This page provides documentation of known vulnerabilities in libdwarf. We are concerned here with cases where corrupt (by accident or intention) DWARF can cause the library to get a fault (crash) which could expose the calling program to interception by malefactors. Dates (where known) are in ISO extended date format.
Some of the bugs reported here have a CVE assigned, for example CVE-2017-9052. These are reported on cve.org (or possibly the earlier cve.mitre.org). Search with "libdwarf" on cve.org for a list.
Git reference path names refer to object files
in the libdwarf regression test base.
The test files can be retrieved via anonymous access:
"git clone https://github.org/davea42/libdwarf-regressiontests"
A few bugs refer to https://bugzilla.redhat.com bug system entries and/or https://bugs.chromium.org in addition to showing the names of test files in the regression test base.
Vulnerabilities listed newest-first.
Vulnerabilities listed oldest-first.
Record count: 189
id: DW201412-001
cve: CVE-2014-9482
fuzzer:
datereported: 2014-12-31
reportedby: Adam Maris
vulnerability: Use after free vulnerability in Dwarfdump
product: dwarfdump
description: The use-after-free has no attached testcase anywhere. Due to recent tool advances (like coverity scan) we are confident this was fixed long ago.
bugzilla.redhat.com/show_bug.cgi?id=1177758 www.openwall.com/lists/oss-security/2014/12/31/3 www.openwall.com/lists/oss-security/2015/01/03/14
datefixed: 2018-01-01
references:
gitfixid:
tarrelease: libdwarf-20160507.tar.gz
id: DW201512-001
cve: CVE-2015-8750
fuzzer:
datereported: 2015-12-26
reportedby: Qixue Xiao (xqx)
vulnerability: Null pointer dereference in libdwarf
product: libdwarf
description: libdwarf 20151114 and earlier allows remote attackers to cause a denial of service (NULL pointer dereference and crash) via a debug_abbrev section marked NOBITS in an ELF file.
bugzilla.redhat.com/show_bug.cgi?id=1294264 www.openwall.com/lists/oss-security/2016/01/07/11
datefixed: 2015-12-31
references: regressiontests/xqx-c/awbug6.elf
gitfixid:
tarrelease: libdwarf-20160507.tar.gz
id: DW201512-002
cve: CVE-2015-8538
fuzzer:
datereported: 2015-12-14
reportedby: Adam Maris
vulnerability: Out-of-bounds read in dwarf_leb.c
product: libdwarf
description: libdwarf 20151114 and earlier allows remote attackers to cause a denial of service (NULL pointer dereference and crash) via a debug_abbrev section marked NOBITS in an ELF file. The CVE report mentions a reproducer object file but such is not present. Due to recent tool advances (like coverity scan) we are confident this was fixed long ago.
bugzilla.redhat.com/show_bug.cgi?id=1291299 www.openwall.com/lists/oss-security/2015/12/10/3
datefixed: 2018-01-01
references:
gitfixid:
tarrelease: libdwarf-20160507.tar.gz
id: DW201601-001
cve: CVE-2016-2091
fuzzer:
datereported: 2016-01-12
reportedby: Qixue Xiao
vulnerability: Out of bound read in dwarf_read_cie_fde_prefix()
product: libdwarf
description: Crashes the calling program. Requires a crafted object file.
*** DWARF CHECK: DW_DLE_DEBUG_FRAME_LENGTH_NOT_MULTIPLE len=0x00000010, len size=0x00000004, extn size=0x00000000, totl length=0x00000014, addr size=0x00000008, mod=0x00000004 must be zero in cie, offset 0x00000000. *** 7 ==53495== Invalid read of size 2 1 ==53495== at 0x4C2F7E0: memcpy@@GLIBC_2.14 (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so) 2 ==53495== by 0x43287F: dwarf_read_cie_fde_prefix (dwarf_frame2.c:934) 3 ==53495== by 0x431305: _dwarf_get_fde_list_internal (dwarf_frame2.c:268) 4 ==53495== by 0x42EB5F: dwarf_get_fde_list_eh (dwarf_frame.c:1101) 5 ==53495== by 0x41BABE: print_frames (print_frames.c:1835) 6 ==53495== by 0x40485B: process_one_file (dwarfdump.c:1323) 7 ==53495== by 0x403529: main (dwarfdump.c:630) www.openwall.com/lists/oss-security/2016/01/19/3 www.openwall.com/lists/oss-security/2016/05/28/8
datefixed: 2016-01-21
references: regressiontests/xqx-b/awbug5.elf
gitfixid: d9d40e4d802e626065ce37ff384dd69c43bc499
tarrelease: libdwarf-20160507.tar.gz
id: DW201601-002
cve: CVE-2016-2050
fuzzer:
datereported: 2016-01-19
reportedby: Qixue Xiao
vulnerability: Out of bound write in get_abbrev_array_info
product: libdwarf
description: Crashes the calling program. Requires a crafted object file.
valgrind ./dwarfdump -ka aw.elf ==5358== Memcheck, a memory error detector ==5358== Copyright (C) 2002-2013, and GNU GPL'd, by Julian Seward et al. ==5358== Using Valgrind-3.10.1 and LibVEX; rerun with -h for copyright info ==5358== Command: ../../llvm-codes/dwarf-20151114/dwarfdump/dwarfdump -ka aw.elf ==5358== ==5358== Invalid write of size 8 ==5358== at 0x40DA25: get_abbrev_array_info (in /home/xqx/test/libdwarf-test/llvm-codes/dwarf-20151114/dwarfdump/dwarfdump) ==5358== by 0x40FD92: print_one_die_section (in /home/xqx/test/libdwarf-test/llvm-codes/dwarf-20151114/dwarfdump/dwarfdump) www.openwall.com/lists/oss-security/2016/01/19/9 www.openwall.com/lists/oss-security/2016/01/25/3
datefixed: 2016-01-21
references: regressiontests/xqx-b/aw.elf
gitfixid: d9d40e4d802e626065ce37ff384dd69c43bc499
tarrelease: libdwarf-20160507.tar.gz
id: DW201605-001
cve: CVE-2016-5044
fuzzer:
datereported: 2016-05-02
reportedby: Yue Liu
vulnerability: A specially crafted DWARF section results in a duplicate free() in libdwarf and the calling application will crash.
product: libdwarf
description: In file dwarf_elf_access.c:1071
WRITE_UNALIGNED(dbg,target_section + offset,
&outval,sizeof(outval),reloc_size);
A crafted ELF file may lead to a large offset value, which bigger than the size of target_section heap chunk, then this WRITE_UNALIGNED() function will write the value of &outval out of the heap chunk. offset is a 64bit unsigned int value, so this is more than a heap overflow bug, but also a Out-of-Bound write bug. So WRITE_UNALIGNED() need more strictly checking to prevent this.
datefixed: 2016-05-04
references: regressiontests/liu/heapoverflow01.elf
https://bugzilla.redhat.com/show_bug.cgi?id=1332141
gitfixid: 98a3da1e8237fe0d45b67ef77f3fa5ed9ff0215f
tarrelease: libdwarf-20160507.tar.gz
id: DW201605-002
cve: CVE-2016-5043
fuzzer:
datereported: 2016-05-02
reportedby: Yue Liu
vulnerability: A specially crafted DWARF section results in a read outside the bounds of in memory data so the calling application can crash.
product: libdwarf
description: Out of bound read bug in libdwarf git code. dwarf_dealloc() did not check the Dwarf_Ptr space argument before using it. This will lead to a out-of-bound read bug.
backtrace: #0 dwarf_dealloc (dbg=dbg@entry=0x655f30, space=0xa0, alloc_type=alloc_type@entry=1) at dwarf_alloc.c:477 #1 0x00002aaaaacf3296 in dealloc_srcfiles (dbg=0x655f30, srcfiles=0x66b8f0, srcfiles_count=17) at dwarf_macro5.c:1025 #2 0x00002aaaaacf50e6 in dealloc_srcfiles (srcfiles_count=<optimized out>, srcfiles=<optimized out>, dbg=<optimized out>) at dwarf_macro5.c:1021 ----- gef> p &r->rd_dbg $14 = (void **) 0x90
datefixed: 2016-05-04
references: regressiontests/liu/outofbound01.elf
https://bugzilla.redhat.com/show_bug.cgi?id=1332144
gitfixid: 98a3da1e8237fe0d45b67ef77f3fa5ed9ff0215f
tarrelease: libdwarf-20160507.tar.gz
id: DW201605-003
cve: CVE-2016-5042
fuzzer:
datereported: 2016-05-02
reportedby: Yue Liu
vulnerability: A specially crafted DWARF section results in an infinite loop that eventually crashes the application.
product: libdwarf
description: In dwarf_get_aranges_list() an invalid count will iterate, reading from memory addresses that increase till it all fails.
datefixed: 2016-05-04
references: regressiontests/liu/infiniteloop.elf
https://bugzilla.redhat.com/show_bug.cgi?id=1332145
gitfixid: 98a3da1e8237fe0d45b67ef77f3fa5ed9ff0215f
tarrelease: libdwarf-20160507.tar.gz
id: DW201605-004
cve: CVE-2016-5041
fuzzer:
datereported: 2016-05-02
reportedby: Yue Liu
vulnerability: A specially crafted DWARF section results in a null dereference reading debugging information entries which crashes the application.
product: libdwarf
description: If no DW_AT_name is present in a debugging information entry using DWARF5 macros a null dereference in dwarf_macro5.c will crash the application.
datefixed: 2016-05-04
references: regressiontests/liu/null01.elf
https://bugzilla.redhat.com/show_bug.cgi?id=1332148
gitfixid: 98a3da1e8237fe0d45b67ef77f3fa5ed9ff0215f
tarrelease: libdwarf-20160507.tar.gz
id: DW201605-005
cve: CVE-2016-5040
fuzzer:
datereported: 2016-05-02
reportedby: Yue Liu
vulnerability: A specially crafted DWARF section results in reading a compilation unit header that crashes the application.
product: libdwarf
description: If the data read for a compilation unit header contains a too large length value the library will read outside of its bounds and crash the application.
datefixed: 2016-05-04
references: regressiontests/liu/null02.elf
https://bugzilla.redhat.com/show_bug.cgi?id=1332149
gitfixid: 98a3da1e8237fe0d45b67ef77f3fa5ed9ff0215f
tarrelease: libdwarf-20160507.tar.gz
id: DW201605-006
cve:
fuzzer:
datereported: 2016-05-05
reportedby: Yue Liu
vulnerability: Two Heap-Overflow bug
product: libdwarf
description: Two test objects showing a heap overflow in libdwarf when using dwarfdump. It seems that these were fixed by the previous git update. Neither gdb nor valgrind find any errors when building with yesterday's commit.
datefixed: 2016-05-04
references: regressiontests/liu/free_invalid_address.elf regressiontests/liu/heapoverflow01b.elf
gitfixid: 98a3da1e8237fe0d45b67ef77f3fa5ed9ff0215f
tarrelease: libdwarf-20160507.tar.gz
id: DW201605-007
cve: CVE-2016-5039
fuzzer:
datereported: 2016-05-05
reportedby: Yue Liu
vulnerability: OOB read bug in get_attr_value()
product: libdwarf
description: Test object shows out of bound read. Object had data all-bits-on so the existing length check did not work due to wraparound. Added a check not susceptible to that error (DW_DLE_FORM_BLOCK_LENGTH_ERROR).
datefixed: 2016-05-06
references: regressiontests/liu/OOB0505_01.elf
gitfixid: eb1472afac95031d0c9dd8c11d527b865fe7deb8
tarrelease: libdwarf-20160507.tar.gz
id: DW201605-008
cve: CVE-2016-5038
fuzzer:
datereported: 2016-05-05
reportedby: Yue Liu
vulnerability: OOB read in dwarf_get_macro_startend_file()
product: libdwarf
description: Test object shows out of bound read. OOB at: line 772 *src_file_name = macro_context->mc_srcfiles[trueindex]; in dwarf_macro5.c A string offset into .debug_str is outside the bounds of the .debug_str section.
datefixed: 2016-05-12
references: regressiontests/liu/OOB0505_02.elf regressiontests/liu/OOB0505_02_02.elf
gitfixid: 82d8e007851805af0dcaaff41f49a2d48473334b
tarrelease: libdwarf-20160923.tar.gz
id: DW201605-009
cve: CVE-2016-5037
fuzzer:
datereported: 2016-05-05
reportedby: Yue Liu
vulnerability: NULL dereference in _dwarf_load_section
product: libdwarf
description: Test object shows null dereverence at line 1010 if(!strncmp("ZLIB",(const char *)src,4)) { in dwarf_init_finish.c The zlib code was not checking for a corrupted length-value.
datefixed: 2016-05-06
references: regressiontests/liu/NULLderefer0505_01.elf
gitfixid: b6ec2dfd850929821626ea63fb0a752076a3c08a
tarrelease: libdwarf-20160507.tar.gz
id: DW201605-010
cve: CVE-2016-5036
fuzzer:
datereported: 2016-05-06
reportedby: Yue Liu
vulnerability: OOB read bug in dump_block
product: libdwarf
description: Test object shows null dereverence at line 186 of dump_block() in print_sections.c Frame code was not noticing frame data corruption.
datefixed: 2016-05-12
references: regressiontests/liu/OOB_read3.elf regressiontests/liu/OOB_read3_02.elf
gitfixid: 82d8e007851805af0dcaaff41f49a2d48473334b
tarrelease: libdwarf-20160923.tar.gz
id: DW201605-011
cve: CVE-2016-5035
fuzzer:
datereported: 2016-05-06
reportedby: Yue Liu
vulnerability: OOB read bug in _dwarf_read_line_table_header
product: libdwarf
description: Test object shows null dereference at line 62 of dwarf_line_table_reader.c. Frame code and linetable code was not noticing data corruption.
datefixed: 2016-05-12
references: regressiontests/liu/OOB_read4.elf
gitfixid: 82d8e007851805af0dcaaff41f49a2d48473334b
tarrelease: libdwarf-20160923.tar.gz
id: DW201605-012
cve: CVE-2016-5034
fuzzer:
datereported: 2016-05-13
reportedby: Yue Liu
vulnerability: OOB write. From relocation records
product: libdwarf
description: Test object shows an invalid write in dwarf_elf_access.c (when doing the relocations). Adding the relocation value to anything overflowed and disguised the bad relocation record. With a 32bit kernel build the test could show a double-free and coredump due to the unchecked invalid writes from relocations.
datefixed: 2016-05-17
references: regressiontests/liu/HeapOverflow0513.elf
gitfixid: 10ca310f64368dc083efacac87732c02ef560a92
tarrelease: libdwarf-20160923.tar.gz
id: DW201605-013
cve: CVE-2016-5033
fuzzer:
datereported: 2016-05-17
reportedby: Yue Liu
vulnerability: OOB read bug in print_exprloc_content
product: libdwarf
description: Test object shows an invalid write in print_exprloc_content.
#0 print_exprloc_content (dbg=dbg@entry=0x654ea0,
die=die@entry=0x65b110, attrib=attrib@entry=0x65b590,
esbp=esbp@entry=0x7fffffffcef0, showhextoo=1) at print_die.c:4182
#1 0x0000000000412fb1 in get_attr_value (dbg=dbg@entry=0x654ea0,
tag=<optimized out>, die=die@entry=0x65b110,
dieprint_cu_goffset=dieprint_cu_goffset@entry=11,
attrib=attrib@entry=0x65b590, srcfiles=srcfiles@entry=0x0,
cnt=cnt@entry=0, esbp=esbp@entry=0x7fffffffcef0, show_form=0,
local_verbose=0) at print_die.c:4972
datefixed: 2015-05-18
references: regressiontests/liu/OOB0517_01.elf
gitfixid: ac6673e32f3443a5d36c2217cb814000930b2c54
tarrelease: libdwarf-20160923.tar.gz
id: DW201605-014
cve: CVE-2016-5032
fuzzer:
datereported: 2016-05-17
reportedby: Yue Liu
vulnerability: OOB read bug in dwarf_get_xu_hash_entry()
product: libdwarf
description: Test object shows an invalid read in dwarf_get _xu_hash_entry, lin 211.
#0 dwarf_get_xu_hash_entry (xuhdr=xuhdr@entry=0x657360,
index=index@entry=2897626028, hash_value=
hash_value@entry=0x7fffffffd5b0,
index_to_sections=index_to_sections@entry=0x7fffffffd5a8,
err=err@entry=0x7fffffffdb08) at dwarf_xu_index.c:211
#1 0x00002aaaaacfd05e in _dwarf_search_fission_for_key (
dbg=0x654a50, error=0x7fffffffdb08, percu_index_out=<synthetic pointer>,
key_in=0x7fffffffd670, xuhdr=0x657360) at dwarf_xu_index.c:363
#2 dwarf_get_debugfission_for_key (dbg=dbg@entry=0x654a50,
key=key@entry=0x7fffffffd670, key_type=key_type@entry=0x2aaaaad15e2a
"tu", percu_out=percu_out@entry=0x65a830,
error=error@entry=0x7fffffffdb08) at dwarf_xu_index.c:577
datefixed: 2015-05-18
references: regressiontests/liu/OOB0517_02.elf
gitfixid: ac6673e32f3443a5d36c2217cb814000930b2c54
tarrelease: libdwarf-20160923.tar.gz
id: DW201605-015
cve: CVE-2016-5031
fuzzer:
datereported: 2016-05-17
reportedby: Yue Liu
vulnerability: OOB read bug in print_frame_inst_bytes()
product: libdwarf
description: Test object shows an invalid read in print_frame_inst_bytes().
