[libunwind] Fix getSLEB128 on large values

[rprichard]

Summary

Previously, for large-enough values, getSLEB128 would attempt to shift
a signed int in the range [0..0x7f] by 28, 35, 42... bits, which is
undefined behavior and likely to fail.

Avoid shifting (-1ULL) by 70 for large values. e.g. For INT64_MAX, the
last two bytes will be:

• 0x7f [bit==56]
• 0x00 [bit==63]

[rprichard]

The shift for negative numbers is also broken.

[mstorsjo]

LGTM

[smeenai]

Would it make sense to pick this to 11.0? CC @hans

[hans]

In D83742#2155006, @smeenai wrote:

Would it make sense to pick this to 11.0? CC @hans

Sounds good to me. Please go ahead and cherry-pick with the -x option, or let me know if you'd prefer me to do it.

[smeenai]

Would it make sense to pick this to 11.

In D83742#2155822, @hans wrote:

In D83742#2155006, @smeenai wrote:

Would it make sense to pick this to 11.0? CC @hans

Sounds good to me. Please go ahead and cherry-pick with the -x option, or let me know if you'd prefer me to do it.

Sorry, I wasn't sure if this was directed to me or Ryan (the diff author). I'd be happy to cherry-pick, but I don't have a way of testing it.

[hans]

Sorry, I wasn't sure if this was directed to me or Ryan (the diff author). I'd be happy to cherry-pick, but I don't have a way of testing it.

I meant you, since you brought it up :-) But if there's no pressing need we can also just let it wait for the next release.

[rprichard]

AFAIK, there's no regression, and no one has hit this issue in practice, so maybe it can wait until the next release? It's probably only an issue when the size of something exceeds 2**32?

(I'm more worried about https://bugs.llvm.org/show_bug.cgi?id=46743, because that is a regression.)

libunwind/src/AddressSpace.hpp

inline int64_t LocalAddressSpace::getSLEB128(pint_t &addr, pint_t end) {
   const uint8_t *pend = (uint8_t *)end;
   int64_t result = 0;
   int bit = 0;
   uint8_t byte;
   do {
     if (p == pend)
       _LIBUNWIND_ABORT("truncated sleb128 expression");
     byte = *p++;
-    result |= ((byte & 0x7f) << bit);
+    result |= (uint64_t)(byte & 0x7f) << bit;
     bit += 7;
   } while (byte & 0x80);
   // sign extend negative numbers
-  if ((byte & 0x40) != 0)
+  if ((byte & 0x40) != 0 && bit < 64)
     result |= (-1ULL) << bit;
   addr = (pint_t) p;
   return result;
 }
 
 inline LocalAddressSpace::pint_t
 LocalAddressSpace::getEncodedP(pint_t &addr, pint_t end, uint8_t encoding,
                                pint_t datarelBase) {