MacKwnoX/NX2navi
1
0
Fork 0
mirror of https://github.com/tangerineturtle1/NX2navi.git synced 2026-11-15 19:19:58 -04:00
Code Issues Packages Projects Releases Wiki Activity

Note the module list from each image's header and GetModuleFileNameA, which takes 0.3 ms where it took 25 ms.

Browse Source 
The review of the ExitProcess note found that platform/win32.c plat_exit_note_modules, which runs on the window thread
at every focus loss, was quadratic in the number of modules: GetModuleInformation and GetModuleBaseNameA each walk the
loader's module list through ReadProcessMemory. Its probe measured 27.5 ms a note with 292 modules.

Each module's size now comes from SizeOfImage in its own image header and its name from GetModuleFileNameA, the part
after the last backslash. The header is read through ReadProcessMemory into a 1 KiB buffer rather than in place: a
module unloaded since EnumProcessModules then fails the read and is left out, where a plain read would fault. The
table's contents are the same.

Measured with a probe beside this change (283 modules loaded, scratchpad rv_fxexit_fix\modtime_rpm): 25.3 ms a note
before, 0.27 ms after, and both ways note the same 283 bases, sizes and names. tests/exit_tls_check.py, whose note
names the program's own module from that table, and the compile check of win32.c pass, with no new -Wall warnings.
The review's optional second note after the device is made is left out: vk_device.c create_device is another
branch's.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
chrg6
2026-11-13 18:35:00 -05:00
parent e4f10a5a02
commit 359591c237
1 changed files with 22 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

26
runtime/platform/win32.c
View File

@@ -1441,6 +1441,10 @@ void plat_exit_note_setup(PlatAppend *log, void (*queued)(const char **p, size_t
en_log = log;
}
// Each module's size comes from its own image header and its name from GetModuleFileNameA. GetModuleInformation and
// GetModuleBaseNameA each walk the whole module list through ReadProcessMemory, so a note took about 27 ms with 290
// modules, on the window thread at every focus loss; this takes about 0.3 ms. The header is read through
// ReadProcessMemory too, which fails for a module unloaded since EnumProcessModules where a plain read would fault.
void plat_exit_note_modules(void) {
HMODULE m[EN_MODS];
DWORD need = 0;
@@ -1449,11 +1453,25 @@ void plat_exit_note_modules(void) {
plat_mutex_lock(&en_lock);
int t = InterlockedCompareExchange(&en_cur, 0, 0) == 0 ? 1 : 0; // the table the callback does not read
for (int i = 0; i < n; i++) {
MODULEINFO mi;
if (!GetModuleInformation(GetCurrentProcess(), m[i], &mi, sizeof mi)) continue;
unsigned char hdr[1024];
SIZE_T got = 0;
if (!ReadProcessMemory(GetCurrentProcess(), m[i], hdr, sizeof hdr, &got) || got != sizeof hdr) continue;
const IMAGE_DOS_HEADER *dos = (const IMAGE_DOS_HEADER *)hdr;
if (dos->e_magic != IMAGE_DOS_SIGNATURE || dos->e_lfanew < 0 ||
(size_t)dos->e_lfanew + sizeof(IMAGE_NT_HEADERS) > sizeof hdr)
continue;
const IMAGE_NT_HEADERS *nt = (const IMAGE_NT_HEADERS *)(hdr + dos->e_lfanew);
if (nt->Signature != IMAGE_NT_SIGNATURE) continue;
char path[MAX_PATH];
DWORD len = GetModuleFileNameA(m[i], path, sizeof path);
if (len >= sizeof path) len = sizeof path - 1;
path[len] = 0;
const char *name = path;
for (DWORD j = 0; j < len; j++)
if (path[j] == '\\' || path[j] == '/') name = path + j + 1;
EnMod *e = &en_mods[t][k++];
e->base = (uintptr_t)mi.lpBaseOfDll, e->size = mi.SizeOfImage;
if (!GetModuleBaseNameA(GetCurrentProcess(), m[i], e->name, sizeof e->name)) e->name[0] = 0;
e->base = (uintptr_t)m[i], e->size = nt->OptionalHeader.SizeOfImage;
snprintf(e->name, sizeof e->name, "%s", name);
}
en_n[t] = k;
InterlockedExchange(&en_cur, t);