1294 for (; len > 0;) {
1295 unsigned char ibyte = *instp; <- $pc
1296 int top = ibyte & 0xc0;
#0 print_frame_inst_bytes (dbg=dbg@entry=0x654c80,
cie_init_inst=<optimized out>, len=503715, data_alignment_factor=-4,
code_alignment_factor=1, addr_size=addr_size@entry=4, offset_size=4,
version=3, config_data=config_data@entry=0x63bda0
<g_config_file_data>) at print_frames.c:1295
#1 0x000000000041b64c in print_one_cie (dbg=dbg@entry=0x654c80,
cie=<optimized out>, cie_index=cie_index@entry=1,
address_size=<optimized out>, config_data=
config_data@entry=0x63bda0 <g_config_file_data>) at print_frames.c:1161
#2 0x000000000041ce92 in print_frames (dbg=0x654c80,
print_debug_frame=print_debug_frame@entry=1, print_eh_frame=0,
config_data=config_data@entry=0x63bda0 <g_config_file_data>)
at print_frames.c:2209
gef> x/10x $r13
0x5e7981: Cannot access memory at address 0x5e7981
gef> p/x $r13
$14 = 0x5e7981
datefixed: 2015-05-18
references: regressiontests/liu/OOB0517_03.elf
gitfixid: ac6673e32f3443a5d36c2217cb814000930b2c54
tarrelease: libdwarf-20160923.tar.gz
id: DW201605-016
cve:
fuzzer:
datereported: 2016-05-19
reportedby: Yue Liu
vulnerability: Invalid dwarf leads to dwarfdump crash in print_frame_inst_bytes.
product: dwarfdump
description: Corrupted dwarf crashes dwarfdump
1297 }
1298 len = len_in;
1299 endpoint = instp + len;
1300 for (; len > 0;) {
1301 unsigned char ibyte = *instp; <- $pc
1302 int top = ibyte & 0xc0;
1303 int bottom = ibyte & 0x3f;
1304 int delta = 0;
1305 int reg = 0;
#0 print_frame_inst_bytes (dbg=dbg@entry=0x655ca0,
cie_init_inst=<optimized out>, len_in=<optimized out>,
data_alignment_factor=-4, code_alignment_factor=4,
addr_size=addr_size@entry=4, offset_size=4, version=3,
config_data=config_data@entry=0x63cda0 <g_config_file_data>)
at print_frames.c:1301
#1 0x000000000041b70c in print_one_cie
(dbg=dbg@entry=0x655ca0, cie=<optimized out>,
cie_index=cie_index@entry=2, address_size=<optimized out>,
config_data=config_data@entry=0x63cda0 <g_config_file_data>)
at print_frames.c:1161
#2 0x000000000041cf52 in print_frames (dbg=0x655ca0,
print_debug_frame=print_debug_frame@entry=1, print_eh_frame=0,
config_data=config_data@entry=0x63cda0 <g_config_file_data>)
at print_frames.c:2229
gef> p/x $r13
$1 = 0x4bcad8
gef> p/x *$r13
Cannot access memory at address 0x4bcad8
datefixed: 2016-05-22
references: regressiontests/liu/OOB_READ0519.elf
gitfixid: 6fa3f710ee6f21bba7966b963033a91d77c952bd
tarrelease: libdwarf-20160923.tar.gz
id: DW201605-017
cve: CVE-2016-5030
fuzzer:
datereported: 2016-05-19
reportedby: Yue Liu
vulnerability: Null dereference bug in _dwarf_calculate_info_section_end_ptr().
product: libdwarf
description: NULL dereference bug in _dwarf_calculate_info_section_end_ptr().
1742 Dwarf_Off off2 = 0; 1743 Dwarf_Small *dataptr = 0; 1744 1745 dbg = context->cc_dbg; 1746 dataptr = context->cc_is_info? dbg->de_debug_info.dss_data: <- $pc 1747 dbg->de_debug_types.dss_data; 1748 off2 = context->cc_debug_offset; 1749 info_start = dataptr + off2; 1750 info_end = info_start + context->cc_length + #0 _dwarf_calculate_info_section_end_ptr (context=context@entry=0x0) at dwarf_query.c:1746 #1 0x00002aaaaace307d in _dwarf_extract_string_offset_via_str_offsets (dbg=dbg@entry=0x655a70, info_data_ptr=0x6629f0 "", attrnum=attrnum@entry=121, attrform=attrform@entry=26, cu_context=0x0, str_sect_offset_out=str_sect_offset_out@entry=0x7fffffffd718, error=error@entry=0x7fffffffd878) at dwarf_form.c:1099 #2 0x00002aaaaacf4ed7 in dwarf_get_macro_defundef (macro_context=macro_context@entry=0x65b790, op_number=op_number@entry=1, line_number=line_number@entry=0x7fffffffd858, index=index@entry=0x7fffffffd860, offset=offset@entry=0x7fffffffd868, forms_count=forms_count@entry=0x7fffffffd7ce, macro_string=macro_string@entry=0x7fffffffd870, error=error@entry=0x7fffffffd878) at dwarf_macro5.c:557 ------ _dwarf_calculate_info_section_end_ptr (context=context@entry=0x0) at dwarf_query.c:1746 1746 dataptr = context->cc_is_info? dbg->de_debug_info.dss_data: gef> p/x $rdi $4 = 0x0
datefixed: 2016-05-22
references: regressiontests/liu/NULLdereference0519.elf
gitfixid: 6fa3f710ee6f21bba7966b963033a91d77c952bd
tarrelease: libdwarf-20160923.tar.gz
id: DW201605-018
cve: CVE-2016-5029
fuzzer:
datereported: 2016-05-22
reportedby: Yue Liu
vulnerability: Null dereference in create_fullest_file_path().
product: libdwarf
description: The null dereference in create_fullest_file_path() causes a crash. This is due to corrupted dwarf and the fix detects this corruption and if that null string pointer happens undetected a static string is substituted so readers can notice the situation.
202 }
203 if (dirno > 0 && fe->fi_dir_index > 0) {
204 inc_dir_name = (char *)
line_context->lc_include_directories[
205 fe->fi_dir_index - 1];
206 incdirnamelen = strlen(inc_dir_name); <- $pc
207 }
208 full_name = (char *) _dwarf_get_alloc(dbg,
#0 create_fullest_file_path (dbg=<optimized out>,
fe=0x68d510, line_context=0x68c4f0, name_ptr_out=<optimized
out>, error=0x7fffffffe2b8) at ./dwarf_line.c:206
#1 0x00007ffff7b6d3f9 in dwarf_filename (context=<optimized
out>, fileno_in=<optimized out>, ret_filename=0x7fffffffe280,
error=0x7fffffffe2b8) at ./dwarf_line.c:1418
#2 dwarf_linesrc (line=<optimized out>,
ret_linesrc=<optimized out>, error=<optimized out>) at
./dwarf_line.c:1436
datefixed: 2016-05-22
references: regressiontests/liu/NULLdereference0522.elf
gitfixid: acae971371daa23a19358bc62204007d258fbc5e
tarrelease: libdwarf-20160923.tar.gz
id: DW201605-019
cve: CVE-2016-5028
fuzzer:
datereported: 2016-05-23
reportedby: Yue Liu
vulnerability: Null dereference in print_frame_inst_bytes (dwarfdump)
product: libdwarf
description: The null dereference is due to a corrupted object file. Libdwarf was not dealing with empty (bss-like) sections since it really did not expect to see such in sections it reads! Now libdwarf catches the object error so dwarfdump sees the section as empty (as indeed it is!).
datefixed: 2016-05-23
references: regressiontests/liu/NULLdeference0522c.elf
gitfixid: a55b958926cc67f89a512ed30bb5a22b0adb10f4
tarrelease: libdwarf-20160923.tar.gz
id: DW201605-020
cve: CVE-2016-5027
fuzzer:
datereported: 2016-04-25
reportedby: Yue Liu,lieanu
vulnerability: NULL dereference in _dwarf_decode_s_leb128
product: libdwarf
description: dwarf_form.c in libdwarf 20160115 allows remote attackers to cause a denial of service (crash) via a crafted elf file Apparently no crafted object file presented. However the code fix is presented in the report at openwall.com. Discovered the CVE November 2021 To attack the code just pass the argument Dwarf_Word * leb128_length as a NULL pointer (that is allowed). The code was fixed in dwarf_leb.c on 2016-04-27 20:00:06.
bugzilla.redhat.com/show_bug.cgi?id=1330237 www.openwall.com/lists/oss-security/2016/05/24/1 www.openwall.com/lists/oss-security/2016/05/25/1
datefixed: 2016-05-27
references:
gitfixid:
tarrelease: libdwarf-20160507.tar.gz
id: DW201609-001
cve:
fuzzer:
datereported: 2016-09-16
reportedby: STARLAB
vulnerability: libdwarf 20160613 die_info_ptr in dwarf_die_deliv.c: 1533 Out-Of_bounds
product: libdwarf
description: At line 1533 of dwarf_die_deliv.c a pointer dereference is done with a pointer pointing past the end of the CU data.
see https://sourceforge.net/p/libdwarf/bugs/2/
==8054==ERROR: AddressSanitizer: heap-buffer-overflow on
address 0xf4c027ab at pc 0x819e4a4 bp 0xff88eb38 sp 0xff88eb30
READ of size 1 at 0xf4c027ab thread T0
#0 0x819e4a3 in dwarf_siblingof_b /home/starlab/fuzzing/dwarf-20160613/libdwarf/dwarf_die_deliv.c:1533
#1 0x8116201 in print_die_and_children_internal /home/starlab/fuzzing/dwarf-20160613/dwarfdump/print_die.c:1157
Bug report on sourceforge.net bug list for libdwarf.
The bad pointer dereference is due to libdwarf
not noticing that the DWARF in that file is corrupt.
In addition
The code was not noticing that it could dereference
a pointer that pointed out of bounds in the end-sibling-list
loop.
The example from the bug report (DW201609-001-poc) has the same problem. dwarfdump now reports DW_DLE_SIBLING_LIST_IMPROPER on both test2.o and DW201609-001-poc.
datefixed: 2016-09-17
references: regressiontests/DW201609-001/test2.o regressiontests/DW201609-001/DW201609-001-poc
gitfixid: 3767305debcba8bd7e1c483ae48c509d25399252
tarrelease: libdwarf-20160923.tar.gz
id: DW201609-002
cve: CVE-2016-7511
fuzzer:
datereported: 2016-09-18
reportedby: Shi Ji (@Puzzor)
vulnerability: libdwarf 20160613 Integer Overflow
product: libdwarf
description: In dwarf_get_size_of_val() with fuzzed DWARF data we get a SEGV.
See https://sourceforge.net/p/libdwarf/bugs/3/
==6825== ERROR: AddressSanitizer: SEGV on unknown address 0x0583903c (pc 0xb61f1a98 sp 0xbfa388b4 bp 0xbfa38d08 T0) AddressSanitizer can not provide additional info. #1 0xb61e3c0b (/usr/lib/i386-linux-gnu/libasan.so.0+0xdc0b) #2 0x80a21b1 in _dwarf_get_size_of_val /home/fuzzing/fuzzing/dwarf-20160613/libdwarf/dwarf_util.c:210 #3 0x8054214 in _dwarf_next_die_info_ptr /home/fuzzing/fuzzing/dwarf-20160613/libdwarf/dwarf_die_deliv.c:1340 #4 0x80557a5 in dwarf_child /home/fuzzing/fuzzing/dwarf-20160613/libdwarf/dwarf_die_deliv.c:1640 #5 0x804b23f in get_die_and_siblings /home/fuzzing/fuzzing/dwarf-20160613/dwarfexample/./simplereader.c:573
_dwarf_make_CU_Context() is insufficiently cautious about the length of a CU being absurd. Unclear as yet if this is the problem but it is a problem and is fixed for next release.
datefixed: 2016-09-23
references: regressiontests/DW201609-002/DW201609-002-poc
gitfixid: 3767305debcba8bd7e1c483ae48c509d25399252
tarrelease: libdwarf-20160923.tar.gz
id: DW201609-003
cve: CVE-2016-7410
fuzzer:
datereported: 2016-09-13
reportedby: https://marc.info/?l=oss-security&m=147391785920048&w=2
vulnerability: libdwarf 20160613 heap-buffer-overflow
product: libdwarf
description: With AddressSanitizer, we found a Heap-Buffer-overflow in the latest release version of dwarfdump. The crash output is as follows:
See also: https://marc.info/?l=oss-security&m=147378394815872&w=2 The testcase poc is from this web page.
==17411==ERROR: AddressSanitizer: heap-buffer-overflow on address
0xf3808904 at pc 0x80a6f76 bp 0xffb95e78 sp 0xffb95a5c
READ of size 4 at 0xf3808904 thread T0
==17411==WARNING: Trying to symbolize code, but external symbolizer is
not initialized!
#0 0x80a6f75 in __interceptor_memcpy ??:?
#1 0x8426c3b in _dwarf_read_loc_section
/home/starlab/fuzzing/dwarf-20160613/libdwarf/./dwarf_loc.c:919
#2 0x84250e2 in _dwarf_get_loclist_count
/home/starlab/fuzzing/dwarf-20160613/libdwarf/./dwarf_loc.c:970
#3 0x8438826 in dwarf_get_loclist_c
/home/starlab/fuzzing/dwarf-20160613/libdwarf/./dwarf_loc2.c:551
#4 0x81a1be8 in get_location_list
/home/starlab/fuzzing/dwarf-20160613/dwarfdump/print_die.c:3523
#5 0x816e1a2 in print_attribute
_dwarf_get_loclist_header_start() is not cautious about values in the header being absurdly large. Unclear as yet if this is the problem but it is a potential problem (fixed for next release).
Address Sanitizer in gcc reproduces the report. In _dwarf_read_loc_section() the simple calculation of loc_section_end was wrong, so end-of section was incorrect for the local reads. With that fixed we get DW_DLE_READ_LITTLEENDIAN_ERROR when libdwarf attempts to read off end of section.
datefixed: 2016-09-23
references: regressiontests/DW201609-003/poc
gitfixid: 3767305debcba8bd7e1c483ae48c509d25399252
tarrelease: libdwarf-20160923.tar.gz
id: DW201609-004
cve: CVE-2016-7510
fuzzer:
datereported: 2016-09-17
reportedby: Puzzor
vulnerability: libdwarf 20160613 Out-of-Bounds read
product: libdwarf
description: read line table program Out-of-Bounds read line_ptr in dwarf_line_table_reader_common.c:1433 Out-of-Bounds read See:
https://bugzilla.redhat.com/show_bug.cgi?id=1377015 https://sourceforge.net/p/libdwarf/bugs/4/
# Address Sanitizer Output ==27763==ERROR: AddressSanitizer: heap-buffer-overflow on address 0xf4603f84 at pc 0x8408ede bp 0xffff6518 sp 0xffff6510 READ of size 1 at 0xf4603f84 thread T0 #0 0x8408edd in read_line_table_program /home/puzzor/test-fuzzing/code/libdwarf/./dwarf_line_table_reader_common.c:1433 #1 0x83f716c in _dwarf_internal_srclines /home/puzzor/test-fuzzing/code/libdwarf/./dwarf_line.c:690 #2 0x841436c in dwarf_srclines_b /home/puzzor/test-fuzzing/code/libdwarf/./dwarf_line.c:944 #3 0x81fbc28 in print_line_numbers_this_cu /home/puzzor/test-fuzzing/code/dwarfdump/print_lines.c:763 #4 0x815c191 in print_one_die_section /home/puzzor/test-fuzzing/code/dwarfdump/print_die.c:850 #5 0x81565c1 in print_infos /home/puzzor/test-fuzzing/code/dwarfdump
datefixed: 2016-09-23
references: regressiontests/DW201609-004/poc
gitfixid: 3767305debcba8bd7e1c483ae48c509d25399252
tarrelease: libdwarf-20160923.tar.gz
id: DW201610-001
cve: CVE-2016-8681
fuzzer:
datereported: 2016-10-02
reportedby: agostino
vulnerability: Out of bounds read
product: libdwarf
description: The _dwarf_get_abbrev_for_code function in dwarf_util.c in libdwarf 20161001 and earlier allows remote attackers to cause a denial of service (out-of-bounds read) by calling the dwarfdump command on a crafted file.
bugzilla.redhat.com/show_bug.cgi?id=1385690 www.securityfocus.com/bid/93592 Duplicate of CVE-2016-8680
datefixed: 2016-10-04
references:
gitfixid: efe48cad0693d6994d9a7b561e1c3833b073a624
tarrelease:
id: DW201610-002
cve: CVE-2016-8680
fuzzer:
datereported: 2016-10-02
reportedby: agostino
vulnerability: Out of bounds read
product: libdwarf
description: The _dwarf_get_abbrev_for_code function in dwarf_util.c in libdwarf 20161001 and earlier allows remote attackers to cause a denial of service (out-of-bounds read) by calling the dwarfdump command on a crafted file.
bugzilla.redhat.com/show_bug.cgi?id=1385690 www.securityfocus.com/bid/93592 Duplicate of CVE-2016-8681
datefixed: 2016-10-04
references:
gitfixid: efe48cad0693d6994d9a7b561e1c3833b073a624
tarrelease:
id: DW201610-003
cve: CVE-2016-8679
fuzzer:
datereported: 2016-10-02
reportedby: agostino
vulnerability: dwarf_get_size_of_val out of bounds read
product: libdwarf
description: The _dwarf_get_size_of_val function in libdwarf/dwarf_util.c in Libdwarf before 20161124 allows remote attackers to cause a denial of service (out-of-bounds read) by calling the dwarfdump command on a crafted file.
www.securityfocus.com/bid/93601 blogs.gentoo.org/ago/2016/10/06/libdwarf-heap-based- buffer-overflow-in-_dwarf_get_size_of_val-dwarf_util-c/
datefixed: 2016-10-04
references:
gitfixid: efe48cad0693d6994d9a7b561e1c3833b073a624
tarrelease:
id: DW201611-001
cve:
fuzzer:
datereported: 2016-11-02
reportedby: Agostino Sarubbo
vulnerability: Memory allocation failure in do_decompress_zlib
product: libdwarf
description: In decompressing a zlib compressed section if the decompressed section size is nonsense (too large) an attempted malloc will fail and could let an exception propagate to callers.
==27994==WARNING: AddressSanitizer failed to allocate 0x62696c2f7273752f bytes ==27994==AddressSanitizer's allocator is terminating the process instead of returning 0 ... #6 0x4c0ab1 in malloc /var/tmp/portage/sys-devel/llvm-3.8.1- #8 0x5b582e in _dwarf_load_section #9 0x5bb479 in dwarf_srcfiles #10 0x5145cd in print_one_die_section
datefixed: 2016-11-04
references: regressiontests/sarubbo-2/00024-libdwarf-memalloc-do_decompress_zlib
gitfixid: 583f8834083b5ef834c497f5b47797e16101a9a6
tarrelease: libdwarf-20170416.tar.gz
id: DW201611-002
cve:
fuzzer:
datereported: 2016-11-02
reportedby: Agostino Sarubbo
vulnerability: heap overflow in get_attr_value
product: libdwarf
description: Libdwarf failed to check for a bogus length in dwarf_form.c (dwarf_formblock()) resulting in a pointer pointing outside of the intended memory region. Anything could happen in the subsequent use of the bogus pointer.
