From 28320acf2c07319d88d61551394576d2b1a6f29f Mon Sep 17 00:00:00 2001
From: gookie <36005381+gookie-dev@users.noreply.github.com>
Date: Sun, 11 Oct 2026 13:59:58 +0200
Subject: [PATCH] perf(libkernel): read a loaded image's unwind tables once in
 sceKernelGetModuleInfoFromAddr

sceKernelGetModuleInfoFromAddr found the size of the image's .eh_frame
on every call by walking all of its CIE and FDE records, checking each
against every PT_LOAD header. The guest libc's unwinder calls it for
each frame it unwinds, so a title that unwinds often walked its whole
.eh_frame hundreds of times per second.

The tables read from a .eh_frame_hdr are now kept per header address
and reused while the image stays loaded. The cache is dropped whenever
dl_iterate_phdr reports that an object was unloaded (dlpi_subs), so an
address reused by another image is read again.

guest_module_info queries a module twice and expects the same tables,
then unloads it, loads it again and expects tables of the same size at
the same offset in the new image.

PPSA16827 (Harmony's Odyssey), title screen, over 500 calls per second
counted under gdb, which slows the title: the process used 9269 CPU
samples per 10 s, 53.7% of them in this walk, and 3306 with this change,
3.2% in the lookup. The title still runs at 27 FPS; its frame rate is
limited elsewhere.
---
 .../prx/libkernel/Module/src/ModuleInfo.cpp   | 23 +++++++++++++++++--
 core/libs/tests/GuestModuleInfo.cpp           | 11 +++++++++
 2 files changed, 32 insertions(+), 2 deletions(-)

diff --git a/core/libs/prx/libkernel/Module/src/ModuleInfo.cpp b/core/libs/prx/libkernel/Module/src/ModuleInfo.cpp
index 66dc83653..b83f9a11f 100644
--- a/core/libs/prx/libkernel/Module/src/ModuleInfo.cpp
+++ b/core/libs/prx/libkernel/Module/src/ModuleInfo.cpp
@@ -22,7 +22,9 @@
 #else
 #include <dlfcn.h>
 #include <link.h>
+#include <mutex>
 #include <unistd.h>
+#include <unordered_map>
 #endif
 
 extern "C" std::int32_t ModuleIdForImage_nid_no_patch(const void* native);
@@ -81,6 +83,24 @@ bool Contains(const dl_phdr_info& image, std::uintptr_t begin, std::uint64_t siz
     return false;
 }
 
+struct FrameTableCache {
+    std::mutex mutex;
+    unsigned long long unloads = 0;
+    std::unordered_map<std::uintptr_t, EhFrame::Tables> tables;
+};
+
+EhFrame::Tables FrameTables(const dl_phdr_info& image, std::uintptr_t header) {
+    static auto& cache = *new FrameTableCache();
+    std::lock_guard lock(cache.mutex);
+    if (cache.unloads != image.dlpi_subs) {
+        cache.tables.clear();
+        cache.unloads = image.dlpi_subs;
+    }
+    if (const auto found = cache.tables.find(header); found != cache.tables.end()) return found->second;
+    const auto contains = [&](std::uintptr_t begin, std::uint64_t size) { return Contains(image, begin, size); };
+    return cache.tables.emplace(header, EhFrame::ReadTables(contains, header, Caller)).first->second;
+}
+
 std::string ImageName(const dl_phdr_info& image) {
     std::string path = image.dlpi_name ? image.dlpi_name : "";
     if (path.empty()) {
@@ -111,8 +131,7 @@ void Fill(const dl_phdr_info& image, TInfo& info) {
                 info.tls_size = ToU32(header.p_memsz, "TLS size");
                 info.tls_align = ToU32(header.p_align, "TLS alignment");
             } else if (header.p_type == PT_GNU_EH_FRAME) {
-                const auto contains = [&](std::uintptr_t begin, std::uint64_t size) { return Contains(image, begin, size); };
-                const auto tables = EhFrame::ReadTables(contains, address, Caller);
+                const auto tables = FrameTables(image, address);
                 info.eh_frame_hdr_addr = address;
                 info.eh_frame_hdr_size = ToU32(header.p_memsz, "eh_frame_hdr size");
                 info.eh_frame_addr = tables.frames;
diff --git a/core/libs/tests/GuestModuleInfo.cpp b/core/libs/tests/GuestModuleInfo.cpp
index b63e20a04..40eacdf81 100644
--- a/core/libs/tests/GuestModuleInfo.cpp
+++ b/core/libs/tests/GuestModuleInfo.cpp
@@ -114,5 +114,16 @@ int main(int argc, char** argv) {
     ModuleInfoEx unset{};
     unset.st_size = 0xf0000;
     Require(sceKernelGetModuleInfoFromAddr(reinterpret_cast<std::uintptr_t>(add), 2, &unset) == 0 && unset.st_size == sizeof(ModuleInfoEx));
+    const auto again = Query(add, 0);
+    Require(again.eh_frame_hdr_addr == info.eh_frame_hdr_addr && again.eh_frame_hdr_size == info.eh_frame_hdr_size);
+    Require(again.eh_frame_addr == info.eh_frame_addr && again.eh_frame_size == info.eh_frame_size);
     Require(dlclose_nid_postfix(module) == 0);
+    void* reloaded = dlopen_nid_postfix(argv[1], 2);
+    Require(reloaded != nullptr);
+    const void* reloadedAdd = dlsym_nid_postfix(reloaded, "GuestModuleAdd");
+    Require(reloadedAdd != nullptr);
+    const auto fresh = Query(reloadedAdd, 0);
+    Require(fresh.eh_frame_size == info.eh_frame_size && fresh.eh_frame_hdr_size == info.eh_frame_hdr_size);
+    Require(fresh.eh_frame_addr - fresh.segments[0].address == info.eh_frame_addr - info.segments[0].address);
+    Require(dlclose_nid_postfix(reloaded) == 0);
 }