Fix possible crash in UdpContext host mocking.#9312
Open
everslick wants to merge 2 commits intoesp8266:masterfrom
Open
Fix possible crash in UdpContext host mocking.#9312everslick wants to merge 2 commits intoesp8266:masterfrom
everslick wants to merge 2 commits intoesp8266:masterfrom
Conversation
I saw this crash in UdpContext.h happen after some hours of regular NTP requests.... ``` [17:38:51] ADE: voltage spike: phase A, factor=0.81, cycles=1 (mock) ClientContext::connect: ::connect(): Connection refused (mock) ClientContext::connect: ::connect(): Connection refused (mock) ClientContext::connect: ::connect(): Connection refused (mock) ClientContext::connect: ::connect(): Connection refused [17:41:14] ADE: voltage spike: phase A, factor=0.86, cycles=2 (mock) ClientContext::connect: ::connect(): Connection refused (mock) ClientContext::connect: ::connect(): Connection refused (mock) ClientContext::connect: ::connect(): Connection refused (mock) ClientContext::connect: ::connect(): Connection refused [17:43:49] ADE: voltage spike: phase B, factor=1.12, cycles=3 (mock) ClientContext::connect: ::connect(): Connection refused (mock) =====> UdpServer port: 2390 <===== [17:44:31] NTP: waiting for response from '81.169.233.252' timed out Program received signal SIGSEGV, Segmentation fault. 0x000055555556848c in std::_Function_base::_M_empty() const () (gdb) where #0 0x000055555556848c in std::_Function_base::_M_empty() const () esp8266#1 0x00005555555685be in std::function<void ()>::operator bool() const () esp8266#2 0x00005555555684b8 in UdpContext::mock_cb() () esp8266#3 0x00005555555683af in check_incoming_udp() () esp8266#4 0x0000555555567fa6 in main () ``` I tracked it down to UdpContext::mock_cb() which is calling a std::function callback that's been destroyed thus used-after-free. I believe what happens is: 1. UdpContext::unref() calls delete this when refcount hits 0 (line 56) 2. The destructor (~UdpContext(), line 47) is empty — it never calls register_udp(_sock, nullptr) to remove itself from the global udps map 3. check_incoming_udp() iterates udps, finds the dangling pointer, calls mock_cb() on freed memory 4. _on_rx (a destroyed std::function) → segfault in _M_empty() The disconnect() method already does the right thing (lines 78-84: closes socket + unregisters), but it's never called before the object is destroyed via unref(). The fix is to make the destructor clean up properly. I tested intensely, no crashes since.
Fixes code-style CI.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
I saw this crash in UdpContext.h happen after some hours of regular NTP requests....
I tracked it down to UdpContext::mock_cb() which is calling a std::function callback that's been destroyed thus used-after-free. I believe what happens is:
The disconnect() method already does the right thing (lines 78-84: closes socket + unregisters), but it's never called before the object is destroyed via unref().
The fix is to make the destructor clean up properly.
I tested intensely, no crashes since.