Bug 24892 - [-fPIE/-pie] FAIL: gdb.arch/amd64-entry-value-param.exp: call 1: p y
Summary: [-fPIE/-pie] FAIL: gdb.arch/amd64-entry-value-param.exp: call 1: p y
Status: RESOLVED FIXED
Alias: None
Product: gdb
Classification: Unclassified
Component: symtab (show other bugs)
Version: HEAD
: P2 normal
Target Milestone: 12.1
Assignee: Not yet assigned to anyone
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2019-08-08 14:33 UTC by Tom de Vries
Modified: 2021-10-05 09:43 UTC (History)
1 user (show)

See Also:
Host:
Target:
Build:
Last reconfirmed:

Attachments
Tentative patch (1.17 KB, patch)
2019-08-09 06:28 UTC, Tom de Vries 		Details | Diff
View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Tom de Vries 2019-08-08 14:33:52 UTC 
With target board unix/-fPIE/-pie and gdb.arch/amd64-entry-value-param.exp, we get:
...
FAIL: gdb.arch/amd64-entry-value-param.exp: call 1: p y
FAIL: gdb.arch/amd64-entry-value-param.exp: call 1: p b
FAIL: gdb.arch/amd64-entry-value-param.exp: call 2: p y
FAIL: gdb.arch/amd64-entry-value-param.exp: call 2: p b
FAIL: gdb.arch/amd64-entry-value-param.exp: call 3: p y
FAIL: gdb.arch/amd64-entry-value-param.exp: call 3: p b
...

Reproduce first FAIL on command line:
...
$ gdb ./build/gdb/testsuite/outputs/gdb.arch/amd64-entry-value-param/amd64-entry-value-param -batch -ex "b foo" -ex run -ex "p y"
Breakpoint 1 at 0x650: file gdb.arch/amd64-entry-value-param.c, line 26.

Breakpoint 1, foo (x=10, y=<optimized out>, z=3) at gdb.arch/amd64-entry-value-param.c:26
26      gdb.arch/amd64-entry-value-param.c: Bestand of map bestaat niet.
$1 = <optimized out>
...

as opposed to this without -fPIE/-pie:
...
$ gdb ./build/gdb/testsuite/outputs/gdb.arch/amd64-entry-value-param/amd64-entry-value-param -batch -ex "b foo" -ex run -ex "p y"
Breakpoint 1 at 0x4004c0: file gdb.arch/amd64-entry-value-param.c, line 26.