0x61300000de1c is located 0 bytes to the right of 348-byte region [0x61300000dcc0,0x61300000de1c) allocated by thread T0 here: #0 0x4c0ad8 in malloc /var/tmp/portage/sys-devel/llvm-3.8.1- r2/work/llvm-3.8.1.src/projects/compiler-rt/lib/asan/asan_malloc_linux.cc:52 #1 0x7f883cfc6206 in __libelf_set_rawdata_wrlock /tmp/portage/dev- libs/elfutils-0.166/work/elfutils-0.166/libelf/elf_getdata.c:318
datefixed: 2016-11-04
references: regressiontests/sarubbo-2/00025-libdwarf-heapoverflow-get_attr_value
gitfixid: 583f8834083b5ef834c497f5b47797e16101a9a6
tarrelease: libdwarf-20170416.tar.gz
id: DW201611-003
cve: CVE-2016-9276
fuzzer:
datereported: 2016-11-02
reportedby: Agostino Sarubbo
vulnerability: Bad aranges length leads to overflow and bad pointer
product: libdwarf
description: in dwarf_arange.c(dwarf_get_aranges_list) an aranges header with corrupt data could, with an overflowing calculation, result in pointers to invalid or inappropriate memory being dereferenced.
datefixed: 2016-11-04
references: regressiontests/sarubbo-2/00026-libdwarf-heapoverflow-dwarf_get_aranges_list
gitfixid: 583f8834083b5ef834c497f5b47797e16101a9a6
tarrelease: libdwarf-20180129.tar.gz
id: DW201611-004
cve: CVE-2016-9275
fuzzer:
datereported: 2016-11-02
reportedby: Agostino Sarubbo
vulnerability: Heap overflow in dwarf_skim_forms()
product: libdwarf
description: If a non-terminated string in a DWARF5 macro section ends a section it can result in accessing memory not in the application (out of bounds read). dwarf_macro5.c(in _dwarf_skim_forms()).
datefixed: 2016-11-04
references: regressiontests/sarubbo-2/00027-libdwarf-heapoverflow-_dwarf_skim_forms
gitfixid: 583f8834083b5ef834c497f5b47797e16101a9a6
tarrelease: libdwarf-20180129.tar.gz
id: DW201611-005
cve: CVE-2016-9558
fuzzer:
datereported: 2016-11-11
reportedby: Agostino Sarubbo
vulnerability: negation of -9223372036854775808 cannot be represented in type
product: libdwarf
description: With the right bit pattern in a signed leb number the signed leb decode would execute an unary minus with undefined effect. This is not known to generate an incorrect value, but it could, one supposes.
datefixed: 2016-11-11
references: regressiontests/sarubbo-2/00050-libdwarf-negate-itself
gitfixid: 4f19e1050cd8e9ddf2cb6caa061ff2fec4c9b5f9
tarrelease: libdwarf-20180129.tar.gz
id: DW201611-006
cve: CVE-2016-9480
fuzzer:
datereported: 2016-11-14
reportedby: Puzzor (Shi Ji)
vulnerability: Heap buffer overflow
product: libdwarf
description: An object with corrupt contents causes a memory reference out of bounds, a heap buffer overflow reference.
heap-buffer-overflow in dwarf_util.c:208 for val_ptr
# Version
bb9a3492ac5713bed9cf3ae58ddb7afa6e9e98f8
(in regression tests here named heap_buf_overflow.o)
# ASAN Output
<0> tag: 17 DW_TAG_compile_unit name: "strstrnocase.c" FORM 0xe "DW_FORM_strp"
<1> tag: 46 DW_TAG_subprogram name: "is_strstrnocase" FORM 0xe "DW_FORM_strp"
=================
==1666==ERROR: AddressSanitizer: heap-buffer-overflow on address
0xb5846db9 at p
c 0x080b3a1b bp 0xbfa75d18 sp 0xbfa75d08
READ of size 1 at 0xb5846db9 thread T0
#0 0x80b3a1a in _dwarf_get_size_of_val /home/puzzor/libdwarf-code/
libdwarf/dwarf_util.c:208
#1 0x8056602 in _dwarf_next_die_info_ptr /home/puzzor/libdwarf-code/
libdwarf/dwarf_die_deliv.c:1353
#2 0x8057f4b in dwarf_child /home/puzzor/libdwarf-code/libdwarf/
dwarf_die_de liv.c:1688
#3 0x804b5fa in get_die_and_siblings simplereader.c:637
#4 0x804b65c in get_die_and_siblings simplereader.c:643
#5 0x804b3f3 in read_cu_list simplereader.c:611
#6 0x804aeae in main simplereader.c:533
#7 0xb6ffe275 in __libc_start_main (/lib/i386-linux-gnu/libc.so.6+0x18275)
#8 0x80491c0 (/home/puzzor/libdwarf-code/dwarfexample/simplereader+
0x80491c 0)
0xb5846db9 is located 0 bytes to the right of 249-byte region
[0xb5846cc0,0xb5846db9)
allocated by thread T0 here:
#0 0xb727fae4 in __interceptor_malloc (/usr/lib/i386-linux-gnu/libasan.so.
3+ 0xc3ae4)
#1 0xb71a9b98 (/usr/lib/i386-linux-gnu/libelf.so.1+0x9b98)
For the orignal bug report see
https://sourceforge.net/p/libdwarf/bugs/5/
datefixed: 2016-11-16
references: regressiontests/puzzor/heap_buf_overflow.o
gitfixid: 5dd64de047cd5ec479fb11fe7ff2692fd819e5e5
tarrelease: libdwarf-20180129.tar.gz
id: DW201611-007
cve: CVE-2016-10255
fuzzer:
datereported: 2016-11-04
reportedby: Agostino Sarubbo
vulnerability: Crash libelf reading fuzzed object.
product: libdwarf
description: This is a weakness in libelf checking. Testing that current libdwarf deals with it properly, though it was never a bug in libdwarf. The CVE mentions libdwarf.
bugzilla.redhat.com/show_bug.cgi?id=1387584 www.openwall.com/lists/oss-security/2017/03/22/1 blogs.gentoo.org/ago/2016/11/04/elfutils-memory-allocation-failure-in-__libelf_set_rawdata_wrlock-elf_getdata-c/
Fixed in gentoo libelf by Agostino Sarubbo.
datefixed: 2016-11-04
references: regressiontests/sarubbo-a/00031-elfutils-memalloc-__libelf_set_rawdata_wrlock
gitfixid:
tarrelease: libdwarf-20180129.tar.gz
id: DW201611-008
cve: CVE-2016-10254
fuzzer:
datereported: 2016-11-04
reportedby: Agostino Sarubbo
vulnerability: Crash libelf reading fuzzed object.
product: libdwarf
description: This is a weakness in libelf checking. Testing that current libdwarf deals with it properly, though it was never a bug in libdwarf. The CVE mentions libdwarf.
blogs.gentoo.org/ago/2016/11/04/elfutils-memory-allocation-failure-in-allocate_elf-common-h/ www.openwall.com/lists/oss-security/2017/03/22/2
Fixed in gentoo libelf by Agostino Sarubbo.
datefixed: 2016-11-04
references: regressiontests/sarubbo-b/00011-elfutils-memalloc-allocate_elf
gitfixid:
tarrelease: libdwarf-20180129.tar.gz
id: DW201703-001
cve: CVE-2017-9055
fuzzer:
datereported: 2017-03-21
reportedby: Marcel Bohme and Van-Thuan Pham
vulnerability: Heap overflow in dwarf_formsdata
product: libdwarf
description: 1/7. In dwarf_formsdata() a few data types were not checked as being in bounds. The test object is intentionally corrupted (fuzzed).
A portion of sanitizer output:
LOCAL_SYMBOLS:
< 1><0x0000002f> DW_TAG_subprogram
==180088==ERROR: AddressSanitizer: heap-buffer-overflow on
address 0x60800000bf72 at pc 0x0000006cab95 bp
0x7fff31425830 sp 0x7fff31425828
READ of size 1 at 0x60800000bf72 thread T0
#0 0x6cab94 in dwarf_formsdata /home/ubuntu/subjects/
build-asan/libdwarf/libdwarf/dwarf_form.c:937:9
#1 0x567daf in get_small_encoding_integer_and_name /home/
ubuntu/subjects/build-asan/libdwarf/dwarfdump/print_die.c:1533:16
#2 0x576f38 in check_for_type_unsigned /home/ubuntu/
subjects/build-asan/libdwarf/dwarfdump/print_die.c:4301:11
#3 0x56ad8c in formxdata_print_value /home/ubuntu/
subjects/build-asan/libdwarf/dwarfdump/print_die.c:4374:39
#4 0x5643be in get_attr_value /home/ubuntu/
subjects/build-asan/libdwarf/dwarfdump/print_die.c:5140:24
#5 0x555f86 in print_attribute /home/ubuntu/subjects/build
...
After fixes applied dwarfdump gets:
ERROR: dwarf_attrlist: DW_DLE_DW_DLE_ATTR_OUTSIDE_SECTION(281)
datefixed: 2017-03-21
references: regressiontests/marcel/crash1
gitfixid: cc37d6917011733d776ae228af4e5d6abe9613c1
tarrelease: libdwarf-20180129.tar.gz
id: DW201703-002
cve: CVE-2017-9054
fuzzer:
datereported: 2017-03-21
reportedby: Marcel Bohme and Van-Thuan Pham
vulnerability: Heap overflow in _dwarf_decode_s_leb128_chk()
product: libdwarf
description: 2/7. In _dwarf_decode_s_leb128_chk() a byte pointer was dereferenced just before was checked as being in bounds. The test object is intentionally corrupted (fuzzed).
A portion of sanitizer output:
.debug_line: line number info for a single cu
==180103==ERROR: AddressSanitizer: heap-buffer-overflow
on address 0x610000007ffc at pc 0x0000007b0f5b
bp 0x7ffe06bbf510 sp 0x7ffe06bbf508
READ of size 1 at 0x610000007ffc thread T0
#0 0x7b0f5a in _dwarf_decode_s_leb128_chk /home/ubuntu/
subjects/build-asan/libdwarf/libdwarf/dwarf_leb.c:304:9
#1 0x7e753e in read_line_table_program /home/ubuntu/
subjects/build-asan/libdwarf/libdwarf/./
dwarf_line_table_reader_common.c:1167:17
#2 0x7d7fe3 in _dwarf_internal_srclines /home/ubuntu/
subjects/build-asan/libdwarf/libdwarf/./dwarf_line.c:690:15
#3 0x7f9dbb in dwarf_srclines_b /home/ubuntu/
subjects/build-asan/libdwarf/libdwarf/./dwarf_line.c:944:12
#4 0x5caaa5 in print_line_numbers_this_cu /home/ubuntu/
subjects/build-asan/libdwarf/dwarfdump/print_lines.c:762:16
After fix applied one gets:
ERROR: dwarf_srclines: DW_DLE_LEB_IMPROPER (329)
Runs off end of section or CU
datefixed: 2017-03-21
references: regressiontests/marcel/crash2
gitfixid: cc37d6917011733d776ae228af4e5d6abe9613c1
tarrelease: libdwarf-20180129.tar.gz
id: DW201703-003
cve:
fuzzer:
datereported: 2017-03-21
reportedby: Marcel Bohme and Van-Thuan Pham
vulnerability: Heap overflow in strcmp
product: libdwarf (libelf)
description: 3/7. An apparent heap overflow that gives the appearance of being in libdwarf is due to libelf call elf_strptr() failing to fully check that its arguments make sense. This is not a bug in libdwarf, it is a libelf bug. The test object is intentionally corrupted (fuzzed). The submission was with Ubuntu 14.04. With Ubuntu 16.04 there is no sanitizer error report. A portion of sanitizer output with Ubuntu 14.04:
==180106==ERROR: AddressSanitizer: heap-buffer-overflow
on address 0x60f00000ef09 at pc 0x000000447300
bp 0x7ffc667dce10 sp 0x7ffc667dc5d0
READ of size 4 at 0x60f00000ef09 thread T0
#0 0x4472ff in __interceptor_strcmp (/home/ubuntu/
subjects/build-asan/libdwarf/dwarfdump/dwarfdump+0x4472ff)
#1 0x79938f in this_section_dwarf_relevant /home/ubuntu/
subjects/build-asan/libdwarf/libdwarf/dwarf_init_finish.c:612:12
#2 0x781064 in _dwarf_setup /home/ubuntu/
subjects/build-asan/libdwarf/libdwarf/dwarf_init_finish.c:722:14
#3 0x77d59c in dwarf_object_init /home/ubuntu/
subjects/build-asan/libdwarf/libdwarf/dwarf_init_finish.c:922:20
#4 0x899d4f in dwarf_elf_init_file_ownership /
With Ubuntu 16.04 libelf one gets: ERROR: dwarf_elf_init: DW_DLE_ELF_STRPTR_ERROR (30) a call to elf_strptr() failed trying to get a section name Fix date is irrelevant, libdwarf no longer uses libelf.
datefixed:
references: regressiontests/marcel/crash3
gitfixid:
tarrelease: libdwarf-20180129.tar.gz
id: DW201703-004
cve:
fuzzer:
datereported: 2017-03-21
reportedby: Marcel Bohme and Van-Thuan Pham
vulnerability: Heap overflow in set_up_section strlen
product: libdwarf (libelf)
description: 4/7. An apparent heap overflow that gives the appearance of being in libdwarf is due to libelf call elf_strptr() failing to fully check that its arguments make sense. This is not a bug in libdwarf, it is a libelf bug. The test object is intentionally corrupted (fuzzed). The submission was with Ubuntu 14.04. With Ubuntu 16.04 there is no sanitizer error report. As of 2023 libdwarf no longer calls or references libelf.
A portion of sanitizer output with Ubuntu 14.04:
==180109==ERROR: AddressSanitizer: heap-buffer-overflow
on address 0x60b00000b000 at pc 0x00000048fd12
bp 0x7fff4ad31ef0 sp 0x7fff4ad316b0
READ of size 16 at 0x60b00000b000 thread T0
#0 0x48fd11 in __interceptor_strlen (/home/ubuntu/
subjects/build-asan/libdwarf/dwarfdump/dwarfdump+0x48fd11)
#1 0x7a84a4 in set_up_section /home/ubuntu/
subjects/build-asan/libdwarf/libdwarf/dwarf_init_finish.c:285:27
#2 0x79aaa5 in enter_section_in_de_debug_sections_array /home/ubuntu/
subjects/build-asan/libdwarf/libdwarf/dwarf_init_finish.c:355:5
#3 0x78170b in _dwarf_setup /home/ubuntu/
subjects/build-asan/libdwarf/libdwarf/dwarf_init_finish.c:746:19
With Ubuntu 16.04 libelf one gets:
ERROR: dwarf_elf_init: DW_DLE_ELF_STRPTR_ERROR (30)
a call to elf_strptr() failed trying to get a section name
datefixed:
references: regressiontests/marcel/crash4
gitfixid:
tarrelease: libdwarf-20180129.tar.gz
id: DW201703-005
cve: CVE-2017-9053
fuzzer:
datereported: 2017-03-21
reportedby: Marcel Bohme and Van-Thuan Pham
vulnerability: Heap overflow in _dwarf_read_loc_expr_op()
product: libdwarf
description: 5/7. A heap overflow in _dwarf_read_loc_expr_op() is due to a failure to check a pointer for being in bounds (in a few places in this function). The test object is intentionally corrupted (fuzzed).
A portion of sanitizer output with Ubuntu 14.04:
==180112==ERROR: AddressSanitizer: heap-buffer-overflow
on address 0x60800000bf72 at pc 0x00000084dd52
bp 0x7ffc12136fd0 sp 0x7ffc12136fc8
READ of size 1 at 0x60800000bf72 thread T0
#0 0x84dd51 in _dwarf_read_loc_expr_op /home/ubuntu/subjects/
build-asan/libdwarf/libdwarf/./dwarf_loc.c:250:9
#1 0x841f16 in _dwarf_get_locdesc_c /home/ubuntu/subjects/
build-asan/libdwarf/libdwarf/./dwarf_loc2.c:109:15
#2 0x837d08 in dwarf_get_loclist_c /home/ubuntu/subjects/
build-asan/libdwarf/libdwarf/./dwarf_loc2.c:685:18
#3 0x57dff2 in get_location_list /home/ubuntu/subjects/
build-asan/libdwarf/dwarfdump/print_die.c:3812:16
After fixes applied dwarfdump says:
ERROR: dwarf_get_loclist_c: DW_DLE_LOCEXPR_OFF_SECTION_END
(343) Corrupt dwarf
datefixed: 2017-03-21
references: regressiontests/marcel/crash5
gitfixid: cc37d6917011733d776ae228af4e5d6abe9613c1
tarrelease: libdwarf-20180129.tar.gz
id: DW201703-006
cve: CVE-2017-9052
fuzzer:
datereported: 2017-03-21
reportedby: Marcel Bohme and Van-Thuan Pham
vulnerability: Heap overflow in dwarf_formsdata
product: libdwarf
description: 6/7. A heap overflow in dwarf_formsdata() is due to a failure to check a pointer for being in bounds (in a few places in this function) and a failure in a check in dwarf_attr_list(). The test object is intentionally corrupted (fuzzed).
A portion of sanitizer output with Ubuntu 14.04:
==180130==ERROR: AddressSanitizer: heap-buffer-overflow
on address 0x61100000589c at pc 0x0000006cab95
bp 0x7fff749aab10 sp 0x7fff749aab08
READ of size 1 at 0x61100000589c thread T0
#0 0x6cab94 in dwarf_formsdata /home/ubuntu/subjects/
build-asan/libdwarf/libdwarf/dwarf_form.c:937:9
#1 0x567daf in get_small_encoding_integer_and_name /home/ubuntu/subjects/
build-asan/libdwarf/dwarfdump/print_die.c:1533:16
#2 0x562f28 in get_attr_value /home/ubuntu/subjects/
build-asan/libdwarf/dwarfdump/print_die.c:5030:24
#3 0x555f86 in print_attribute /home/ubuntu/subjects/
build-asan/libdwarf/dwarfdump/print_die.c:3357:13
After fixes applied dwarfdump says:
ERROR: dwarf_attrlist: DW_DLE_DW_DLE_ATTR_OUTSIDE_SECTION(281)
datefixed: 2017-03-21
references: regressiontests/marcel/crash6
gitfixid: cc37d6917011733d776ae228af4e5d6abe9613c1
tarrelease: libdwarf-20180129.tar.gz
id: DW201703-007
cve:
fuzzer:
datereported: 2017-03-21
reportedby: Marcel Bohme and Van-Thuan Pham
vulnerability: Heap overflow in strncmp (libelf bug)
product: libdwarf (libelf)
description: 7/7. A heap overflow in strncmp() is due to libelf failing to check arguments to elf_ strptr. This is not a bug in libdwarf, it is a libelf bug. A pointer for being in bounds (in a few places in this function) and a failure in a check in dwarf_attr_list(). The test object is intentionally corrupted (fuzzed).
