return tdep->ppc_gp0_regnum + num;
> + else if (32 <= num && num <= 63)
> + return tdep->ppc_fp0_regnum + (num - 32);
> + else if (77 <= num && num < 77 + 32)
> + return tdep->ppc_vr0_regnum + (num - 77);
> + else
> + switch (num)
> + {
> + case 65:
> + return tdep->ppc_lr_regnum;
> + case 66:
> + return tdep->ppc_ctr_regnum;
> + case 76:
> + return tdep->ppc_xer_regnum;
> + case 109:
> + return tdep->ppc_vrsave_regnum;
> + case 110:
> + return tdep->ppc_vrsave_regnum - 1; /* vscr */
> + }
> +
> + /* Unknown DWARF register number. */
> + return -1;
> +}
> +
> +/* Translate a .eh_frame register to DWARF register, or adjust a
> + .debug_frame register. */
> +
> +static int
> +rs6000_linux_adjust_frame_regnum (struct gdbarch *gdbarch, int num,
> + int eh_frame_p)
> +{
> + /* Linux uses the same numbering for .debug_frame numbering as .eh_frame. */
> + return num;
> +}
> +
> static void
> ppc_linux_init_abi (struct gdbarch_info info,
> struct gdbarch *gdbarch)
> @@ -2135,6 +2179,15 @@ ppc_linux_init_abi (struct gdbarch_info info,
> set_gdbarch_stap_is_single_operand (gdbarch, ppc_stap_is_single_operand);
> set_gdbarch_stap_parse_special_token (gdbarch,
> ppc_stap_parse_special_token);
> + /* Linux DWARF register mapping is different from the other OSes. */
> + set_gdbarch_dwarf2_reg_to_regnum (gdbarch,
> + rs6000_linux_dwarf2_reg_to_regnum);
> + /* Note on Linux the mapping for the DWARF registers and the stab registers
> + use the same numbers. Install rs6000_linux_dwarf2_reg_to_regnum for the
> + stab register mappings as well. */
> + set_gdbarch_stab_reg_to_regnum (gdbarch,
> + rs6000_linux_dwarf2_reg_to_regnum);
> + dwarf2_frame_set_adjust_regnum (gdbarch, rs6000_linux_adjust_frame_regnum);
>
> if (tdep->wordsize == 4)
> {
> --
> 2.37.2
next prev parent reply other threads:[~2023-11-08 10:54 UTC|newest]
Thread overview: 27+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-10-12 14:51 Carl Love
2023-10-12 14:58 ` [Patch 1/2] " Carl Love
2023-10-13 20:34 ` Keith Seitz
2023-10-13 21:00 ` Carl Love
2023-10-16 11:12 ` Ulrich Weigand
2023-10-16 14:31 ` Andrew Burgess
2023-10-16 15:51 ` Carl Love
2023-10-19 15:54 ` Carl Love
2023-10-24 8:50 ` Andrew Burgess
2023-10-24 16:05 ` Carl Love
2023-10-20 18:08 ` [PATCH 1/2, ver2] " Carl Love
2023-10-24 9:30 ` Andrew Burgess
2023-10-25 13:24 ` Ulrich Weigand
2023-10-30 9:45 ` Andrew Burgess
2023-10-30 16:44 ` Ulrich Weigand
2023-10-30 17:16 ` Carl Love
2023-10-30 17:25 ` [PATCH 1/2, ver3] " Carl Love
2023-11-06 18:24 ` Carl Love
2023-11-08 10:54 ` Andrew Burgess [this message]
2023-10-12 15:00 ` [PATCH 2/2] " Carl Love
2023-10-13 20:35 ` Keith Seitz
2023-10-13 21:00 ` Carl Love
2023-10-16 11:13 ` Ulrich Weigand
2023-10-16 14:36 ` Andrew Burgess
2023-10-16 15:51 ` Carl Love
2023-10-20 18:08 ` Carl Love
2023-10-24 8:53 ` Andrew Burgess
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=874jhwtt08.fsf@redhat.com \
--to=aburgess@redhat.com \
--cc=Ulrich.Weigand@de.ibm.com \
--cc=cel@linux.ibm.com \
--cc=cel@us.ibm.com \
--cc=gdb-patches@sourceware.org \
--cc=keiths@redhat.com \
/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).