[libunwind] Fix wrong end argument passed to decodeEHHdr() #68813
Conversation
|
@llvm/pr-subscribers-libunwind Author: Alexander Richardson (arichardson) ChangesIt previously took a start+end pint_t, but all but one callsite were actually passing start+length arguments. This should not have any functional change since the end argument is almost always ignored. I noticed this while debugging some incorrect error messages being printed while running the testsuite baremetal (using binaries that did not have a valid eh_frame_hdr section): the tests print Full diff: https://github.com/llvm/llvm-project/pull/68813.diff 1 Files Affected:
diff --git a/libunwind/src/EHHeaderParser.hpp b/libunwind/src/EHHeaderParser.hpp
index ed4317c89055c9e..2162178e10bb2fb 100644
--- a/libunwind/src/EHHeaderParser.hpp
+++ b/libunwind/src/EHHeaderParser.hpp
@@ -35,7 +35,7 @@ template <typename A> class EHHeaderParser {
uint8_t table_enc;
};
- static bool decodeEHHdr(A &addressSpace, pint_t ehHdrStart, pint_t ehHdrEnd,
+ static bool decodeEHHdr(A &addressSpace, pint_t ehHdrStart, size_t ehHdrSize,
EHHeaderInfo &ehHdrInfo);
static bool findFDE(A &addressSpace, pint_t pc, pint_t ehHdrStart,
uint32_t sectionLength,
@@ -53,8 +53,9 @@ template <typename A> class EHHeaderParser {
template <typename A>
bool EHHeaderParser<A>::decodeEHHdr(A &addressSpace, pint_t ehHdrStart,
- pint_t ehHdrEnd, EHHeaderInfo &ehHdrInfo) {
+ size_t ehHdrSize, EHHeaderInfo &ehHdrInfo) {
pint_t p = ehHdrStart;
+ pint_t ehHdrEnd = ehHdrStart + ehHdrSize;
uint8_t version = addressSpace.get8(p++);
if (version != 1) {
_LIBUNWIND_LOG("unsupported .eh_frame_hdr version: %" PRIu8 " at %" PRIx64,
@@ -106,7 +107,7 @@ bool EHHeaderParser<A>::findFDE(A &addressSpace, pint_t pc, pint_t ehHdrStart,
pint_t ehHdrEnd = ehHdrStart + sectionLength;
EHHeaderParser<A>::EHHeaderInfo hdrInfo;
- if (!EHHeaderParser<A>::decodeEHHdr(addressSpace, ehHdrStart, ehHdrEnd,
+ if (!EHHeaderParser<A>::decodeEHHdr(addressSpace, ehHdrStart, sectionLength,
hdrInfo))
return false;
|
All but one callsite were actually passing start+length arguments. This should not have any functional change since the end argument is almost always ignored. I noticed this while debugging some incorrect error messages being printed while running the testsuite baremetal (using binaries that did not have a valid eh_frame_hdr section): the tests print `libunwind: unsupported .eh_frame_hdr version: 20 at 8000d308` because libunwind is reading nonsense data for .eh_frame_hdr.
arichardson
force-pushed
the
libunwind-eh-hdr-arg
branch
from
f5b7f1f to
ff360ee
Compare
arichardson
changed the title
I'd like to keep the two commits separate since they are addressing different bugs. |
This would be a lot easier if I could chose rebase+merge in the PR flow. |
All but one callsite were actually passing start+length arguments.
This should not have any functional change since the end argument is
almost always ignored.
I noticed this while debugging some incorrect error messages being
printed while running the testsuite baremetal (using binaries that did
not have a valid eh_frame_hdr section): the tests print
libunwind: unsupported .eh_frame_hdr version: 20 at https://github.com/arichardson/upstream-llvm-project/commit/8000d308146ebf49cb364cb600e28a0a42e22c83becauselibunwind is reading nonsense data for .eh_frame_hdr.