Breakpoint 1, foo (x=x@entry=10, y=2, z=3) at gdb.arch/amd64-entry-value-param.c:26
26      gdb.arch/amd64-entry-value-param.c: Bestand of map bestaat niet.
$1 = 2
...
Comment 1 Tom de Vries 2019-08-08 15:01:42 UTC 
Setting a breakpoint on throw_error, we get:
...
Thread 1 "gdb" hit Breakpoint 2, throw_error (error=NO_ENTRY_VALUE_ERROR, 
    fmt=0xb57318 "DW_OP_entry_value resolving cannot find DW_TAG_call_site %s in %s")
    at src/gdb/gdbsupport/common-exceptions.c:222
222       va_start (args, fmt);
(gdb) bt
#0  throw_error (error=NO_ENTRY_VALUE_ERROR, 
    fmt=0xb57318 "DW_OP_entry_value resolving cannot find DW_TAG_call_site %s in %s")
    at src/gdb/gdbsupport/common-exceptions.c:222
#1  0x0000000000479158 in call_site_for_pc (gdbarch=0x1a9a730, pc=93824992233083)
    at src/gdb/block.c:237
#2  0x000000000059b023 in call_site_find_chain_1 (gdbarch=0x1a9a730, caller_pc=93824992233083, 
    callee_pc=93824992233040) at src/gdb/dwarf2loc.c:1097
#3  0x000000000059b3fc in call_site_find_chain (gdbarch=0x1a9a730, caller_pc=93824992233083, 
    callee_pc=93824992233040) at src/gdb/dwarf2loc.c:1201
#4  0x000000000058c41e in dwarf2_tailcall_sniffer_first (this_frame=0x1a4db00, tailcall_cachep=0x1a4dc08, 
    entry_cfa_sp_offsetp=0x0) at src/gdb/dwarf2-frame-tailcall.c:391
#5  0x000000000058f3f8 in dwarf2_frame_prev_register (this_frame=0x1a4db00, this_cache=0x1a4db18, regnum=16)
    at src/gdb/dwarf2-frame.c:1236
#6  0x0000000000629105 in frame_unwind_register_value (next_frame=0x1a4db00, regnum=16)
    at src/gdb/frame.c:1229
#7  0x0000000000628ca4 in frame_register_unwind (next_frame=0x1a4db00, regnum=16, optimizedp=0x7fffffffc5bc, 
    unavailablep=0x7fffffffc5b8, lvalp=0x7fffffffc5a8, addrp=0x7fffffffc5b0, realnump=0x7fffffffc5ac, 
    bufferp=0x7fffffffc5e8 "") at src/gdb/frame.c:1132
#8  0x0000000000628f95 in frame_unwind_register (next_frame=0x1a4db00, regnum=16, buf=0x7fffffffc5e8 "")
    at src/gdb/frame.c:1188
#9  0x000000000069a36e in i386_unwind_pc (gdbarch=0x1a9a730, next_frame=0x1a4db00)
    at src/gdb/i386-tdep.c:1967
#10 0x000000000063bf53 in gdbarch_unwind_pc (gdbarch=0x1a9a730, next_frame=0x1a4db00)
    at src/gdb/gdbarch.c:3085
#11 0x0000000000628616 in frame_unwind_pc (this_frame=0x1a4db00) at src/gdb/frame.c:928
#12 0x000000000062b536 in get_frame_pc (frame=0x1b8ef80) at src/gdb/frame.c:2399
#13 0x000000000062b614 in get_frame_address_in_block (this_frame=0x1b8ef80)
    at src/gdb/frame.c:2429
#14 0x00000000006d8352 in inline_frame_sniffer (self=0xbe83e0 <inline_frame_unwind>, this_frame=0x1b8ef80, 
    this_cache=0x1b8ef98) at src/gdb/inline-frame.c:204
#15 0x0000000000626491 in frame_unwind_try_unwinder (this_frame=0x1b8ef80, this_cache=0x1b8ef98, 
    unwinder=0xbe83e0 <inline_frame_unwind>) at src/gdb/frame-unwind.c:107
#16 0x00000000006265fe in frame_unwind_find_by_frame (this_frame=0x1b8ef80, this_cache=0x1b8ef98)
    at src/gdb/frame-unwind.c:165
#17 0x0000000000627a57 in compute_frame_id (fi=0x1b8ef80) at src/gdb/frame.c:546
#18 0x000000000062a777 in get_prev_frame_if_no_cycle (this_frame=0x1a4db00)
    at src/gdb/frame.c:1927
#19 0x000000000062ae87 in get_prev_frame_always_1 (this_frame=0x1a4db00)
    at src/gdb/frame.c:2108
#20 0x000000000062aeaf in get_prev_frame_always (this_frame=0x1a4db00)
    at src/gdb/frame.c:2124
#21 0x000000000062b43a in get_prev_frame (this_frame=0x1a4db00) at src/gdb/frame.c:2376
#22 0x000000000059bbb4 in value_of_dwarf_reg_entry (type=0x1a164e0, frame=0x1a4db00, 
    kind=CALL_SITE_PARAMETER_DWARF_REG, kind_u=...) at src/gdb/dwarf2loc.c:1449
#23 0x000000000059bd5f in value_of_dwarf_block_entry (type=0x1a164e0, frame=0x1a4db00, block=0x1a30085 "U\aK", 
    block_len=1) at src/gdb/dwarf2loc.c:1501
#24 0x000000000059ffff in locexpr_read_variable_at_entry (symbol=0x1a64610, frame=0x1a4db00)