A portion of sanitizer output with Ubuntu 14.04:
==180133==ERROR: AddressSanitizer: heap-buffer-overflow
on address 0x60d00000cff1 at pc 0x0000004476f4
bp 0x7fff87dd7dd0 sp 0x7fff87dd7590
READ of size 8 at 0x60d00000cff1 thread T0
#0 0x4476f3 in __interceptor_strncmp (/home/ubuntu/subjects/
build-asan/libdwarf/dwarfdump/dwarfdump+0x4476f3)
#1 0x7992ae in this_section_dwarf_relevant /home/ubuntu/subjects/
build-asan/libdwarf/libdwarf/dwarf_init_finish.c:608:13
#2 0x781064 in _dwarf_setup /home/ubuntu/subjects/
build-asan/libdwarf/libdwarf/dwarf_init_finish.c:722:14
#3 0x77d59c in dwarf_object_init /home/ubuntu/subjects/
build-asan/libdwarf/libdwarf/dwarf_init_finish.c:922:20
With Ubuntu 16.04 libelf dwarfdump gets:
ERROR: dwarf_elf_init: DW_DLE_ELF_STRPTR_ERROR (30)
a call to elf_strptr() failed trying to get a section name
Fix date is irrelevant, libdwarf no longer uses libelf.
datefixed: 2017-07-06
references: regressiontests/marcel/crash7
gitfixid: cc37d6917011733d776ae228af4e5d6abe9613c1
tarrelease: libdwarf-20180129.tar.gz
id: DW201706-001
cve: CVE-2017-9998
fuzzer:
datereported: 2017-06-28
reportedby: team OWL337
vulnerability: Addition overflow in libdwarf leads to segmentation violation
product: libdwarf
description: A fuzzed object with a resulting invalid value can overflow when added to a valid pointer (depending on how the runtime memory is laid out) and thereafter a dereference results in a segmentation violation).
see https://bugzilla.redhat.com/show_bug.cgi?id=1465756 for contact information of those finding the bug. Fabian Wolff sent email and provided the link to the web page.
datefixed: 2017-07-06
references: regressiontests/wolff/POC1
gitfixid: e91681e8841291f57386f26a90897fd1dcf92a6e
tarrelease: libdwarf-20180129.tar.gz
id: DW201709-001
cve:
fuzzer:
datereported: 2017-09-19
reportedby: Agostino Sarubbo
vulnerability: Incorrect abbrev section could crash caller.
product: libdwarf
description: A fuzzed object with a resulting invalid abbrev section where the end of section follows an abbrev tag would dereference a non-existent has-child byte.
datefixed: 2017-09-26
references: regressiontests/sarubbo-3/1.crashes.bin
gitfixid: bcc2e33908e669bacd397e3c941ffd1db3005d17
tarrelease: libdwarf-20180129.tar.gz
id: DW201711-001
cve:
fuzzer:
datereported: 2017-11-01
reportedby: Agostino Sarubbo
vulnerability: Incorrect frame section could crash caller
product: libdwarf
description: A carefully crafted object with a resulting invalid frame section with DW_CFA_advance_loc1 implying data off-the-end-of-section will dereference an invalid pointer. A segmentation fault and core dump is possible. Corrected code checks now.
datefixed: 2017-11-02
references: regressiontests/sarubbo-8/1.crashes.bin
gitfixid: 44349d7991e44dd3751794f76537cabcf65ee28d
tarrelease: libdwarf-20180129.tar.gz
id: DW201711-002
cve:
fuzzer:
datereported: 2017-11-08
reportedby: Agostino Sarubbo
vulnerability: Incorrect line table section could crash caller
product: libdwarf
description: An carefully crafted object with a invalid line table section crafted to end early at a particular point resulted in dereferencing outside the line table from libdwarf/dwarf_line_table_reader_common.c . A segmentation-fault/core-dump is possible.
datefixed: 2017-11-08
references: regressiontests/sarubbo-9/3.crashes.bin
gitfixid: a1644f4dde7dd5990537ff7ad22a9e94b8723186
tarrelease: libdwarf-20180129.tar.gz
id: DW201712-001
cve:
fuzzer:
datereported: 2017-12-01
reportedby: Agostino Sarubbo
vulnerability: Incorrect frame section could let caller crash
product: libdwarf
description: A carefully crafted object with an invalid frame section can result in passing back data to a caller of dwarf_get_fde_augmentation_data() is erroneous and will result in the caller reference off the end of the frame section. A segmentation-fault/core-dump is possible.
datefixed: 2017-12-01
references: sarubbo-10/1.crashes.bin
gitfixid: 329ea8e56bc9550260cae6e2e9756bfbe7e2ff6d
tarrelease: libdwarf-20180129.tar.gz
id: DW201801-001
cve:
fuzzer:
datereported: 2018-01-28
reportedby: Agostino Sarubbo
vulnerability: Incorrect frame section can crash dwarfdump
product: dwarfdump
description: A carefully crafted object with an invalid frame section set of initial-instructions can crash the frame-instructions decode in dwarfdump. In addition, a couple places in libdwarf are not as careful in checking frame data as they should be. A segmentation-fault/core-dump is possible.
datefixed: 2018-01-29
references: sarubbo-11/testcase{1,2,3,4,5}.bin
gitfixid: 7af0ecddfafed88446969cbf8c888356ad485d99
tarrelease: libdwarf-20180129.tar.gz
id: DW201907-001
cve: CVE-2019-14249
fuzzer:
datereported: 2019-07-23
reportedby: unknown
vulnerability: Denial of service with zero size section group
product: libdwarf
description: dwarf_elf_load_headers.c in libdwarf before 2019-07-05 allows attackers to cause a denial of service (division by zero) via an ELF file with a zero-size section group (SHT_GROUP), as demonstrated by dwarfdump.
datefixed: 2019-07-05
references:
gitfixid: cb7198abde46c2ae29957ad460da6886eaa606ba
tarrelease: libdwarf-0.4.1.tar.xz
id: DW202010-001
cve: CVE-2020-27545
fuzzer:
datereported: 2020-10-10
reportedby: Casper Sun
vulnerability: A carefully corrupted line table can crash calling app
product: libdwarf
description: A carefully crafted object with an invalid line table could cause libdwarf to dereference a pointer reading a single byte outside of the intended .debug_line section and potentially outside of memory visible to the library. A segmentation fault is possible. The code testing for the error was coded incorrectly so an invalid dereference could occur. Now the test code is correct and the error is detected resulting in a normal error return.
This should be visible after redhat makes it public. Filed on bugzilla.redhat 22 November 2021. bugzilla.redhat.com/show_bug.cgi?id=2025694
datefixed: 2020-10-17
references: regressiontests/c-sun/poc
gitfixid: 95f634808c01f1c61bbec56ed2395af997f397ea
tarrelease: libdwarf-0.4.1.tar.xz
id: DW202010-002
cve: CVE-2020-28162
fuzzer:
datereported: 2020-10-27
reportedby: Casper Sun
vulnerability: dwarfdump crashes if the nest of C scopes is too deep
product: dwarfdump
description: An object file where the DIEs depth of nesting exceeds the limit of 800 levels due to corruption or a compiler bug can result in exhausting the die stack array and writing past its end. A segmentation fault is possible. The code at the point of error was not adjusting the array index properly so an invalid dereference could occur. Now the test code is correct and the array overflow is detected resulting in a normal error return. Additional places where this could occur were identified and the proper test added.
Unable to enter in bugzilla.redhat.com so CVE can be completed by Fedora (as CNA) as dwarfdump is not part of Fedora
datefixed: 2020-10-28
references: regressiontests/c-sun2/globaloverflow
gitfixid: a7fa8edd640b74daf8e7a442dcec96640875b4fb
tarrelease: libdwarf-0.4.1.tar.xz
id: DW202010-003
cve: CVE-2020-28163
fuzzer:
datereported: 2020-10-27
reportedby: Casper Sun
vulnerability: Passing null to %s due to corrupt line table header.
product: libdwarf
description: If a DWARF5 line table header has an invalid FORM for a pathname, the fi_file_name field may be null and printing it via %s can result in referencing memory at address 0, possibly generating segmentation violation or application crash. Now in case of null we provide a fixed string of <no file name> and for the form code we print the value and <unknown form> so there are no unpredictable effects.
This should be visible after redhat makes it public. Filed on bugzilla.redhat 23 November 2021. bugzilla.redhat.com/show_bug.cgi?id=2026000
datefixed: 2020-10-28
references: regressiontests/c-sun2/nullpointer
gitfixid: faf99408e3f9f706fc3809dd400e831f989778d3
tarrelease: libdwarf-0.4.1.tar.xz
id: DW202111-001
cve:
fuzzer: oss-fuzz-40663
datereported: 2021-11-03
reportedby: David Korczynski
vulnerability: Timeout in fuzz_init_path
product: libdwarf
description: Corrupted MachO object can crash caller Two fields in the MachO file header were not checked for sanity so nonsense large values could lead to excessive malloc and or a caller segmentation violation. Verified by oss-fuzz as fixed. The testcase has illegal libdwarf call and improper include statements.
https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=40663
datefixed: 2021-11-04
references: regressiontests/ossfuzz40663/clusterfuzz-testcase-minimized-fuzz_init_path-6122542432124928
gitfixid: b40f7e291216e771185f62292dd6304b5a662926
tarrelease: libdwarf-0.4.1.tar.xz
id: DW202111-002
cve:
fuzzer: oss-fuzz-40669
datereported: 2021-11-03
reportedby: David Korczynski
vulnerability: Out-of-memory in fuzz_init_path
product: libdwarf
description: Corrupted MachO object can crash caller.b Two fields in the MachO file header were not checked for sanity so nonsense large values could lead to excessive malloc and or a caller segmentation violation. Fixed by DW202111-001. Verified as fixed by oss-fuzz
https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=40669
datefixed: 2021-11-04
references: regressiontests/ossfuzz40669/clusterfuzz-testcase-minimized-fuzz_init_path-5399726397194240 regressiontests/clusterfuzz-testcase-fuzz_init_path-5399726397194240
gitfixid: b40f7e291216e771185f62292dd6304b5a662926
tarrelease: libdwarf-0.4.1.tar.xz
id: DW202111-003
cve:
fuzzer: oss-fuzz-40671
datereported: 2021-11-03
reportedby: David Korczynski
vulnerability: Direct-leak in _dwarf_get_debug
product: libdwarf
description: The test code is calling a libdwarf-internal function (which is against the rules, only libdwarf function names beginning with dwarf_ are callable. When building libdwarf as an archive there is no means to enforce this rule) doc/libdwarf.mm/pdf now documents this rule.
https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=40671
datefixed: 2021-11-05
references: regressiontests/oss40671/clusterfuzz-testcase-fuzz_init_path-5455557297831936 regressiontests/oss40671/clusterfuzz-testcase-minimized-fuzz_init_path-5455557297831936
gitfixid: b40f7e291216e771185f62292dd6304b5a662926
tarrelease: libdwarf-0.4.1.tar.xz
id: DW202111-004
cve:
fuzzer: oss-fuzz-40673
datereported: 2021-11-03
reportedby: David Korczynski
vulnerability: Null-dereference READ in dwarf_object_init_b
product: libdwarf
description: The macho object has corrupted headers and now mentions that and stops. Verified as fixed by oss-fuzz 2021-11-03
https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=40673
datefixed: 2021-11-05
references: regressiontests/ossfuzz40673/clusterfuzz-testcase-minimized-fuzz_init_path-6240961391362048.fuzz
gitfixid: 94dece3ce0f030d06da442a103bd6a5301410b25
tarrelease: libdwarf-0.4.1.tar.xz
id: DW202111-005
cve:
fuzzer: oss-fuzz-40674
datereported: 2021-11-03
reportedby: David Korczynski
vulnerability: Heap-buffer-overflow in _dwarf_elf_setup_all_section_groups
product: libdwarf
description: Object file has corrupt section group information. Results in buffer overflow.
https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=40674#c6
datefixed: 2021-11-07
references: regressiontests/ossfuzz40674/clusterfuzz-testcase-minimized-fuzz_init_path-6557751518560256
gitfixid: 94dece3ce0f030d06da442a103bd6a5301410b25
tarrelease: libdwarf-0.4.1.tar.xz
id: DW202111-008
cve:
fuzzer: oss-fuzz-40731
datereported: 2021-11-03
reportedby: David Korczynski
vulnerability: Out-of-memory in fuzz_init_binary
product: libdwarf
description: The fuzzed macho64 object has corrupted headers. The library notices and reports an error.
https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=40731
datefixed: 2021-11-07
references: regressiontests/ossfuzz40731/clusterfuzz-testcase-fuzz_init_binary-5983147574034432
gitfixid: 94dece3ce0f030d06da442a103bd6a5301410b25
tarrelease: libdwarf-0.4.1.tar.xz
id: DW202111-009
cve:
fuzzer: oss-fuzz-40729
datereported: 2021-11-05
reportedby: David Korczynski
vulnerability: Timeout - fuzz_init_binary
product: libdwarf
description: The object file (macho 64 bit) has some header fuzzing that was not caught reading the object until the macho reader tried a gigantic malloc.. Now the library code catches the error before malloc and returns an error code.
https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=40729
datefixed: 2021-11-07
references: regressiontests/ossfuzz40729/clusterfuzz-testcase-minimized-fuzz_init_binary-4791627277795328
gitfixid: 94dece3ce0f030d06da442a103bd6a5301410b25
tarrelease: libdwarf-0.4.1.tar.xz
id: DW202111-010
cve:
fuzzer: oss-fuzz-40627
datereported: 2021-11-02
reportedby: David Korczynski
vulnerability: Abrt in _dwarf_error_string
product: libdwarf
description: The Elf object file has some corruption. The read now stops with an error.
https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=40627
datefixed: 2021-11-07
references: regressiontests/ossfuzz40627/clusterfuzz-testcase-fuzz_init_path-5186858573758464 regressiontests/ossfuzz40627/clusterfuzz-testcase-minimized-fuzz_init_path-5186858573758464
gitfixid: 94dece3ce0f030d06da442a103bd6a5301410b25
tarrelease: libdwarf-0.4.1.tar.xz
id: DW202111-011
cve:
fuzzer: oss-fuzz-40799
datereported: 2021-11-02
reportedby: David Korczynski
vulnerability: Out-of-memory in fuzz_init_path
product: libdwarf
description: A corrupted object. Gigantic section sizes or offsets were provoking a large malloc. Now these are detected and no malloc is attempted (an error is returned).
https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=40799
datefixed: 2021-11-07
references: regressiontests/ossfuzz40799/clusterfuzz-testcase-fuzz_init_path-5245778948390912 regressiontests/ossfuzz40799/clusterfuzz-testcase-minimized-fuzz_init_path-5245778948390912
gitfixid: 94dece3ce0f030d06da442a103bd6a5301410b25
tarrelease: libdwarf-0.4.1.tar.xz
id: DW202111-012
cve:
fuzzer: oss-fuzz-40801
datereported: 2021-11-07
reportedby: David Korczynski
vulnerability: Timeout in fuzz_init_path
product: libdwarf
description: A corrupted object. libdwarf detects it quickly now.
https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=40801
datefixed: 2021-11-07
references: regressiontests/ossfuzz801/clusterfuzz-testcase-fuzz_init_path-5443517279764480 regressiontests/ossfuzz40801/clusterfuzz-testcase-minimized-fuzz_init_path-5443517279764480
gitfixid: 94dece3ce0f030d06da442a103bd6a5301410b25
tarrelease: libdwarf-0.4.1.tar.xz
id: DW202111-013
cve:
fuzzer: oss-fuzz-40802
datereported: 2021-11-07
reportedby: David Korczynski
vulnerability: Null-dereference READ in dwarf_object_init_b
product: libdwarf
description: A corrupted object. The error handling code in dwarf_object_init_b was not properly dealing with a NULL pointer Dwarf_Error *errp in the test code.
https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=40802
datefixed: 2021-11-19
references: regressiontests/ossfuzz40802/ clusterfuzz-testcase-fuzz_init_binary-5538015955517440.fuzz regressiontests/ossfuzz40802/clusterfuzz-testcase-minimized-fuzz_init_binary-5538015955517440.fuzz
gitfixid: adf4dae25b39039f1821b095688c00f3010e1d37
tarrelease: libdwarf-0.4.1.tar.xz
id: DW202111-014
cve:
fuzzer: oss-fuzz-40895
datereported: 2021-11-10
reportedby: David Korczynski
vulnerability: Out-of-memory in fuzz_init_binary
product: libdwarf
description: A corrupted object. Some Elf section sizes are larger than the file size.
https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=40895
datefixed: 2021-11-12
references: regressiontests/ossfuzz40895/clusterfuzz-testcase-fuzz_init_binary-4805508242997248 regressiontests/ossfuzz40895/clusterfuzz-testcase-minimized-fuzz_init_binary-4805508242997248
gitfixid: b7a119dc07c502c1334bcbf8dd04ca0e4d5f6ab6
tarrelease: libdwarf-0.4.1.tar.xz
id: DW202111-015
cve:
fuzzer: oss-fuzz-40896
datereported: 2021-11-10
reportedby: David Korczynski
vulnerability: Out-of-memory in fuzz_init_path
product: libdwarf
description: A corrupted object. Several Elf section sizes and section offsets are larger than the file size.
https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=40896
datefixed: 2021-11-12
references: regressiontests/ossfuzz40896/clusterfuzz-testcase-fuzz_init_path-5337872492789760 regressiontests/ossfuzz40896/clusterfuzz-testcase-minimized-fuzz_init_path-5337872492789760
gitfixid: b7a119dc07c502c1334bcbf8dd04ca0e4d5f6ab6
tarrelease: libdwarf-0.4.1.tar.xz
id: DW202111-016
cve:
fuzzer: oss-fuzz-41240
datereported: 2021-11-20
reportedby: David Korczynski
vulnerability: Out-of-memory in fuzz_init_path
product: libdwarf
description: A corrupted object. The PE object section header for section .gnu_debuglink is corrupted. A very large number is in the VirtualSize field. Attempting a malloc for the section could succeed or might fail, resulting in Denial Of Service.
https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=41240
datefixed: 2021-11-21
references: regressiontests/ ossfuzz41240/clusterfuzz-testcase-minimized-fuzz_init_path-5929343686148096
gitfixid: a120c808234060c3c9b1872ab9a059aa1ac70b1d
tarrelease: libdwarf-0.4.1.tar.xz
id: DW202205-001
cve:
fuzzer:
datereported: 2022-05-26
reportedby: Casper Sun
vulnerability: buffer overflow in dwarf_globals.c
product: libdwarf
description: A carefully corrupted .debug_pubnames section would cause libdwarf to read outside of a buffer containing the section contents. That could cause a segmentation violation or other major error, terminating the calling application and resulting in Denial Of Service. The bug has been present for many years.
