[valgrind] [Bug 460142] New: Auxiliary stack traces
https://bugs.kde.org/show_bug.cgi?id=460142 Bug ID: 460142 Summary: Auxiliary stack traces Classification: Developer tools Product: valgrind Version: unspecified Platform: Other OS: Linux Status: REPORTED Severity: wishlist Priority: NOR Component: memcheck Assignee: jsew...@acm.org Reporter: simon.rich...@hogyros.de Target Milestone: --- SUMMARY I'm trying to narrow down a use-after-free bug with interned strings in Python. It is fairly obvious what is happening, someone gave a dynamically allocated string to the Python interpreter with a promise that its lifetime would exceed that of the interpreter (so the string data could be shared), and later deallocated the string. The three stack traces I get (allocation, deallocation and use) show what is happening, but it is difficult for me to find the point where the string is given to Python -- the function to intern the string is called quite often, so I can't just easily break there. What would be nice would be a user request to attach an auxiliary stack trace to a block of memory so more lifetime events can be traced -- in my case I could for example save the stack trace for the call that adds the string to the list of interned strings, so I can see which context made the lifetime promise that later on wasn't kept. -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 415621] epoll_ctl reports for uninitialized padding
https://bugs.kde.org/show_bug.cgi?id=415621 --- Comment #2 from Simon Richter --- Not yet, I'd need guidance on what would be the desired behaviour for uninitialized padding in the middle of a struct (which is harmless but might leak data if it is blindly copied around). Is there a flag for hiding these warnings, or a separate reporting level? -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 415621] New: epoll_ctl reports for uninitialized padding
https://bugs.kde.org/show_bug.cgi?id=415621 Bug ID: 415621 Summary: epoll_ctl reports for uninitialized padding Product: valgrind Version: unspecified Platform: Other OS: Linux Status: REPORTED Severity: normal Priority: NOR Component: memcheck Assignee: jsew...@acm.org Reporter: simon.rich...@hogyros.de Target Milestone: --- SUMMARY On platforms where the epoll_event struct is padded (>= 64 bit architectures except amd64), the padding is reported uninitialized unless the program explicitly uses memset. STEPS TO REPRODUCE 1. On powerpc64, compile and run #include int main(int, char **) { epoll_event evt; evt.events = 0; evt.data.ptr = nullptr; epoll_ctl(0, EPOLL_CTL_ADD, 0, ); } OBSERVED RESULT The padding is reported as uninitialized. EXPECTED RESULT No error, or an indication that this is mostly harmless. -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 413119] ioctl wrapper for DRM_IOCTL_I915_GEM_MMAP
https://bugs.kde.org/show_bug.cgi?id=413119 Simon Richter changed: What|Removed |Added CC||simon.rich...@hogyros.de -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 413119] ioctl wrapper for DRM_IOCTL_I915_GEM_MMAP
https://bugs.kde.org/show_bug.cgi?id=413119 --- Comment #2 from Simon Richter --- Created attachment 123299 --> https://bugs.kde.org/attachment.cgi?id=123299=edit Patch to handle ioctl -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 413119] ioctl wrapper for DRM_IOCTL_I915_GEM_MMAP
https://bugs.kde.org/show_bug.cgi?id=413119 --- Comment #3 from Simon Richter --- Created attachment 123300 --> https://bugs.kde.org/attachment.cgi?id=123300=edit Patch to handle legacy ioctl -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 413119] ioctl wrapper for DRM_IOCTL_I915_GEM_MMAP
https://bugs.kde.org/show_bug.cgi?id=413119 --- Comment #1 from Simon Richter --- Created attachment 123298 --> https://bugs.kde.org/attachment.cgi?id=123298=edit Patch to update ioctl argument definition -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 413119] New: ioctl wrapper for DRM_IOCTL_I915_GEM_MMAP
https://bugs.kde.org/show_bug.cgi?id=413119 Bug ID: 413119 Summary: ioctl wrapper for DRM_IOCTL_I915_GEM_MMAP Product: valgrind Version: unspecified Platform: unspecified OS: Linux Status: REPORTED Severity: normal Priority: NOR Component: memcheck Assignee: jsew...@acm.org Reporter: simon.rich...@hogyros.de Target Milestone: --- SUMMARY This adds a wrapper for the DRM_IOCTL_I915_GEM_MMAP ioctl, which creates a mapping for a DRM buffer object in the client's address space. STEPS TO REPRODUCE 1. Run an OpenGL application with direct rendering enabled on an Intel GPU under valgrind OBSERVED RESULT Memory accesses to command and texture buffers are flagged as errors EXPECTED RESULT Command buffers can be written normally SOFTWARE/OS VERSIONS Linux: 4.19.37 ADDITIONAL INFORMATION There are two versions of this ioctl that are distinguished by different argument sizes. -- You are receiving this mail because: You are watching all bug changes.