    at src/gdb/dwarf2loc.c:3543
#25 0x0000000000866ad8 in read_frame_arg (fp_opts=..., sym=0x1a64610, frame=0x1a4db00, argp=0x7fffffffcb80, 
    entryargp=0x7fffffffcba0) at src/gdb/stack.c:496
#26 0x00000000008675b7 in print_frame_args (fp_opts=..., func=0x1a645a0, frame=0x1a4db00, num=-1, 
    stream=0x1a6e070) at src/gdb/stack.c:818
#27 0x000000000086873e in print_frame (fp_opts=..., frame=0x1a4db00, print_level=0, print_what=SRC_AND_LOC, 
    print_args=1, sal=...) at src/gdb/stack.c:1332
#28 0x0000000000867d54 in print_frame_info (fp_opts=..., frame=0x1a4db00, print_level=0, print_what=SRC_AND_LOC, 
    print_args=1, set_current_sal=1) at src/gdb/stack.c:1031
#29 0x00000000008662d6 in print_stack_frame (frame=0x1a4db00, print_level=0, print_what=SRC_AND_LOC, 
    set_current_sal=1) at src/gdb/stack.c:305
#30 0x00000000006d1423 in print_stop_location (ws=0x7fffffffd020)
    at src/gdb/infrun.c:7809
#31 0x00000000006d148a in print_stop_event (uiout=0x1a18a00, displays=true)
    at src/gdb/infrun.c:7826
#32 0x00000000008fae0f in tui_on_normal_stop (bs=0x1ab9fe0, print_frame=1)
    at src/gdb/tui/tui-interp.c:99
#33 0x0000000000445d3a in std::_Function_handler<void (bpstats*, int), void (*)(bpstats*, int)>::_M_invoke(std::_Any_data const&, bpstats*&&, int&&) (__functor=..., __args#0=@0x7fffffffd140: 0x1ab9fe0, 
    __args#1=@0x7fffffffd13c: 1) at /usr/include/c++/8/bits/std_function.h:297
#34 0x00000000006d5a14 in std::function<void (bpstats*, int)>::operator()(bpstats*, int) const (this=0x131b220, 
    __args#0=0x1ab9fe0, __args#1=1) at /usr/include/c++/8/bits/std_function.h:687
#35 0x00000000006d51be in gdb::observers::observable<bpstats*, int>::notify (
    this=0x11f72a0 <gdb::observers::normal_stop>, args#0=0x1ab9fe0, args#1=1)
    at src/gdb/gdbsupport/observable.h:106
#36 0x00000000006d1b5e in normal_stop () at src/gdb/infrun.c:8098
#37 0x00000000006c8b87 in fetch_inferior_event (client_data=0x0)
    at src/gdb/infrun.c:3781
#38 0x00000000006b087b in inferior_event_handler (event_type=INF_REG_EVENT, client_data=0x0)
    at src/gdb/inf-loop.c:43
#39 0x000000000070223c in handle_target_event (error=0, client_data=0x0)
    at src/gdb/linux-nat.c:4352
#40 0x000000000060de3d in handle_file_event (file_ptr=0x1a5b670, ready_mask=1)
    at src/gdb/event-loop.c:731
#41 0x000000000060e3e0 in gdb_wait_for_event (block=0) at src/gdb/event-loop.c:857
#42 0x000000000060d2b1 in gdb_do_one_event () at src/gdb/event-loop.c:321
#43 0x00000000008d7cbb in wait_sync_command_done () at src/gdb/top.c:510
#44 0x00000000008d7d33 in maybe_wait_sync_command_done (was_sync=0) at src/gdb/top.c:527
#45 0x00000000008d8144 in execute_command (p=0x7fffffffe134 "", from_tty=0)
    at src/gdb/top.c:629
#46 0x0000000000726e66 in catch_command_errors (command=0x8d7d4c <execute_command(char const*, int)>, 
    arg=0x7fffffffe131 "run", from_tty=0) at src/gdb/main.c:372
#47 0x0000000000728065 in captured_main_1 (context=0x7fffffffda70)
    at src/gdb/main.c:1136
#48 0x00000000007281a5 in captured_main (data=0x7fffffffda70) at src/gdb/main.c:1161
#49 0x0000000000728210 in gdb_main (args=0x7fffffffda70) at src/gdb/main.c:1186
#50 0x00000000004123f9 in main (argc=13, argv=0x7fffffffdb78) at src/gdb/gdb.c:32
...
Comment 2 Tom de Vries 2019-08-08 15:06:35 UTC 
Hmm, AFAIU, we call read_call_site_scope and try to update the pc to have the relocated value:
...
  baseaddr = ANOFFSET (objfile->section_offsets, SECT_OFF_TEXT (objfile));
  ...
  pc = attr_value_as_address (attr) + baseaddr;
  pc = gdbarch_adjust_dwarf2_addr (gdbarch, pc);
...
but unfortunately, objfile_relocate1 has not been called yet, so baseaddr is 0.
Comment 3 Tom de Vries 2019-08-09 06:28:41 UTC 
Created attachment 11943 [details]
Tentative patch

This patch makes the test-case pass with -fPIE/-pie.