datefixed: 2022-05-29
references: regressiontests/sleicasper/bufferoverflow
gitfixid: 8151575a6ace77d005ca5bb5d71c1bfdba3f7069
tarrelease: libdwarf-0.4.1.tar.xz
id: DW202206-001
cve:
fuzzer:
datereported: 2022-06-15
reportedby: Casper Sun
vulnerability: buffer overflow in dwarf_form.c
product: libdwarf
description: A carefully corrupted .debug_info section would cause libdwarf to read outside of a buffer containing a Dwarf_Sig8 symbolic reference. That could cause a segmentation violation or other major error, terminating the calling application and resulting in Denial Of Service. This failure to check for buffer overflow has been present since DWARF4 when DW_FORM_ref_sig8 was added to libdwarf.
datefixed: 2022-06-15
references: regressiontests/sleicasper2/buffer-overflow-dwarf-form
gitfixid: 7ef09e1fc9ba07653dd078edb2408631c7969162
tarrelease: libdwarf-0.4.1.tar.xz
id: DW202207-001
cve:
fuzzer: ossfuzz
datereported: 2022-05-01
reportedby: David Korczynski
vulnerability: buffer overflow in dwarf_form.c
product: libdwarf
description: A carefully corrupted string would cause libdwarf to read outside of a buffer containing the string (one past the end) when checking the string to determine if it is a full path in processing a .gnu.debuglink section. That could cause a segmentation violation or other major error, terminating the calling application and resulting in Denial Of Service.
datefixed: 2022-07-23
references: regressiontests/ossfuzz47150/clusterfuzz-testcase-minimized-fuzz_init_path-6727387238236160.fuzz
gitfixid: 24dff940cc4c71a9c3cb5475aee231b19163a12c
tarrelease: libdwarf-0.5.0.tar.xz
id: DW202208-001
cve:
fuzzer: unspecified
datereported: 2022-08-27
reportedby: Han Zheng
vulnerability: Double free on corrupted frame data.
product: libdwarf
description: A carefully corrupted object file would cause libdwarf to do a double free in handling an error condition in dwarf_expand_frame_instructions(). (in libdwarf/dwarf_frame.c) That could cause a segmentation violation or other major error, terminating the calling application and resulting in Denial Of Service.
datefixed: 2022-08-27
references: regressiontests/hanzheng/fuzzedobject
gitfixid: 428235e3d132fb62faf7732735fdbb034d6264b4
tarrelease: libdwarf-0.5.0.tar.xz
id: DW202212-001
cve:
fuzzer: oss-fuzz
datereported: 2022-12-28
reportedby: David Korczynski
vulnerability: Denial of Service with fuzzed object.
product: libdwarf
description: The fuzzed testcase has at least four major errors which libdwarf did not catch, leading to unpredictable library behavior, possibly including crashing the calling program. Things not noticed before the fix (and now resulting in error being reported): A) The object has just 2 sections, too few to be real. at least 3 sections are needed to contain DWARF information of any kind. B) Section zero has non-zero contents, in violation of the Elf object specification. C) The header says section strings are in section zero (a violation of the Elf specification). D) Section 1 masquerades as .note.gnu.debug-id and the description size is gigantic (as is the section, which fits the description field length).
datefixed: 2023-01-09
references: regressiontests/ossfuzz54724/clusterfuzz-54724-poc
gitfixid: 45f6d778811553a835916b60845933e6dda63b7f
tarrelease: libdwarf-0.6.0.tar.xz
id: DW202301-001
cve:
fuzzer:
datereported: 2023-01-24
reportedby: Steve Kaufmann
vulnerability: Denial of Service with DW_FORM_strx3
product: libdwarf
description: Any use of DW_FORM_strx3 or DW_FORM_addrx3 would get libdwarf very confused and incorrect return values and or a library crash might result.
datefixed: 2023-01-24
references: regressiontests/kaufmann2/ct-bad.o
gitfixid: 97e90eb7ab98df60b8da0bdc2ac855711c4db804
tarrelease: libdwarf-0.6.0.tar.xz
id: DW202303-001
cve:
fuzzer: oss-fuzz id: 56465
datereported: 2023-03-01
reportedby: David Korczynski
vulnerability: Denial of Service with corrupt attribute
product: libdwarf
description: With a particular data corruption dwarf_attrlist() failed to return an error and the library would dereference a stale pointer. This also provoked a memory leak.
datefixed: 2023-03-02
references: regressiontests/ossfuzz56465/fuzz_die_cu_offset-5866690199289856
gitfixid: 948352178dc791796ed574a961191844d8322493
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-002
cve:
fuzzer: oss-fuzz id: 56548
datereported: 2023-03-01
reportedby: David Korczynski
vulnerability: Denial of service with corrupt line table header
product: libdwarf
description: With any one of a small set of corrupted data fields in a line table header that were not checked for sanity the library could try to malloc a giant space, which could take a long time to succeed or fail. After that almost anything could happen. Same bug fix as 56443
datefixed: 2023-03-03
references: regressiontests/ossfuzz56548/fuzz_findfuncbypc-5073632331431936
gitfixid: 89d3beccd161657760585967255bbabf67e5b4c9
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-003
cve:
fuzzer: oss-fuzz id: 56636
datereported: 2023-03-03
reportedby: David Korczynski
vulnerability: Denial of service with corrupt .debug_addr section
product: libdwarf
description: A corrupted .debug_addr header was not properly checked, and the calculation setting up the table in memory was not correctly set up. Calling dwarf_debug_addr_by_index() could crash the calling application.
datefixed: 2023-03-03
references: regressiontests/ossfuzz56636/fuzz_debug_addr_access-4801779658522624.fuzz
gitfixid: a3ab3f16ab67f4d976561fe0d863e1ed8b71f3c6
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-004
cve:
fuzzer: oss-fuzz id: 56666
datereported: 2023-03-04
reportedby: David Korczynski
vulnerability: Denial of service with corrupt gnu_index section
product: libdwarf
description: A corrupted .debug_gnu_index header was not properly checked, and the calculation setting up the table in memory was not correctly set up.
datefixed: 2023-03-08
references: regressiontests/ossfuzz56666/fuzz_gnu_index-4803574417981440
gitfixid: 64eaaa58703258cab02896e798664a1bb11a3d5c
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-005
cve:
fuzzer: oss-fuzz id: 56676
datereported: 2023-03-04
reportedby: David Korczynski
vulnerability: Denial of service with corrupt frame section.
product: libdwarf
description: A call to dwarf_expand_frame_instructions() with corrupt data gets a segmentation violation.
datefixed: 2023-03-14
references: regressiontests/ossfuzz56676/fuzz_set_frame_all-5081006119190528.fuzz
gitfixid: e564c9350c104f16eb2223d7b29082e3deb5d2fb
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-006
cve:
fuzzer: oss-fuzz id: 56456
datereported: 2023-03-04
reportedby: David Korczynski
vulnerability: Denial of service (crash) looking for gdbindex section.
product: libdwarf
description: The logic was wrong in a couple places (fixed now) and almost nothing was checked for validity. Now we check, so to libdwarf do not result in a crash of the library. The bugs have been there since the code was written in 2014.
datefixed: 2023-03-14
references: regressiontests/ossfuzz56456/fuzz_gdbindex-5240324382654464
gitfixid: e564c9350c104f16eb2223d7b29082e3deb5d2fb
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-007
cve:
fuzzer: oss-fuzz id: 56735
datereported: 2023-03-04
reportedby: David Korczynski
vulnerability: Denial of service with corrupt debug_macro section.
product: libdwarf
description: We were not checking .debug macro data for a corrupted internal macro length field. Now we check.
datefixed: 2023-03-08
references: regressiontests/ossfuzz56735/fuzz_macro_dwarf5-6718585377783808
gitfixid: bb99fe7ddb2bc6601bcb0ee30ced6a8cc8cb0564
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-008
cve:
fuzzer: oss-fuzz id: 56530
datereported: 2023-03-01
reportedby: David Korczynski
vulnerability: Denial of Service with corrupt attribute
product: libdwarf
description: With a particular data corruption dwarf_attrlist() failed to return an error and the library would dereference a stale pointer. This also provoked a memory leak. Same bug as DW202303-001
datefixed: 2023-03-02
references: ossfuzz56530/fuzz_findfuncbypc-6272642689925120
gitfixid: 948352178dc791796ed574a961191844d8322493
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-009
cve:
fuzzer: oss-fuzz id: 56443
datereported: 2023-03-01
reportedby: David Korczynski
vulnerability: Denial of service with corrupt line table header
product: libdwarf
description: With any one of a small set of corrupted data fields in a line table header that were not checked for sanity the library could try to malloc a giant space, which could take a long time to succeed or fail. After that almost anything could happen. Same as ossfuzz 56548. With the bug in 56548 fixed, now we still have a bug, there is a leak from _dwarf_special_no_dbg_error_malloc() -fsanitize does not show the bug here, but valgrind does
datefixed: 2023-05-29
references: regressiontests/ossfuzz56443/fuzz_crc_32-4750941179215872
gitfixid: 241fe0cb415569975c451d1f2d62fb2b2147cd72
tarrelease: libdwarf-0.8.0.tar.xz
id: DW202303-010
cve:
fuzzer: ossfuzz
datereported: 2023-03-07
reportedby: Youngseok Choi
vulnerability: Stack overflow dwarfdump
product: dwarfdump
description: A wholly corrupted speically constructed dwarfdump.conf which is entirely inappropriate ascii really made a mess of one's screen. Now after a couple lines of garbage we give up on that conf file. And we print the garbage sanitized to avoid messing up one's screen. The fix also avoids buffer overflow.
datefixed: 2023-03-25
references: regressiontests/choi010/poc_file
gitfixid: cb8dd45770f2e1f440aab60adac0256f268fc16e
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-011
cve:
fuzzer: ossfuzz id: 56453
datereported: 2023-02-28
reportedby: David Korczynski
vulnerability: Null pointer dereference
product: libdwarf
description: A null pointer dereference in reading a fuzzed object file. This is related to checks for a correct value from the READ_AREA_LENGTH macro. The missing checks have been missing a very long time.
datefixed: 2023-03-07
references: regressiontests/ossfuzz56453
gitfixid: 86671059c1c240ae56433fa94993dcd28df2ae7d
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-012
cve:
fuzzer: ossfuzz id: 56456
datereported: 2023-02-28
reportedby: David Korczynski
vulnerability: Heap Buffer Overflow
product: libdwarf
description: A heap buffer overflow reading a fuzzed object file Code reading the gdbindex section was not fully checking for valid offsets and pointers Duplicate of DW202303-006
datefixed: 2023-03-14
references: regressiontests/ossfuzz56456/fuzz_gdbindex-5240324382654464
gitfixid: e564c9350c104f16eb2223d7b29082e3deb5d2fb
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-013
cve:
fuzzer: ossfuzz id: 56460
datereported: 2023-02-28
reportedby: David Korczynski
vulnerability: Heap Buffer Overflow
product: libdwarf
description: A heap buffer overflow reading a fuzzed object file A revision of dwarf_str_offsets (which was new in DWARF5) affected several source files
datefixed: 2023-03-12
references: regressiontests/ossfuzz56460/fuzz_str_offsets-5376904040677376-5240324382654464
gitfixid: 0343c63bd04d387924974e6da60d8471fdf945a9
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-014
cve:
fuzzer: ossfuzz id: 56478
datereported: 2023-02-28
reportedby: David Korczynski
vulnerability: Heap Buffer Overflow in read_single_rle_entry
product: libdwarf
description: A heap buffer overflow reading a fuzzed object file A revision of dwarf_str_offsets (which was new in DWARF5) affected several source files
datefixed: 2023-03-12
references: regressiontests/ossfuzz56478/fuzz_rng-5030515398017024
gitfixid: 0343c63bd04d387924974e6da60d8471fdf945a9
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-015
cve:
fuzzer: ossfuzz id: 56489
datereported: 2023-02-28
reportedby: David Korczynski
vulnerability: Heap Buffer Overflow in read_single_rle_entry
product: libdwarf
description: A heap buffer overflow reading a fuzzed object file A revision of dwarf_str_offsets (which was new in DWARF5) affected several source files
datefixed: 2023-03-12
references: regressiontests/ossfuzz56489/fuzz_srcfiles-5091530466787328
gitfixid: 0343c63bd04d387924974e6da60d8471fdf945a9
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-016
cve:
fuzzer: ossfuzz id: 56476
datereported: 2023-02-28
reportedby: David Korczynski
vulnerability: Heap Buffer Overflow in dwarf_get_rnglist_offset_value
product: libdwarf
description: A heap buffer overflow reading a fuzzed object file A revision of dwarf_str_offsets (which was new in DWARF5) affected several source files
datefixed: 2023-03-12
references: regressiontests/ossfuzz56476/fuzz_rng-5008229349588992/
gitfixid: 0343c63bd04d387924974e6da60d8471fdf945a9
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-017
cve:
fuzzer: ossfuzz id: 56450
datereported: 2023-02-28
reportedby: David Korczynski
vulnerability: Stack Buffer Overflow
product: libdwarf
description: Stack buffer overflow in dwarf_dietype_offset. Reading a corrupted object file. The bug was in the test code, not libdwarf. Having fixed that, valgrind finds that some memory is not freed by dwarf_finish(). In dwarf_alloc.c March 20 _dwarf_free_all_of_one_debug() if the Dwarf_Debug was normal we failed to call _dwarf_free_static_errlist() and that left memory allocated from a bogus earlier call to the library (a situation libdwarf should handle and now does).
datefixed: 2023-05-30
references: regressiontests/ossfuzz56450/fuzz_die_cu_attrs-4953133005799424
gitfixid:
tarrelease: libdwarf-0.8.0.tar.xz
id: DW202303-018
cve:
fuzzer: ossfuzz id: 56807
datereported: 2023-02-28
reportedby: David Korczynski
vulnerability: Memory leak reading line table
product: libdwarf
description: A memory leak in _dwarf_read_line_table() reading a particular corrupted object. One return failed to free a local malloc. An very old bug, encountered reading corrupted DWARF line tables.
datefixed: 2023-03-20
references: regressiontests/ossfuzz56807fuzz_srcfiles-4626047380619264
gitfixid: 1ff4365bd64a4e4f9ab717b3f62589c2ba6637a5
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-019
cve:
fuzzer: ossfuzz id: 56454
datereported: 2023-02-28
reportedby: David Korczynski
vulnerability: Stack buffer overflow
product: libdwarf
description: The overflow was in dwarf_get_version_of_die() One return failed to free a local malloc due to the particular corruption in the text object DWARF.
datefixed: 2023-03-20
references: regressiontests/ossfuzz56454/
gitfixid: 1ff4365bd64a4e4f9ab717b3f62589c2ba6637a5
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-020
cve:
fuzzer: ossfuzz id: 56458
datereported: 2023-02-28
reportedby: David Korczynski
vulnerability: Heap buffer overflow
product: libdwarf
description: The overflow encountered when reading a corrupted line table header in read_a_name_table_header in dwarf_debugnames.c There was insufficient checking for out of bounds values.
datefixed: 2023-03-20
references: regressiontests/ossfuzz56458/fuzz_findfuncbypc-5073632331431936
gitfixid: 1ff4365bd64a4e4f9ab717b3f62589c2ba6637a5
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-021
cve:
fuzzer: ossfuzz id: 56487
datereported: 2023-02-28
reportedby: David Korczynski
vulnerability: Memory Leak
product: testing harness
description: The leak was due to the test driver fuzz/fuzz_rng.c failing to call dwarf_finish at the point of returning as a result of the corrupted binary returning an error condition.
datefixed: 2023-03-20
references: regressiontests/ossfuzz56487/clusterfuzz-testcase-fuzz_rng-6655451078197248
gitfixid: 1ff4365bd64a4e4f9ab717b3f62589c2ba6637a5
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-022
cve:
fuzzer: ossfuzz id: 56497
datereported: 2023-02-28
reportedby: David Korczynski
vulnerability: Memory Leak
product: test harness
description: The leak was due to the test driver fuzz/fuzz_rng.c failing to call dwarf_finish at the point of returning as a result of the corrupted binary returning an error condition.
datefixed: 2023-03-20
references: regressiontests/ossfuzz56497/
gitfixid: 1ff4365bd64a4e4f9ab717b3f62589c2ba6637a5
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-023
cve:
fuzzer: ossfuzz id: 56568
datereported: 2023-02-28
reportedby: David Korczynski
vulnerability: Test build failure
product: Test harness
description: The fuzz_dnames.c testcase build failed. A result of removing two functions from the API, dwarf_dnames_abbrev_by_code() and dwarf_dnames_abbrev_form_by_index(). Removed 2023-02-23. The test source no longer uses those two functions. The first was slow and hard to use. The second was unusable and never worked. The documentation (libdwarf.pdf) gives alternates in the library that work.
datefixed: 2023-03-20
references:
gitfixid: 2eced75af9903ab778c3b237ec7be3ddc93ea6ec
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-024
cve:
fuzzer: ossfuzz id: 56807
datereported: 2023-02-28
reportedby: David Korczynski
vulnerability: Memory Leak in dwarf_check_lineheader_b()
product: libdwarf
description: The fuzzed test object file resulted in a memory leak calling dwarf_check_lineheader_b() as called from fuzz_srcfiles.c Two error conditions in dwarf_line_table_reader_common.h were missing a required free().
datefixed: 2023-03-24
references: regressiontests/ossfuzz56807/fuzz_srcfiles-4626047380619264
gitfixid: 484f50ef8be0506be2e4b5fbad489868db5c7985
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-025
cve:
fuzzer: ossfuzz id: 56895
datereported: 2023-02-28
reportedby: David Korczynski
vulnerability: Heap buffer overflow reading compilation unit header.
product: libdwarf
description: Calling dwarf_next_cu_header_d() on the fuzzed test object results in a library crash in fuzz_die_cu_attrs_loclist.c due to a failure to precisely test for a too-short Compilation Unit header. Now a DW_DLV_ERROR is returned. A very old bug.
datefixed: 2023-03-24
references: regressiontests/ossfuzz56895/fuzz_macro_dwarf5-5080340952907776
gitfixid: 771cfcca1ef6a4a7eb9595d700fc72020d0ed72e
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-026
cve:
fuzzer: ossfuzz id: 56897
datereported: 2023-02-28
reportedby: David Korczynski
vulnerability: Heap buffer overflow reading the rnglists section
product: libdwarf
description: dwarf_get_rnglist_offset_index_value() could fail on a corrupt object due to imprecise calculations of entry offsets. Fixed by a major update of the code in dwarf_rnglists.c
datefixed: 2023-03-22
references: regressiontests/ossfuzz56897/fuzz_rng-5105415777288192
gitfixid: b9393bb9b6399a34f8616a272d030bdd004a5ef5
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-027
cve:
fuzzer: ossfuzz id: 56906
datereported: 2023-03-09
reportedby: David Korczynski
vulnerability: Heap Buffer overflow reading rnglists section.
