tequilaOS/platform_bionic
2
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

riscv64: fix ifuncs, improve the ifunc tests.

Browse source 
Talking futher to the person doing the glibc risc-v ifunc work, they
clarified that glibc _is_ passing hwcap as the first argument, and the
null pointer is actually the second argument.

https://sourceware.org/pipermail/libc-alpha/2023-August/150967.html

So since our whole purpose here was source compatibility, let's do what
they're actually doing, and let's add some tests. I've also added a test
that __riscv_hwprobe() works from an ifunc resolver because that's one
place where it might well be used. That said, one other thing that came
out of the discussion is that I actually went away and looked at a
sample of top apps to see how many are using ifuncs currently. The
result? Zero. So although this _might_ be interesting long term
(especially if clang gets riscv64 FMV), I think we've done more than we
need to with riscv64 ifuncs for now!

Test: ran locally, both dynamic and static tests
Change-Id: Ie2044d9f4e47c32c00ad381f045c537f4df38b08
This commit is contained in:
Elliott Hughes 2023-08-07 12:12:36 -07:00
parent e722c756a5
commit 620a72252e
2 changed files with 32 additions and 6 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

12
libc/bionic/bionic_call_ifunc_resolver.cpp
View file

@ -58,12 +58,12 @@ ElfW(Addr) __bionic_call_ifunc_resolver(ElfW(Addr) resolver_addr) {
}
return reinterpret_cast<ifunc_resolver_t>(resolver_addr)(hwcap);
#elif defined(__riscv)
// This argument and its value is just a placeholder for now,
// but it means that if we do pass something in future (such as
// getauxval() and/or hwprobe key/value pairs), callees will be able to
// recognize what they're being given.
typedef ElfW(Addr) (*ifunc_resolver_t)(void*);
return reinterpret_cast<ifunc_resolver_t>(resolver_addr)(nullptr);
// The pointer argument is currently unused, but reserved for future
// expansion. If we pass nullptr from the beginning, it'll be easier
// to recognize if/when we pass actual data (and matches glibc).
typedef ElfW(Addr) (*ifunc_resolver_t)(uint64_t, void*);
static uint64_t hwcap = getauxval(AT_HWCAP);
return reinterpret_cast<ifunc_resolver_t>(resolver_addr)(hwcap, nullptr);
#else
typedef ElfW(Addr) (*ifunc_resolver_t)(void);
return reinterpret_cast<ifunc_resolver_t>(resolver_addr)();

26
tests/ifunc_test.cpp
View file

@ -60,6 +60,26 @@ extern "C" fn_ptr_t hwcap_resolver(unsigned long hwcap) {
return ret42;
}
#elif defined(__riscv)
#include <sys/hwprobe.h>
static uint64_t g_hwcap;
static riscv_hwprobe g_hwprobes[] = {{.key = RISCV_HWPROBE_KEY_IMA_EXT_0}};
extern "C" fn_ptr_t hwcap_resolver(uint64_t hwcap, void* null) {
// Check hwcap like arm32/arm64.
g_hwcap = hwcap;
// For now, the pointer argument is reserved for future expansion.
if (null != NULL) abort();
// Ensure that __riscv_hwprobe() can be called from an ifunc.
if (__riscv_hwprobe(g_hwprobes, 1, 0, nullptr, 0) != 0) return nullptr;
return ret42;
}
#else
extern "C" fn_ptr_t hwcap_resolver() {
@ -81,6 +101,12 @@ TEST(ifunc, hwcap) {
EXPECT_EQ(getauxval(AT_HWCAP2), g_arg._hwcap2);
#elif defined(__arm__)
EXPECT_EQ(getauxval(AT_HWCAP), g_hwcap);
#elif defined(__riscv)
EXPECT_EQ(getauxval(AT_HWCAP), g_hwcap);
riscv_hwprobe probes[] = {{.key = RISCV_HWPROBE_KEY_IMA_EXT_0}};
ASSERT_EQ(0, __riscv_hwprobe(probes, 1, 0, nullptr, 0));
EXPECT_EQ(probes[0].value, g_hwprobes[0].value);
#endif
}