product: libdwarf
description: Calling dwarf_get_rnglist_rle() on a corrupted object file could result in a library crash.
datefixed: 2023-03-22
references: regressiontests/ossfuzz56906/fuzz_rng-6031783801257984.fuzz
gitfixid: b9393bb9b6399a34f8616a272d030bdd004a5ef5
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-028
cve:
fuzzer: ossfuzz id: 56958
datereported: 2023-03-12
reportedby: David Korczynski
vulnerability: Out of memory crash.
product: libdwarf
description: Failing to check for error conditions in a fuzzed object correctly lead to a giant malloc that could not succeed.
datefixed: 2023-03-22
references: regressiontests/ossfuzz56958/fuzz_stack_frame_access-6097292873826304
gitfixid: b9393bb9b6399a34f8616a272d030bdd004a5ef5
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-029
cve:
fuzzer: ossfuzz id: 56993
datereported: 2023-03-12
reportedby: David Korczynski
vulnerability: Leaked Memory
product: libdwarf
description: Calling dwarf_get_macro_context on a particular fuzzed object file results in a memory leak when a particular error in the corrupted section is detected. The malloc was done by the line table reader code. The bug was there for many years
datefixed: 2023-03-22
references: regressiontests/ossfuzz56993/fuzz_macro_dwarf5-5770464300761088
gitfixid: 5fde5e404a98c6727889cf14d8f93ec2138a6fa
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-030
cve:
fuzzer: ossfuzz id: 57027
datereported: 2023-03-12
reportedby: David Korczynski
vulnerability: Infinite loop reading a gnu index section.
product: libdwarf
description: Calling dwarf_next_cu_header_d() on a corrupted object file results in an infinite loop and (eventually) a crash in dwarf_xu_index.c attempting to resolve an 8 byte hash key. The bug existed from the first version of this source file, which was in 2017 as the data involved DWARF5, new in 2017.
datefixed: 2023-03-22
references: regressiontests/ossfuzz57027/fuzz_stack_frame_access-5123569972805632
gitfixid: 774f98e596df9dd8f3cb92ec76243caaa4287039
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-031
cve:
fuzzer: ossfuzz id: 57048
datereported: 2023-03-14
reportedby: David Korczynski
vulnerability:
product: libdwarf
description: Calling dwarf_next_cu_header_d() on a corrupted object file results in an infinite loop and (eventually) a crash in dwarf_xu_index.c attempting to resolve an 8 byte hash key. The bug existed from the first version of this source file. The same bug as DW202303-030.
datefixed: 2023-03-22
references: regressiontests/ossfuzz57048/fuzz_findfuncbypc-4647942385696768
gitfixid: 774f98e596df9dd8f3cb92ec76243caaa4287039
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-032
cve:
fuzzer: ossfuzz id: 57107
datereported: 2023-03-14
reportedby: David Korczynski
vulnerability: Infinite loop till out of memory.
product: libdwarf
description: The infinite loop reading a fuzzed object file was caused by letting internal function _dwarf_find_CU_Context_given_sig() unconditionally do too much in the middle of setting up a CU_Context (by letting it start more CU_contexts(). The implicit infinite loop has been there a few years, depending on the correctness of object files DWARF4/DWARF5 being read.
datefixed: 2023-03-23
references: regressiontests/ossfuzz57107/fuzz_die_cu_attrs_loclist-4991396240293888
gitfixid: 0c92ef5b66c5bbcacae03fbf355b12713151c098
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-033
cve:
fuzzer: ossfuzz id: 57149
datereported: 2023-03-15
reportedby: David Korczynski
vulnerability: Stack overflow
product: libdwarf
description: Infinite loop till out of memory Similar to 57107 DW202303-032 but revealed there were more places in find_cu_die_base_fields that needed to call the internal _dwarf_internal_global_formref_b() function. The bug was present since 2017, when DWARF5 support for new 'base' fields was created.
datefixed: 2023-03-24
references: regressiontests/ossfuzz57149/fuzz_srcfiles-6213793811398656
gitfixid: 7165918c8594061c3f5ba7dd4df7c4555c68ec78
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-034
cve:
fuzzer: ossfuzz id: 57193
datereported: 2023-03-15
reportedby: David Korczynski
vulnerability: Infinite loop till out of memory.
product: libdwarf
description: The infinite loop reading a fuzzed object file was caused by letting internal function _dwarf_find_CU_Context_given_sig() unconditionally do too much in the middle of setting up a CU_Context (by letting it start more CU_contexts(). The implicit infinite loop has been there a few years, depending on the correctness of object files DWARF4/DWARF5 being read. Same bug as ossfuzz 57107, ossfuzz 57149.
datefixed: 2023-03-24
references: regressiontests/ossfuzz57193/fuzz_die_cu_offset-5215024489824256
gitfixid: 7165918c8594061c3f5ba7dd4df7c4555c68ec78
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-035
cve:
fuzzer: ossfuzz id: 57292
datereported: 2023-03-15
reportedby: David Korczynski
vulnerability: Out of Memory
product: libdwarf
description: Another case of the infinite loop due to _dwarf_find_CU_Context_given_sig(). See DW202303-034 57193 and others listed here.
datefixed: 2023-03-24
references: regressiontests/ossfuzz57292/fuzz_die_cu_print-5412313393135616
gitfixid: 7165918c8594061c3f5ba7dd4df7c4555c68ec78
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-036
cve:
fuzzer: ossfuzz id: 57300
datereported: 2023-03-15
reportedby: David Korczynski
vulnerability: Out of Memory
product: libdwarf
description: Another case of the infinite loop due to _dwarf_find_CU_Context_given_sig(). See DW202303-034 57193 See DW202303-034 57149 See DW202303-034 57193 See DW202303-034 57292
datefixed: 2023-03-24
references: regressiontests/ossfuzz57300/fuzz_die_cu-4752724662288384
gitfixid: 7165918c8594061c3f5ba7dd4df7c4555c68ec78
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-037
cve:
fuzzer: ossfuzz id: 57300
datereported: 2023-03-21
reportedby: David Korczynski
vulnerability: Out of Memory
product: libdwarf
description: Another case of the infinite loop due to _dwarf_find_CU_Context_given_sig(). See DW202303-034 57193 and others listed here.
datefixed: 2023-03-24
references: regressiontests/ossfuzz57300/fuzz_die_cu-4752724662288384
gitfixid: 7165918c8594061c3f5ba7dd4df7c4555c68ec78
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-038
cve:
fuzzer: ossfuzz id: 57335
datereported: 2023-03-22
reportedby: David Korczynski
vulnerability: Null dereference in dwarf_hasform()
product: libdwarf
description: Passing a null dw_return_bool pointer dereferenced zero, but now, instead we return DW_DLV_ERROR with error code DW_DLE_INVALID_NULL_ARGUMENT The test driver fuzz_die_cu_attrs.c passed in a NULL argument (the test was not modified except for adding a comment) but libdwarf now checks for a null argument.
datefixed: 2023-03-24
references: regressiontests/ossfuzz57335/fuzz_die_cu_attrs-6235345560928256.fuzz
gitfixid: e4053c9a0f25db0bed28372d9b77a50a0307dc10
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-039
cve:
fuzzer: ossfuzz id: 56480
datereported: 2023-02-28
reportedby: David Korczynski
vulnerability: Reading a compilation unit never finishes
product: libdwarf
description: Reading a DWARF compilation unit header in a corrupted object caused an infinite loop of repeated calls (growing the stack at each call) in libdwarf. Now the library properly reflects a NO ENTRY case avoiding the loop and the test case returns an unrelated error due to other corruption. Arguably the loop was due to corruption too, but it should not have gotten stuck in the loop (and now it will not get stuck). See also ossfuzz id 56540. ossfuzz id 56487. ossfuzz id 56497. ossfuzz 57408
datefixed: 2023-03-20
references: regressiontests/ossfuzz56480/fuzz_die_cu_print-5264022485467136
gitfixid: 1ff4365bd64a4e4f9ab717b3f62589c2ba6637a5
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-040
cve:
fuzzer: ossfuzz
datereported: 2023-03-07
reportedby: Youngseok Choi
vulnerability: Reading from zero page
product: dwarfdump
description: The command: dwarfdump --file-abi= causes big problems. Denial of service. For many of such commands with an '=' and nothing following the dwoptarg variable is not set at all, leading to a segmentation violation.
datefixed: 2023-03-25
references: regressiontests/choi011/README
gitfixid: fd92b647e5e3a524be94b3b06c9efd14a8292946
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-041
cve:
fuzzer: ossfuzz
datereported: 2023-03-08
reportedby: Youngseok Choi
vulnerability: Heap Use after Free
product: dwarfdump
description: The command: dwarfdump --check-frame-extended choi012/poc_file_06 is now working, apparently fixed by earlier bug fixes in libdwarf.
datefixed: 2023-03-25
references: regressiontests/choi012/poc_file_06
gitfixid: fd92b647e5e3a524be94b3b06c9efd14a8292946
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-042
cve:
fuzzer: ossfuzz
datereported: 2023-03-08
reportedby: Youngseok Choi
vulnerability: Memory Leak
product: dwarfdump
description: The command: dwarfdump --check-all choi012/poc_file_07 leaked memory in dwarfdump. A single instance of print_error_and_continue() failed to return DW_DLV_ERROR when libdwarf reported an error. Before recent fixes to libdwarf this would also generate heap-use-after-free.
datefixed: 2023-03-25
references: regressiontests/choi012/poc_file_07
gitfixid: 4a8a201cdb3408a2cfdc2946418b51b884140a2c
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-043
cve:
fuzzer: ossfuzz
datereported: 2023-03-08
reportedby: Youngseok Choi
vulnerability: Heap Use After Free
product: libdwarf
description: The command: dwarfdump choi012/poc_file_08 was resulting in a heap use after free. It no longer does, due to earlier libdwarf fixes.
datefixed: 2023-03-25
references: regressiontests/choi012/poc_file_08
gitfixid: 4a8a201cdb3408a2cfdc2946418b51b884140a2c
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-044
cve:
fuzzer: ossfuzz
datereported: 2023-03-08
reportedby: Youngseok Choi
vulnerability: Heap buffer overflow
product: dwarfdump
description: The command: dwarfdump --search-regex=\\ (with no file named) resulted in a heap use after free. No object file is required for the test. The dd_regex.c regex compiler failed to notice when a trailing backslash caused the pattern compiler to step past the pattern string. Now we notice and issue an error.
datefixed: 2023-03-25
references: regressiontests/baselines/choi013.base
gitfixid: 3269f43d2a044bfcce71d30ce214a305473d1ea3
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-045
cve:
fuzzer: ossfuzz
datereported: 2023-03-08
reportedby: Youngseok Choi
vulnerability: Heap buffer overflow, regex
product: dwarfdump
description: The command: dwarfdump --search-regex= with a simple, really long, string. (with no file named) resulted in a buffer overflow in a buffer in dd_regex.c. No object file is required for the test. The code now checks stores to the internal array holding the non-deterministic finite automata (nfa)
datefixed: 2023-03-25
references: regressiontests/baselines/choi014.base
gitfixid: bb8fab9e5e4e40b1268b31d90882c2ab93653eaf
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-046
cve:
fuzzer: ossfuzz
datereported: 2023-03-08
reportedby: Youngseok Choi
vulnerability: global buffer overflow
product: dwarfdump
description: The command: dwarfdump --search-regex=t[ą--]e Note the non-ascii character. The dwarfdump regex only allows ascii in search patterns, not the rest of UTF-8.
datefixed: 2023-03-26
references: regressiontests/choi012/poc_file_10
gitfixid: 9eac0c8bbae3fadb2be3d5ee15b9c44f42d2f966
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-047
cve:
fuzzer: ossfuzz
datereported: 2023-03-08
reportedby: Youngseok Choi
vulnerability: Double Free
product: dwarfdump
description: The command: dwarfdump --check-unique --check-abbrev results in a double free. The table of unique errors contained makename() data, so that aspect caused a double free as makename gets destructed independently. Fixed now,letting makename destructor do the work.
datefixed: 2023-03-24
references: regressiontests/choi015/poc_file_04
gitfixid: df64db4740f1b480e602b1112107d51f0d269828
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-048
cve:
fuzzer: ossfuzz
datereported: 2023-03-08
reportedby: Youngseok Choi
vulnerability: Memory leak (was double free).
product: dwarfdump
description: The command: dwarfdump --file-name=<file> -kG -ka <objectfile> results in a a memory leak. In certain error cases we failed to fclose() a FILE * used to read dwarfdump.conf. Earlier changes fixed the double free, this fixes the memory leak..
datefixed: 2023-03-24
references: regressiontests/choi015/poc_file_03
gitfixid: 24f5697aecd77092de20f0f7e7d91fbc1f2b3da0
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-049
cve:
fuzzer: ossfuzz id: 57408
datereported: 2023-03-24
reportedby: David Korczynski
vulnerability: Stack Overflow, _dwarf_create_a_new_cu_context...
product: libdwarf
description: involves find_sig8_target_as_global_offset() and is the same problem as seen in other guises earlier. The case becomes an infinite loop, so eventually the stack gets exhausted. Fixed. See also ossfuzz id 56540. ossfuzz id 56487. ossfuzz id 56497. ossfuzz 57480
datefixed: 2023-03-26
references: regressiontests/ossfuzz57408/fuzz_die_cu-4702098356043776
gitfixid: 24f5697aecd77092de20f0f7e7d91fbc1f2b3da0
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-050
cve:
fuzzer: ossfuzz id: 57429
datereported: 2023-03-24
reportedby: David Korczynski
vulnerability: invalid free() in test source
product: libdwarf
description: The test source violates libdwarf requirements. fuzz/fuzz_die_cu_attrs.c was doing free on a name pointer returned from dwarf_diename. The documentation clearly states that pointer should not have a free() done. fix id below is fixing the test source.
datefixed: 3023-03-28
references: regressiontests/ossfuzz57429/fuzz_die_cu_attrs-4845537731149824
gitfixid: 2b19bc239f3cedd1b2461e4265d90633277ce704
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-051
cve:
fuzzer: ossfuzz id: 57437
datereported: 2023-03-24
reportedby: David Korczynski
vulnerability: Heap double free
product: libdwarf
description: In a specific error case reading a fuzzed object and calling dwarf_srcfiles local data was freed twice. The bug was fixed earlier, and involved src/lib/libdwarf/dwarf_line_table_reader_common.h.
datefixed: 2023-03-28
references: regressiontests/ossfuzz57437/fuzz_srcfiles-5281689109921792
gitfixid: c25a14c3fd5522aff0b1d2a77d7ee66b7c529779
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-052
cve:
fuzzer: ossfuzz id: 57442
datereported: 2023-03-24
reportedby: David Korczynski
vulnerability: Heap buffer overflow
product: libdwarf
description: Corrupt .debug_rngslists leads to crash when a rnglists header has a length indicating a longer section than we really have. Now we check more carefully for that situation. The bug existed from 2017, when DWARF5 support was added to the library.
datefixed: 2023-03-28
references: regressiontests/ossfuzz57442/fuzz_rng-5974595378479104
gitfixid: 271b9b8367a8151fcd98723d73382ec56f05c810
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-053
cve:
fuzzer: ossfuzz id: 57443
datereported: 2023-03-24
reportedby: David Korczynski
vulnerability: Double free in _dwarf_read_line_table_header
product: libdwarf
description: The same bug seen earlier A double free when a particular error is in the line table header. Fixed already. gitfixid is more recent than truly required.
datefixed: 2023-03-28
references: regressiontests/ossfuzz57443/fuzz_srcfiles-6015429578719232
gitfixid: c25a14c3fd5522aff0b1d2a77d7ee66b7c529779
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-054
cve:
fuzzer: ossfuzz id: 57463
datereported: 2023-03-24
reportedby: David Korczynski
vulnerability: dereference null in test code
product: none, test code bug
description: The fault was in test code. fuzz_die_cu_attrs.c Since fixed (earlier today). No code change here.
datefixed: 2023-03-24
references: regressiontests/ossfuzz57463/fuzz_die_cu_attrs-5158380196200448
gitfixid: e4053c9a0f25db0bed28372d9b77a50a0307dc10
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-055
cve:
fuzzer: ossfuzz id: 57485
datereported: 2023-03-27
reportedby: David Korczynski
vulnerability:
product: None, test code bug
description: Abort in fuzz_die_cu_attrs.c The fault was in test code. Since fixed (earlier today). No code change here.
datefixed: 2023-03-28
references: regressiontests/ossfuzz57485/
gitfixid: 2b19bc239f3cedd1b2461e4265d90633277ce704
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-056
cve:
fuzzer: ossfuzz id: 57516
datereported: 2023-03-28
reportedby: David Korczynski
vulnerability: Null dereference in dwarf_hasattr()
product: libdwarf
description: With a corrupted attribute dwarf_hasattr() could try to access an implicit_const abbrev value indexing off of a NULL library internal pointer. Because the abbrev section had no actual implicit const value due to the corruption, so the internal array for holding such was not present. The pointer abl_implicit_const was NULL. Now we test the pointer for NULL and if NULL report an error. This lack of a NULL check has existed for many years.
datefixed: 2023-03-29
references: regressiontests/ossfuzz57516/fuzz_die_cu_attrs-6171488289161216
gitfixid: 5dc3de5ce70331692a2700b218fb79e0d4d81c23
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-057
cve:
fuzzer:
datereported: 2023-03-26
reportedby: Pedro Navarro
vulnerability: Unable to read large object sections
product: libdwarf
description: A section 2GB+ in size could not be read by libdwarf. Such is a Denial of Service. Simply turning the big read into however many are needed (each below 2GB) was simple to do. The limitation in the 'read' libc function (really a Linux kernel limitation) is well documented but we had not noticed before now. Few object files are so large.
datefixed: 2023-03-28
references:
gitfixid: 8bf96199a0e130483cceca6bfacfbe4127441ab1
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-058
cve:
fuzzer: ossfuzz id: 57527
datereported: 2023-03-29
reportedby: David Korczynski
vulnerability: reading off end of valid data can crash library
product: libdwarf
description: A line table header truncated by a fuzzer to just the right length (anywhere within a 12 byte area) would cause memory references outside of valid data. Now we check there is object data present before referring to that area and if not, return an error.
datefixed: 2023-03-30
references: regressiontests/ossfuzz57527/fuzz_srcfiles-4599045397282816
gitfixid: 36e4063ade31c9ea6ea5df973d2045b36877885b
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202303-059
cve:
fuzzer: ossfuzz id: 57562
datereported: 2023-03-30
reportedby: David Korczynski
vulnerability: Infinite loop reading DIEs
product: libdwarf
description: Caller looping on dwarf_siblingof_b() (wanting to touch all siblings) could be put in an infinite loop. A DW_AT_sibling attribute with a corrupted attribute value meant the caller never sees the DW_DLV_NO_ENTRY return signalling all siblings have been seen.
datefixed: 2023-04-01
references: regressiontests/ossfuzz57562/fuzz_findfuncbypc-6681114772373504
gitfixid: 21b076f652992c03f145f6edeb623918e17693f8
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202304-001
cve:
fuzzer: ossfuzz id: 57711
datereported: 2023-04-04
reportedby: David Korczynski
vulnerability: dereference null pointer
product: libdwarf
description: Crash in libdwarf on dwarf_srcfiles() call. A dereference off a null pointer due to corrupt file numbers not being noticed. Any such crash left an incomplete and misleading stack trace. The large numbers treated as Dwarf_Signed were part of the problem. Now in libdwarf we check Dwarf_Signed for negative values and issue an error if the value less than 0. So later casts to Dwarf_Unsigned work as intended. The libdwarf problems have been in the library for a very long time.
datefixed: 2023-04-06
references: regressiontests/ossfuzz57711/fuzz_srcfiles-4695324781576192
gitfixid: da0d1efbeddcff23c25704bd9672e98314928b19
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202304-002
cve:
fuzzer: ossfuzz id: 57766
datereported: 2023-04-07
reportedby: David Korczynski
vulnerability: Heap Buffer Overflow
product: libdwarf
description: Crash in libdwarf on reading an attribute due to failing to check that an index into .debug_str_offsets is sane. So we read far outside the relevant table. The checks have been incomplete since this DWARF5 section code was written. Two functions in libdwarf dwarf_form.c had the same problem.
datefixed: 2023-04-09
references: regressiontests/ossfuzz57766/fuzz_die_cu_print-5295062170075136
gitfixid: 761da806fc950c6b26c1763e8989a814e9b16a59
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202304-003
cve:
fuzzer: ossfuzz id: 57887
datereported: 2023-04-10
reportedby: David Korczynski
vulnerability: Reading outside the intended section data.
product: libdwarf
description: Crash in libdwarf on reading .debug_addr given a bogus index entry. Due to failing to correctly check that the index is out of range. The index was close to overflowing Dwarf_Unsigned so testing values *after* arithmetic done on the incoming index was too late: so we read outside the .debug_addr table. The checks have been incomplete since this DWARF5 section code was written. libdwarf/dwarf_query.c
datefixed: 2023-04-11
references: regressiontests/ossfuzz57887/fuzz_die_cu-4866423964172288
gitfixid: 1729d9af3f690bece912ae0f625b312566d0ae25
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202304-004
cve:
fuzzer: ossfuzz id: 58026
datereported: 2023-04-15
reportedby: David Korczynski
vulnerability: Segv on unknown address reading frame data.
product: libdwarf
description: On reading a corrupt frame register number the library could crash with a segmentation violation. This bug has been present in the code for 25 years. The conversion of an impossibly large (carefully constructed) register number to a Dwarf_Half or unsigned int the result looked reasonable, invalidating some tests for reasonableness. Now we do all the tests on the full Dwarf_Unsigned register number(s) and retain the value in the long form everywhere. Fixed 2023-04-15. Once that is fixed there is still a leak found by valgrind. Tht test code fuzz/fuzz_set_frame_all.c does a local malloc and in some cases returned without free-ing it locally. Now that local malloc has the necessary local free.
datefixed: 2023-05-30
references: regressiontests/ossfuzz58026/fuzz_set_frame_all-4582976972521472.fuzz
gitfixid: 21b33d13024d18b09e32914ca5718a5c81d1ad67
tarrelease: libdwarf-0.8.0.tar.xz
id: DW202305-001
cve:
fuzzer: ossfuzz id: 58769
datereported: 2023-05-09
reportedby: David Korczynski
vulnerability: Excessive malloc reading experimental line table
product: libdwarf
description: A fuzzed line table in the non-standard (experimental) two-level line table format had several unchecked values and one was larger than made any sense. The failure was due to oss-fuzz limiting malloc to 3GB. The failure was appropriate as the fuzzed values were inappropriate. We now check for sensible values. See libdwarf/dwarf_line_table_reader_common.h The code was in libdwarf starting in 2021.
datefixed: 2023-05-09
references: regressiontests/ossfuzz58769/fuzz_macro_dwarf5-5460713058205696
gitfixid: edc241bd0bf22c94d2d496f3cb761e60f066cd14
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202305-002
cve:
fuzzer: ossfuzz id: 58797
datereported: 2023-05-10
reportedby: David Korczynski
vulnerability: Memory Leak reading experimental line table
product: libdwarf
description: A fuzzed line table in the non-standard (experimental) two-level line table format had several unchecked values and one was larger than made any sense, and detecting the error revealed a memory leak. Caused by the incomplete fix to DW202305-001 (yesterday). This is was a failure to run some crucial tests, which would have exposed the problem before DW202305-001 was completed. Incomplete testing.
datefixed: 2023-05-10
references: regressiontests/ossfuzz58797/fuzz_macro_dwarf5-4872686367801344
gitfixid: eeb935200f78b8509e6b1837f6825b9d551b9f7d
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202305-003
cve:
fuzzer: ossfuzz id: 59091
datereported: 2023-05-19
reportedby: David Korczynski
vulnerability: Incorrect section bound check
product: libdwarf
description: A fuzzed line table in the non-standard (experimental) two-level line table format exposed a failure as the test was v > sectionend whereas it has to be v >= sectionend as end pointers are always one-past the end of the area. This was incorrect since the experimental table support was added in 2021.
datefixed: 2023-05-19
references: regressiontests/ossfuzz59091/fuzz_macro_dwarf5-5135813562990592
gitfixid: 4017ab8b92195641e6876b388cebe2d3307634f5
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202305-004
cve:
fuzzer: ossfuzz id: 56446
datereported: 2023-02-28
reportedby: David Korczynski
vulnerability: Incorrect section bound check
product: libdwarf test code
description: The test program fuzz_dnames.c passed a non-null pointer containing garbage content. The fix is to initialize (in fuzz_dnames.c) the local variable to null (0).
datefixed: 2023-05-23
references: regressiontests/ossfuzz56446/fuzz_dnames-4784811358420992
gitfixid: 6fac1021c67d72da6b65f99ad815978d40b4c1e8
tarrelease: libdwarf-0.8.0.tar.xz
id: DW202305-005
cve:
fuzzer: ossfuzz id: 56462
datereported: 2023-02-28
reportedby: David Korczynski
vulnerability: Unpredictable crash or erroneous data returned
product: libdwarf
description: A call to dwarf_set_frame_undefined_value() dwarf_set_frame_rule_initial_value() dwarf_set_frame_same_value() dwarf_set_frame_cfa_value() dwarf_set_frame_rule_table_size() with unusable values was not being caught. Now if the set of values violates the required relationships an error is returned on requesting actual frame data. The problem has existed for many years (fixed May 23). Once that is fixed valgrind shows leaks. That is because fuzz/fuzz_set_frame_all.c fails to call dwarf_finish() and, instead, simple exit()s at several places. Updated the test source to return from its functions and only exit() from main() after the dwarf_finish() call.
datefixed: 2023-05-30
references: regressiontests/ossfuzz56462/fuzz_set_frame_all-5424385441005568
gitfixid: 21b33d13024d18b09e32914ca5718a5c81d1ad67
tarrelease: libdwarf-0.8.0.tar.xz
id: DW202305-006
cve:
fuzzer: ossfuzz id: 56472
datereported: 2023-02-27
reportedby: David Korczynski
vulnerability: Crash on null pointer argument.
product: libdwarf
description: A call to any dwarf_get_fission or any API entry with _xu_ in the name (functions for DWARF5 Debug Fission, called Split Dwarf in DWARF5) would crash the caller if any relevant argument was null. The problem has existed since the code was written in 2021. Once that is fixed valgrind complains about using an uninitialized value. fuzz/fuzz_simplereader_tu.c calls libdwarf with the declation being Dwarf_Die die; no initializer present. Bad behavior, even a library crash is likely.
datefixed: 2023-05-30
references: regressiontests/ossfuzz56472/fuzz_simplereader_tu-6614412934119424
gitfixid: 8b17d41a31c33e0b3b9727a8044e0093a754d6d7
tarrelease: libdwarf-0.8.0.tar.xz
id: DW202305-007
cve:
fuzzer: ossfuzz id: 56474
datereported: 2023-02-28
reportedby: David Korczynski
vulnerability: Null pointer dereference.
product: libdwarf
description: Calling dwarf_highpc_b() lead to crash. The function was dereferencing an argument before the argument was checked. Now it is checked for null before any dereference. In addition, the test code, fuzz/fuzz_die_cu_attrs_loclist.c, called dwarf_highpc_b() with a Dwarf_Die that is uninitialized local variable (hence contents unpredictable). C code has no way to catch such a caller error. This is a bug in the test code. We are changed the test code local data pointer variable to be initialized with the value 0.
datefixed: 2023-05-23
references: regressiontests/ossfuzz56474/fuzz_die_cu_attrs_loclist-4719938125561856
gitfixid: b3df2530732ea515cda5a85438871e15c6723ead
tarrelease: libdwarf-0.8.0.tar.xz
id: DW202305-008
cve:
fuzzer: ossfuzz id: 56492
datereported: 2023-02-28
reportedby: David Korczynski
vulnerability: Timeout (exceeds 50 seconds)
product: fuzz testing
description: The problem is a bug in the test source: fuzz/fuzz_macro_dwarf5.c. One must not blame the fuzzer author for examplep5(), the code was based on doc/checkexamples.c and there examplep() was really just a sketch. The testcase here no longer specifies an infinite loop.
datefixed: 2023-05-23
references: regressiontests/ossfuzz56492/fuzz_macro_dwarf5-6497277180248064
gitfixid: 97a78122268c9a74701f2dd3115f902309e9a484
tarrelease: libdwarf-0.8.0.tar.xz
id: DW202305-009
cve:
fuzzer: ossfuzz id: 56451
datereported: 2023-02-28
reportedby: David Korczynski
vulnerability: Write to memory fails
product: libdwarf
description: One problem is a bug in the test source: fuzz/fuzz_dnames.c. It calls dwarf_dnames_abbrevtable() incorrectly. The caller is required to provide arrays dw_idxattr_array and dw_form_array and pass a pointer to such. The code was just passing in a pointer to nothing. The library code has no possible way to determine the passed in pointers are usable. In addition, dwarf_dnames_abbrevtable() did not check that pointers passed in were non-null before use, but now it does.
datefixed: 2023-05-30
references: regressiontests/ossfuzz56451/fuzz_dnames-4986494365597696
gitfixid: 12a612fc8db38fc26cd5e6064f09a6f825891c7c
tarrelease: libdwarf-0.8.0.tar.xz
id: DW202305-010
cve:
fuzzer: ossfuzz id: 59478
datereported: 2023-05-31
reportedby: David Korczynski
vulnerability: Memory leak in dwarf_expand_frame_instructions()
product: libdwarf
description: Fuzzing provoked one of four error cases that could leak locally allocated memory from _dwarf_exec_frame_instructor() (called by dwarf_expand_frame_instructions). The code did free(localregtab) but needed to do FREELOCALMALLOC, a macro specific to this function which cleans up all local allocations. All four places have been corrected. Called enough times with fuzzed data could result in filling memory leading to the library being unable to work for the caller and instead just returning errors. This bug has been present in the code for many years.
datefixed: 2023-05-31
references: regressiontests/ossfuzz59478/fuzz_set_frame_all-5300774457180160
gitfixid: 8ef9c8fb613e59f534e789e91a73088eaa5b8a5a
tarrelease: libdwarf-0.8.0.tar.xz
id: DW202306-001
cve:
fuzzer: ossfuzz id: 59517
datereported: 2023-06-01
reportedby: David Korczynski
vulnerability: Signed Integer Overflow
product: libdwarf
description: Nine different places in dwarf_frame.c multiplied a factored value by a (usually) small integer without checking if the factored value read from the object could possibly be real. So the factored value when multiplied by the factor could overflow. In some of the cases the factored value is signed, some it is unsigned. This sanity checking of factored frame offset values never existed before in the library.
datefixed: 2023-06-08
references: regressiontests/ossfuzz59517/fuzz_set_frame_all-5741671019839488
gitfixid: f664f93d456284130afbd3c2e35b39e5f2740366
tarrelease: libdwarf-0.8.0.tar.xz
id: DW202306-002
cve:
fuzzer: ossfuzz id: 59519
datereported: 2023-06-01
reportedby: David Korczynski
vulnerability: Integer Overflow
product: libdwarf
description: Integer Overflow in _dwarf_exec_frame_instr() called by dwarf_expand_frame_instructions. We now check for overflows in add and multiply here. Similar to ossfuzz 59517
datefixed: 2023-06-08
references: regressiontests/ossfuzz59519/fuzz_set_frame_all-4670829255065600
gitfixid: f664f93d456284130afbd3c2e35b39e5f2740366
tarrelease: libdwarf-0.8.0.tar.xz
id: DW202306-003
cve:
fuzzer: shinibufa
datereported: 2023-06-09
reportedby: shinibufa (github)
vulnerability: use after free
product: libdwarf
description: Heap use-after-free dwarf_query.c
datefixed: 2023-05-19
references: regressiontests/shinibufa/fuzzed_input_file
gitfixid: 4017ab8b92195641e6876b388cebe2d3307634f5
tarrelease: libdwarf-0.7.0.tar.xz
id: DW202306-004
cve:
fuzzer: ossfuzz id: 59595
datereported: 2023-06-09
reportedby: shinibufa (github)
vulnerability: Signed Integer overflow
product: libdwarf
description: Signed Integer Overflow. In _dwarf_exec_frame_instr(), called by dwarf_expand_frame_instructions(), there was a DW_CFA_LLVM_def_aspace_cfa_sf and we failed to check for overflow. The test case had a overflow. Now we do that check.
datefixed: 2023-06-10
references: regressiontests/ossfuzz59595/fuzz_set_frame_all-5319697747542016
gitfixid: e8c726e2be644df2706342b7a80633d07ecd7fb4
tarrelease: libdwarf-0.8.0.tar.xz
id: DW202306-005
cve:
fuzzer: ossfuzz id: 59717
datereported: 2023-06-01
reportedby: David Korczynski
vulnerability: Integer Overflow
product: libdwarf
description: Integer Overflow in _dwarf_exec_frame_instr() called by dwarf_expand_frame_instructions. We now check for overflows in add and multiply here. Similar to ossfuzz 59517
datefixed: 2023-06-08
references:
gitfixid: f664f93d456284130afbd3c2e35b39e5f2740366
tarrelease: libdwarf-0.8.0.tar.xz
id: DW202306-006
cve:
fuzzer: ossfuzz id: 59727
datereported: 2023-06-01
reportedby: David Korczynski
vulnerability: Integer Overflow
product: libdwarf
description: Integer Overflow in _dwarf_exec_frame_instr() called by dwarf_expand_frame_instructions. We now check for overflows in add and multiply here. Similar to ossfuzz 59517
datefixed: 2023-06-08
references:
gitfixid: f664f93d456284130afbd3c2e35b39e5f2740366
tarrelease: libdwarf-0.8.0.tar.xz
id: DW202306-007
cve:
fuzzer: ossfuzz id: 59602
datereported: 2023-06-04
reportedby: David Korczynski
vulnerability: Buffer overflow read
product: libdwarf
description: In _dwarf_read_loc_expr_op() we read one byte past available data as the required check for past-end was missing.
datefixed: 2023-06-10
references: regressiontests/ossfuzz59602/fuzz_die_cu_attrs_loclist-6737086749999104
gitfixid: c8c54ba5c79b0a2687f0fa2ac331479506c3210f
tarrelease: libdwarf-0.8.0.tar.xz
id: DW202306-008
cve:
fuzzer: ossfuzz id: 59699
datereported: 2023-06-08
reportedby: David Korczynski
vulnerability: Read past end of section
product: libdwarf
description: In reading a CIE prefix of a fuzzed object we read past the end of the section due to a failure to check a byte pointer before we dereference it in _dwarf_read_cie_fde_prefix().
datefixed: 2023-05-10
references: regressiontests/ossfuzz59699/fuzz_stack_frame_access-6523659305746432
gitfixid: c5b909630bb566cdbf68fae4091f049f3b22ff11
tarrelease: libdwarf-0.8.0.tar.xz
id: DW202306-009
cve:
fuzzer: ossfuzz id: 59775
datereported: 2023-06-11
reportedby: David Korczynski
vulnerability: Fuzzed object results in read past end of section.
product: libdwarf
description: A fuzzed object results in reading one byte past the end of a .eh_frame section in internal function _dwarf_read_loc_expr-op(). Now we check for that before we dereference a pointer (to read the particular single-byte field).
datefixed: 2023-06-13
references: regressiontests/ossfuzz59775/fuzz_die_cu_attrs_loclist-4504718844755968
gitfixid: 9cae1be75ec333d2b8ab8800df4850ed77a8b025
tarrelease: libdwarf-0.8.0.tar.xz
id: DW202306-010
cve:
fuzzer: ossfuzz id: 59950
datereported: 2023-06-18
reportedby: David Korczynski
vulnerability:library reads outside frame section
product: libdwarf
description: A fuzzed object results in adding a too large value (from CIE frame augmentation data) to a pointer, having failed to check the value for reasonableness. That add overflows so dereferencing the pointer in dwarf_frame.c could lead to a crash in the library or getting nonsense information returned to the caller. This bug has been present for many years.
datefixed: 2023-06-19
references: regressiontests/ossfuzz59950/fuzz_set_frame_all-6613067367317504
gitfixid: b7437c9e4923906e9b3f3860a0c8a8289cff0a91
tarrelease: libdwarf-0.8.0.tar.xz
id: DW202306-011
cve:
fuzzer: ossfuzz id: 60090
datereported: 2023-06-24
reportedby: David Korczynski
vulnerability: Read from invalid memory address
product: libdwarf
description: A fuzzed object results in an addition overflow in reading CIE data leading to a read from an invalid address. An almost-correct check for an overflow in case of a fuzzed aug_irix_exception_table augmentation leads to a crash. The bug was incorrect coding of a test (for an absurd value) written a few weeks ago.
datefixed: 2023-06-26
references: regressiontests/ossfuzz60090/fuzz_set_frame_all-5757752673435648
gitfixid: 6f75899f1f90fa87e52da0df09ddaa2e5be778f9
tarrelease: libdwarf-0.8.0.tar.xz
id: DW202307-001
cve:
fuzzer: ossfuzz id: 60506
datereported: 2023-07-09
reportedby: David Korczynski
vulnerability: Read from outside section
product: libdwarf
description: A fuzzed object results in reading outside a line table due to a corruption in a non-standard (experimental) line table format. A corrupted offset was not checked for sanity. The bug has been in the code since the experimental line table support was added in 2015.
datefixed: 2023-07-11
references: regressiontests/ossfuzz60506/fuzz_srcfiles-6494439909228544.fuzz
gitfixid: c8c5073f35b1efdcc610ecf369c78f87fdd34714
tarrelease: libdwarf-0.8.0.tar.xz
id: DW202308-001
cve:
fuzzer: ossfuzz id: 59576
datereported: 2023-06-04
reportedby: David Korczynski
vulnerability: Read from outside frame section
product: libdwarf
description: A fuzzed object results in reading outside a frame section due to a comparison being > when it should have bin >= at about line 1615 in dwarf_frame2. Could result in crash or incorrect frame data returned. Somehow we lost track of this open bug. The bug has been in the code since the augmentation was first implemented in the library.
datefixed: 2023-08-26
references: regressiontests/ossfuzz59576/fuzz_set_frame_all-5867083595120640
gitfixid: e53adc90ffd6d5d0fad61546b0041990aefd970b
tarrelease: libdwarf-0.8.0.tar.xz
id: DW202309-001
cve:
fuzzer: ossfuzz id: 62547
datereported: 2023-09-22
reportedby: David Korczynski
vulnerability: Heap use after free
product: libdwarf
description: Calling dwarf_get_fde_for_die() causes the problem as its special handling in a user calling for fde destruction is wrong when dwarf_finish() is calling for fde destruction. dwarf_finish() can refer to freed memory in trying to delete a CIE twice. The use after free has a dependence on the order nodes are seen in the de_alloc_tree tdestroy() walk of the table (the order is not predictable). Broken in release 0.8.0 and all previous releases.
datefixed: 2023-09-23
references: regressiontests/ossfuzz62547/fuzz_stack_frame_access-5263709637050368
gitfixid: cd741379bd0203a0875b413542d5f982606ae637
tarrelease: libdwarf-0.9.0.tar.xz
id: DW202309-002
cve:
fuzzer: ossfuzz id: 62833
datereported: 2023-09-30
reportedby: David Korczynski
vulnerability: Memory leak in _dwarf_macho_setup()
product: libdwarf
description: The September 29 addition of Mach-O universal binary support resulted in memory leaks due to revising memory alloc/free incompletely when fuzzed object files were encountered.. This bug was in the public repository for two days (in all-new code, for Apple Universal Binary objects).
datefixed: 2023-10-01
references: regressiontests/ossfuzz62833/fuzz_set_frame_all-4521858130903040
gitfixid: gitfixid: f7c7e84e5a77915bb6570215887118d8e7759122
tarrelease: libdwarf-0.9.0.tar.xz
id: DW202309-003
cve:
fuzzer: ossfuzz id: 62834
datereported: 2023-09-30
reportedby: David Korczynski
vulnerability: Memory leak in _dwarf_macho_setup()
product: libdwarf
description: The September 29 addition of Mach-O universal binary support resulted in memory leaks due to revising memory alloc/free incompletely when fuzzed object files were encountered.. This bug was in the public repository for two days (in all-new code, for Apple Universal Binary objects).
datefixed: 2023-10-01
references: regressiontests/ossfuzz62834/fuzz_init_path-4573857635500032
gitfixid: f7c7e84e5a77915bb6570215887118d8e7759122
tarrelease: libdwarf-0.9.0.tar.xz
id: DW202309-004
cve:
fuzzer: ossfuzz id: 62842
datereported: 2023-09-30
reportedby: David Korczynski
vulnerability: Heap buffer overflow
product: libdwarf
description: The heap buffer overflow in _dwarf_memcpy_swap_bytes was due a failure to check for a valid size field (a fuzzed value) in a count of array elements.. Now we check for a sensible count. This bug was in the public repository for two days (in all-new code, for Apple Universal Binary objects).
datefixed: 2023-10-01
references: regressiontests/ossfuzz62842/fuzz_findfuncbypc-4964619766333440.fuzz
gitfixid: f7c7e84e5a77915bb6570215887118d8e7759122
tarrelease: libdwarf-0.9.0.tar.xz
id: DW202310-001
cve:
fuzzer: ossfuzz id: 62943
datereported: 2023-10-02
reportedby: David Korczynski
vulnerability: Heap buffer overflow
product: libdwarf
description: The heap buffer overflow was due to a failure to do initial sanity checks on a universal object. The object involved was not large enough to have a complete universal header. This bug was in the public repository for three days (in all-new code, for Apple Universal Binary objects).
datefixed: 2023-10-03
references: regressiontests/ossfuzz62943/fuzz_init_path-5486726493372416
gitfixid: aea77dad8745d9aad5275c3226e4e3156effa71f
tarrelease: libdwarf-0.9.0.tar.xz
id: DW202310-002
cve:
fuzzer: ossfuzz id: 63024
datereported: 2023-10-06
reportedby: David Korczynski
vulnerability: Heap buffer overflow
product: libdwarf
description: A copy-paste error lead to a heap buffer overflow. Named the wrong struct in calling calloc(). The function with the bug was added seven days ago.
datefixed: 2023-10-07
references: regressiontests/ossfuzz63024/fuzz_init_path-5486726493372416
gitfixid: 3a658bd1dd7437948cecbf82bb9b24f5f6122a7d
tarrelease: libdwarf-0.9.0.tar.xz
id: DW202311-001
cve:
fuzzer: ossfuzz id: 56452
datereported: 2023-11-24
reportedby: David Korczynski
vulnerability: Null dereference
product: libdwarf
description: Passing a null Dwarf_Debug to dwarf_add_debuglink_global_path() lead to library crash. The code was not checking for a valid Dwarf_Debug Argument. The bug was present when the function was created in 2021 Moreover, oss fuzz originally noted the bug on 02 March 2023 but I can find no trace of a notification of the bug arriving before 24 November 2023. Fixing this sort of thing for all functions, here is the last commit id... ef77596af000719c04bd3e40b97139247ff3efb4
datefixed: 2023-11-25
references: regressiontests/ossfuzz56452/fuzz_debuglink-cs4231a-5927365017731072
gitfixid: 1f6988307748f427566e3266695bb72d5384bf3d
tarrelease: libdwarf-0.9.0.tar.xz
id: DW202311-002
cve:
fuzzer: ossfuzz id: 64496
datereported: 2023-02-28
reportedby: David Korczynski
vulnerability: Null dereference from dwarf_gnu_debuglink()
product: libdwarf
description: If the Dwarf_Debug was opened with dwarf_init_object_b() there is no pathname known to libdwarf and the library was dereferencing a null pointer as a result. With the library bug fixed the fuzz/fuzz_debuglink.c test case was violating the rules of use of the function resulting in memory leakage. The documentation has been improved on this function.
datefixed: 2023-11-25
references: regressiontests/ossfuzz64496/fuzz_debuglink-615437663823462
gitfixid: d76cce559b898f7059ce5ffd82f3cfd58cb392fe
tarrelease: libdwarf-0.9.0.tar.xz
id: DW202402-001
cve:
fuzzer: ossfuzz id: 66646
datereported: 2024-02-12
reportedby: David Korczynski
vulnerability: Reference memory outside of section.
product: libdwarf
description: The data pointer for DW_FORM_ref1 was not being validated as pointing into the section before being dereferenced. Here reading a corrupted DWARF section. A library crash is likely. This test failure happened to be on a DW_AT_abstract_origin attribute, but the problem applies in many other situations. Nearly the other forms had checks. This check has been missing for many years. Similarly the requisite check was missing from DW_FORM_block1 for many years and that too was fixed so all the FORMs have checks now.
datefixed: 2024-02-13
references: regressiontests/ossfuzz66646/fuzz_findfuncbypc-5178544143532032
gitfixid: f21e2f7687f3dca183026a1fb72ca1f0dcf8befa
tarrelease: libdwarf-0.9.2.tar.xz
id: DW202402-002
cve: CVE-2024-2002
fuzzer: hongg
datereported: 2024-02-16
reportedby: ifygecko
vulnerability: crashes randomly on fuzzed object
product: libdwarf
description: In a multiply-corrupted DWARF object libdwarf may try to dealloc(free) an allocation twice. Results are unpredictable and various. This has been a possibility since we added code to prevent leaks when generating 'unattached' Dwarf_Error records (where there is no Dwarf_Debug available at the point of error). The problem was introduced in libdwarf-0.1.0 in 2021.
datefixed: 2024-02-17
references: regressiontests/hongg2024-02-16/SIGABRT-a.fuzz SIGABRT-b.fuzz SIGABRT-c.fuzz SIGSEGV-d.fuzz SIGSEGV-e.fuzz SIGSEGV-f.fuzz SIGSEGV-g.fuzz SIGSEGV-h.fuzz SIGSEGV-i.fuzz SIGSEGV-k.fuzz
gitfixid: 404e6b1b14f60c81388d50b4239f81d461b3c3ad
tarrelease: libdwarf-0.9.2.tar.xz
id: DW202402-003
cve:
fuzzer: hongg
datereported: 2024-02-18
reportedby: ifygecko
vulnerability: crashes randomly reading fuzzed locllist
product: libdwarf
description: A carefully corrupted loclists entry can cause libdwarf to read outside of its allowed areas in dwarf_loclists.c due to lack of a sanity check. A segmentation error and libdwarf crash is likely. Similar code in dwarf_rnglists.c and that now has the additional checks. The bugs have been present in both since the code was created in June 2020.
datefixed: 2024-02-18
references: regressiontests/hongg2024-02-18/SIGSEGV-m.fuzz
gitfixid: 5cfbd87dff4fc3c3b595bb92ed886934945b372c
tarrelease: libdwarf-0.9.2.tar.xz
id: DW202403-001
cve:
fuzzer: ossfuzz id: 67490
datereported: 2024-03-18
reportedby: David Korczynski
vulnerability: Reads past end of line table
product: libdwarf
description: A carefully corrupted line table header can cause libdwarf to read outside of its allowed areas in a .debug_line section reading the file names part of the header. The failure to check for end-of-section before reading past end-of-section at the very last byte in section (at a very few specific points in the line table reader code where a valid line table header would not require a test) has been present for many years.
datefixed: 2024-02-19
references: regressiontests/ossfuzz67490/fuzz_srcfiles-5195296927711232
gitfixid: 2930f3121ee6b07da405103934c329bbeca0382f
tarrelease: libdwarf-0.9.2.tar.xz
id: DW202406-001
cve:
fuzzer: ossfuzz id: 69639
datereported: 2024-06-14
reportedby: David Korczynski
vulnerability: Memory Leak reading .debug_loclists
product: libdwarf
description: During startup (reading initial fields from .debug_loclists) libdwarf was allocating an array of integers but in the normal case was failing to free the array. This bug in DWARF5 handling has been in the code for a week or so. See also DW202406-002.
datefixed: 2024-06-15
references: regressiontests/ossfuzz69639/fuzz_die_cu_offset-6001910176350208
gitfixid: 32d832900ebe2e61ec07e82625a561415be05424
tarrelease: libdwarf-0.10.1.tar.xz
id: DW202406-002
cve:
fuzzer: ossfuzz id: 69641
datereported: 2024-06-14
reportedby: David Korczynski
vulnerability: Memory Leak reading .debug_loclists
product: libdwarf
description: During startup (reading initial fields from .debug_loclists) libdwarf was allocating an array of integers but in the normal case was failing to free the array. This bug in DWARF5 handling has been in the code for a week or so. See also DW202406-001.
datefixed: 2024-06-15
references: regressiontests/ossfuzz69641/fuzz_die_cu_attrs_loclist-6271271030030336
gitfixid: 32d832900ebe2e61ec07e82625a561415be05424
tarrelease: libdwarf-0.10.1.tar.xz
id: DW202407-001
cve:
fuzzer: ossfuzz id: 70244
datereported: 2024-07-09
reportedby: David Korczynski
vulnerability: Memory leaks reading rnglists
product: libdwarf
description: Libdwarf was failing to free() some allocations in reading .debug_rnglists. This bug has been present for a week or so.
datefixed: 2024-07-09
references: regressiontests/ossfuzz70244/fuzz_die_cu_attrs_loclist-4958134427254784
gitfixid: 906a4428a5d92e17948da4249cfccbe8f5ae8005
tarrelease:
id: DW202407-002
cve:
fuzzer: ossfuzz id: 70246
datereported: 2024-07-09
reportedby: David Korczynski
vulnerability: Huge malloc request could crash caller
product: libdwarf
description: Libdwarf was not checking a field in a .debug_rnglists header for sanity before doing a malloc based on the field value. This bug has been present for a week or so.
datefixed: 2024-07-10
references: regressiontests/ossfuzz70246/fuzz_macro_dwarf5-5128935898152960
gitfixid: d7c4efdcc7952b38a237a36ccedf364018e0fb1c
tarrelease:
id: DW202407-003
cve:
fuzzer: ossfuzz id: 70256
datereported: 2024-07-10
reportedby: David Korczynski
vulnerability: Huge malloc request could crash caller
product: libdwarf
description: Libdwarf was not checking a field in a .debug_rnglists header for sanity before doing a malloc based on the field value. This bug has been present for a week or so.
datefixed: 2024-07-10
references: regressiontests/ossfuzz70256/fuzz_rng-483822291655065
gitfixid: d7c4efdcc7952b38a237a36ccedf364018e0fb1c
tarrelease:
id: DW202407-004
cve:
fuzzer: ossfuzz id: 70263
datereported: 2024-07-10
reportedby: David Korczynski
vulnerability: Huge malloc request could crash caller
product: libdwarf
description: Libdwarf was not checking a field in a .debug_rnglists header for sanity before doing a malloc based on the field value. This bug has been present for a week or so.
datefixed: 2024-07-10
references: regressiontests/ossfuzz70263/fuzz_die_cu-4960441042796544
gitfixid: d7c4efdcc7952b38a237a36ccedf364018e0fb1c
tarrelease:
id: DW202407-005
cve:
fuzzer: ossfuzz id: 70266
datereported: 2024-07-10
reportedby: David Korczynski
vulnerability: Huge malloc request could crash caller
product: libdwarf
description: Libdwarf was not checking a field in a .debug_rnglists header for sanity before doing a malloc based on the field value. This bug has been present for a week or so.
datefixed: 2024-07-10
references: regressiontests/ossfuzz70266/fuzz_findfuncbypc-6093996460408832
gitfixid: d7c4efdcc7952b38a237a36ccedf364018e0fb1c
tarrelease:
id: DW202407-006
cve:
fuzzer: ossfuzz id: 70277
datereported: 2024-07-10
reportedby: David Korczynski
vulnerability: Huge malloc request could crash caller
product: libdwarf
description: Libdwarf was not checking a field in a .debug_rnglists header for sanity before doing a malloc based on the field value. This bug has been present for a week or so.
datefixed: 2024-07-10
references: regressiontests/ossfuzz70277/
gitfixid: d7c4efdcc7952b38a237a36ccedf364018e0fb1c
tarrelease:
id: DW202407-007
cve:
fuzzer: ossfuzz id: 70278
datereported: 2024-07-10
reportedby: David Korczynski
vulnerability: Huge malloc request could crash caller
product: libdwarf
description: Libdwarf was not checking a field in a .debug_rnglists header for sanity before doing a malloc based on the field value. This bug has been present for a week or so.
datefixed: 2024-07-10
references: regressiontests/ossfuzz70278/
gitfixid: d7c4efdcc7952b38a237a36ccedf364018e0fb1c
tarrelease:
id: DW202407-008
cve:
fuzzer: ossfuzz id: 70282
datereported: 2024-07-10
reportedby: David Korczynski
vulnerability: Huge malloc request could crash caller
product: libdwarf
description: Libdwarf was not checking a field in a .debug_rnglists header for sanity before doing a malloc based on the field value. This bug has been present for a week or so.
datefixed: 2024-07-10
references: regressiontests/ossfuzz70282/
gitfixid: d7c4efdcc7952b38a237a36ccedf364018e0fb1c
tarrelease:
id: DW202407-009
cve:
fuzzer: ossfuzz id: 70287
datereported: 2024-07-10
reportedby: David Korczynski
vulnerability: Huge malloc request could crash caller
product: libdwarf
description: Libdwarf was not checking a field in a .debug_rnglists header for sanity before doing a malloc based on the field value. This bug has been present for a week or so.
datefixed: 2024-07-10
references: regressiontests/ossfuzz70287/
gitfixid: d7c4efdcc7952b38a237a36ccedf364018e0fb1c
tarrelease:
id: DW202407-010
cve:
fuzzer: ossfuzz id: 70721
datereported: 2024-07-27
reportedby: David Korczynski
vulnerability: Heap Use After Free
product: libdwarf
description: Libdwarf was referencing freed space attempting to free up a compilation-unit DIE in the process of creating a context for a compilation-unit DIE, given a particular corruption of the DWARF data being read. This bug has been present for several years.
datefixed: 2024-07-27
references: regressiontests/ossfuzz70721/fuzz_macro_dwarf5-4907954017468416
gitfixid: 6fa96f95e043bac9b98ca6f7a9a542dae8f46cd
tarrelease:
id: DW202407-011
cve:
fuzzer: ossfuzz id: 70753
datereported: 2024-07-28
reportedby: David Korczynski
vulnerability: InfiniteRecursion reading CUs leads to crash
product: libdwarf
description: The code added in git fix id 6b40a0ed378826273080a7b11e7274c2b61d018b two days ago involved adding code setting up compilation unit data and in find_cu_die_base_fields() we called dwarf_global_formref() whereas we must call _dwarf_internal_global_formref_b() to avoid an infinite recursion. A crash was only possible when reading corrupt DWARF.
datefixed: 2024-07-28
references: regressiontests/ossfuzz70753/fuzz_die_cu_offset-6598270743281664
gitfixid: 1b79d618bf5aab2bda9be495c531b13e94ae056a
tarrelease:
id: DW202407-012
cve:
fuzzer: ossfuzz id: 70763
datereported: 2024-07-28
reportedby: David Korczynski
vulnerability: InfiniteRecursion reading CUs leads to crash
product: libdwarf
description: The code added in git fix id 6b40a0ed378826273080a7b11e7274c2b61d018b two days ago involved adding code setting up compilation unit data and in find_cu_die_base_fields() we called dwarf_global_formref() whereas we must call _dwarf_internal_global_formref_b() to avoid an infinite recursion. A crash was only possible when reading corrupt DWARF.
datefixed: 2024-07-28
references: regressiontests/ossfuzz70763/fuzz_macro_dwarf5-5161075908083712
gitfixid: 1b79d618bf5aab2bda9be495c531b13e94ae056a
tarrelease: