* [PULL 01/14] linux-user: Fix setreuid and setregid to use direct syscalls
2024-11-15 20:58 [PULL 00/14] tcg + linux-user patch queue Richard Henderson
@ 2024-11-15 20:58 ` Richard Henderson
2024-11-15 20:58 ` [PULL 02/14] accel/tcg: Fix user-only probe_access_internal plugin check Richard Henderson
` (13 subsequent siblings)
14 siblings, 0 replies; 17+ messages in thread
From: Richard Henderson @ 2024-11-15 20:58 UTC (permalink / raw)
To: qemu-devel; +Cc: Helge Deller, qemu-stable, Helge Deller, Ilya Leoshkevich
From: Helge Deller <deller@kernel.org>
The commit fd6f7798ac30 ("linux-user: Use direct syscalls for setuid(),
etc") added direct syscall wrappers for setuid(), setgid(), etc since the
system calls have different semantics than the libc functions.
Add and use the corresponding wrappers for setreuid and setregid which
were missed in that commit.
This fixes the build of the debian package of the uid_wrapper library
(https://cwrap.org/uid_wrapper.html) when running linux-user.
Cc: qemu-stable@nongnu.org
Signed-off-by: Helge Deller <deller@gmx.de>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Ilya Leoshkevich <iii@linux.ibm.com>
Message-ID: <Zyo2jMKqq8hG8Pkz@p100>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
linux-user/syscall.c | 20 ++++++++++++++++----
1 file changed, 16 insertions(+), 4 deletions(-)
diff --git a/linux-user/syscall.c b/linux-user/syscall.c
index 59b2080b98..0279f23576 100644
--- a/linux-user/syscall.c
+++ b/linux-user/syscall.c
@@ -7233,12 +7233,24 @@ static inline int tswapid(int id)
#else
#define __NR_sys_setgroups __NR_setgroups
#endif
+#ifdef __NR_sys_setreuid32
+#define __NR_sys_setreuid __NR_setreuid32
+#else
+#define __NR_sys_setreuid __NR_setreuid
+#endif
+#ifdef __NR_sys_setregid32
+#define __NR_sys_setregid __NR_setregid32
+#else
+#define __NR_sys_setregid __NR_setregid
+#endif
_syscall1(int, sys_setuid, uid_t, uid)
_syscall1(int, sys_setgid, gid_t, gid)
_syscall3(int, sys_setresuid, uid_t, ruid, uid_t, euid, uid_t, suid)
_syscall3(int, sys_setresgid, gid_t, rgid, gid_t, egid, gid_t, sgid)
_syscall2(int, sys_setgroups, int, size, gid_t *, grouplist)
+_syscall2(int, sys_setreuid, uid_t, ruid, uid_t, euid);
+_syscall2(int, sys_setregid, gid_t, rgid, gid_t, egid);
void syscall_init(void)
{
@@ -11932,9 +11944,9 @@ static abi_long do_syscall1(CPUArchState *cpu_env, int num, abi_long arg1,
return get_errno(high2lowgid(getegid()));
#endif
case TARGET_NR_setreuid:
- return get_errno(setreuid(low2highuid(arg1), low2highuid(arg2)));
+ return get_errno(sys_setreuid(low2highuid(arg1), low2highuid(arg2)));
case TARGET_NR_setregid:
- return get_errno(setregid(low2highgid(arg1), low2highgid(arg2)));
+ return get_errno(sys_setregid(low2highgid(arg1), low2highgid(arg2)));
case TARGET_NR_getgroups:
{ /* the same code as for TARGET_NR_getgroups32 */
int gidsetsize = arg1;
@@ -12264,11 +12276,11 @@ static abi_long do_syscall1(CPUArchState *cpu_env, int num, abi_long arg1,
#endif
#ifdef TARGET_NR_setreuid32
case TARGET_NR_setreuid32:
- return get_errno(setreuid(arg1, arg2));
+ return get_errno(sys_setreuid(arg1, arg2));
#endif
#ifdef TARGET_NR_setregid32
case TARGET_NR_setregid32:
- return get_errno(setregid(arg1, arg2));
+ return get_errno(sys_setregid(arg1, arg2));
#endif
#ifdef TARGET_NR_getgroups32
case TARGET_NR_getgroups32:
--
2.43.0
^ permalink raw reply related [flat|nested] 17+ messages in thread* [PULL 02/14] accel/tcg: Fix user-only probe_access_internal plugin check
2024-11-15 20:58 [PULL 00/14] tcg + linux-user patch queue Richard Henderson
2024-11-15 20:58 ` [PULL 01/14] linux-user: Fix setreuid and setregid to use direct syscalls Richard Henderson
@ 2024-11-15 20:58 ` Richard Henderson
2024-11-15 20:58 ` [PULL 03/14] linux-user: Tolerate CONFIG_LSM_MMAP_MIN_ADDR Richard Henderson
` (12 subsequent siblings)
14 siblings, 0 replies; 17+ messages in thread
From: Richard Henderson @ 2024-11-15 20:58 UTC (permalink / raw)
To: qemu-devel; +Cc: qemu-stable, Alex Bennée
The acc_flag check for write should have been against PAGE_WRITE_ORG,
not PAGE_WRITE. But it is better to combine two acc_flag checks
to a single check against access_type. This matches the system code
in cputlb.c.
Cc: qemu-stable@nongnu.org
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/2647
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: 20241111145002.144995-1-richard.henderson@linaro.org
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
---
accel/tcg/user-exec.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/accel/tcg/user-exec.c b/accel/tcg/user-exec.c
index aa8af52cc3..06016eb030 100644
--- a/accel/tcg/user-exec.c
+++ b/accel/tcg/user-exec.c
@@ -800,7 +800,7 @@ static int probe_access_internal(CPUArchState *env, vaddr addr,
if (guest_addr_valid_untagged(addr)) {
int page_flags = page_get_flags(addr);
if (page_flags & acc_flag) {
- if ((acc_flag == PAGE_READ || acc_flag == PAGE_WRITE)
+ if (access_type != MMU_INST_FETCH
&& cpu_plugin_mem_cbs_enabled(env_cpu(env))) {
return TLB_MMIO;
}
--
2.43.0
^ permalink raw reply related [flat|nested] 17+ messages in thread* [PULL 03/14] linux-user: Tolerate CONFIG_LSM_MMAP_MIN_ADDR
2024-11-15 20:58 [PULL 00/14] tcg + linux-user patch queue Richard Henderson
2024-11-15 20:58 ` [PULL 01/14] linux-user: Fix setreuid and setregid to use direct syscalls Richard Henderson
2024-11-15 20:58 ` [PULL 02/14] accel/tcg: Fix user-only probe_access_internal plugin check Richard Henderson
@ 2024-11-15 20:58 ` Richard Henderson
2024-11-15 20:58 ` [PULL 04/14] tests/tcg: Test that sigreturn() does not corrupt the signal mask Richard Henderson
` (11 subsequent siblings)
14 siblings, 0 replies; 17+ messages in thread
From: Richard Henderson @ 2024-11-15 20:58 UTC (permalink / raw)
To: qemu-devel; +Cc: Ilya Leoshkevich, qemu-stable
From: Ilya Leoshkevich <iii@linux.ibm.com>
Running qemu-i386 on a system running with SELinux in enforcing mode
(more precisely: s390x trixie container on Fedora 40) fails with:
qemu-i386: tests/tcg/i386-linux-user/sigreturn-sigmask: Unable to find a guest_base to satisfy all guest address mapping requirements
00000000-ffffffff
The reason is that main() determines mmap_min_addr from
/proc/sys/vm/mmap_min_addr, but SELinux additionally defines
CONFIG_LSM_MMAP_MIN_ADDR, which is normally larger: 32K or 64K, but,
in general, can be anything. There is no portable way to query its
value: /boot/config, /proc/config and /proc/config.gz are distro- and
environment-specific.
Once the identity map fails, the magnitude of guest_base does not
matter, so fix by starting the search from 1M or 1G.
Cc: qemu-stable@nongnu.org
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/2598
Suggested-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Message-ID: <20241023002558.34589-1-iii@linux.ibm.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
linux-user/elfload.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/linux-user/elfload.c b/linux-user/elfload.c
index 6cef8db3b5..d6ad77d27d 100644
--- a/linux-user/elfload.c
+++ b/linux-user/elfload.c
@@ -2898,7 +2898,7 @@ static uintptr_t pgb_try_itree(const PGBAddrs *ga, uintptr_t base,
static uintptr_t pgb_find_itree(const PGBAddrs *ga, IntervalTreeRoot *root,
uintptr_t align, uintptr_t brk)
{
- uintptr_t last = mmap_min_addr;
+ uintptr_t last = sizeof(uintptr_t) == 4 ? MiB : GiB;
uintptr_t base, skip;
while (true) {
--
2.43.0
^ permalink raw reply related [flat|nested] 17+ messages in thread* [PULL 04/14] tests/tcg: Test that sigreturn() does not corrupt the signal mask
2024-11-15 20:58 [PULL 00/14] tcg + linux-user patch queue Richard Henderson
` (2 preceding siblings ...)
2024-11-15 20:58 ` [PULL 03/14] linux-user: Tolerate CONFIG_LSM_MMAP_MIN_ADDR Richard Henderson
@ 2024-11-15 20:58 ` Richard Henderson
2024-11-15 20:58 ` [PULL 05/14] target/i386: fix hang when using slow path for ptw_setl Richard Henderson
` (10 subsequent siblings)
14 siblings, 0 replies; 17+ messages in thread
From: Richard Henderson @ 2024-11-15 20:58 UTC (permalink / raw)
To: qemu-devel; +Cc: Ilya Leoshkevich
From: Ilya Leoshkevich <iii@linux.ibm.com>
Add a small test to prevent regressions.
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Message-ID: <20241108145237.37377-2-iii@linux.ibm.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
tests/tcg/multiarch/sigreturn-sigmask.c | 51 +++++++++++++++++++++++++
tests/tcg/multiarch/Makefile.target | 3 ++
2 files changed, 54 insertions(+)
create mode 100644 tests/tcg/multiarch/sigreturn-sigmask.c
diff --git a/tests/tcg/multiarch/sigreturn-sigmask.c b/tests/tcg/multiarch/sigreturn-sigmask.c
new file mode 100644
index 0000000000..e6cc904898
--- /dev/null
+++ b/tests/tcg/multiarch/sigreturn-sigmask.c
@@ -0,0 +1,51 @@
+/*
+ * Test that sigreturn() does not corrupt the signal mask.
+ * Block SIGUSR2 and handle SIGUSR1.
+ * Then sigwait() SIGUSR2, which relies on it remaining blocked.
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ */
+#include <assert.h>
+#include <pthread.h>
+#include <signal.h>
+#include <stdlib.h>
+#include <unistd.h>
+
+int seen_sig = -1;
+
+static void signal_func(int sig)
+{
+ seen_sig = sig;
+}
+
+static void *thread_func(void *arg)
+{
+ kill(getpid(), SIGUSR2);
+ return NULL;
+}
+
+int main(void)
+{
+ struct sigaction act = {
+ .sa_handler = signal_func,
+ };
+ pthread_t thread;
+ sigset_t set;
+ int sig;
+
+ assert(sigaction(SIGUSR1, &act, NULL) == 0);
+
+ assert(sigemptyset(&set) == 0);
+ assert(sigaddset(&set, SIGUSR2) == 0);
+ assert(sigprocmask(SIG_BLOCK, &set, NULL) == 0);
+
+ kill(getpid(), SIGUSR1);
+ assert(seen_sig == SIGUSR1);
+
+ assert(pthread_create(&thread, NULL, thread_func, NULL) == 0);
+ assert(sigwait(&set, &sig) == 0);
+ assert(sig == SIGUSR2);
+ assert(pthread_join(thread, NULL) == 0);
+
+ return EXIT_SUCCESS;
+}
diff --git a/tests/tcg/multiarch/Makefile.target b/tests/tcg/multiarch/Makefile.target
index 78b83d5575..18d3cf4ae0 100644
--- a/tests/tcg/multiarch/Makefile.target
+++ b/tests/tcg/multiarch/Makefile.target
@@ -42,6 +42,9 @@ munmap-pthread: LDFLAGS+=-pthread
vma-pthread: CFLAGS+=-pthread
vma-pthread: LDFLAGS+=-pthread
+sigreturn-sigmask: CFLAGS+=-pthread
+sigreturn-sigmask: LDFLAGS+=-pthread
+
# The vma-pthread seems very sensitive on gitlab and we currently
# don't know if its exposing a real bug or the test is flaky.
ifneq ($(GITLAB_CI),)
--
2.43.0
^ permalink raw reply related [flat|nested] 17+ messages in thread* [PULL 05/14] target/i386: fix hang when using slow path for ptw_setl
2024-11-15 20:58 [PULL 00/14] tcg + linux-user patch queue Richard Henderson
` (3 preceding siblings ...)
2024-11-15 20:58 ` [PULL 04/14] tests/tcg: Test that sigreturn() does not corrupt the signal mask Richard Henderson
@ 2024-11-15 20:58 ` Richard Henderson
2024-11-15 20:58 ` [PULL 06/14] cpu: ensure we don't call start_exclusive from cpu_exec Richard Henderson
` (9 subsequent siblings)
14 siblings, 0 replies; 17+ messages in thread
From: Richard Henderson @ 2024-11-15 20:58 UTC (permalink / raw)
To: qemu-devel; +Cc: Pierrick Bouvier
From: Pierrick Bouvier <pierrick.bouvier@linaro.org>
When instrumenting memory accesses for plugin, we force memory accesses
to use the slow path for mmu [1]. This create a situation where we end
up calling ptw_setl_slow. This was fixed recently in [2] but the issue
still could appear out of plugins use case.
Since this function gets called during a cpu_exec, start_exclusive then
hangs. This exclusive section was introduced initially for security
reasons [3].
I suspect this code path was never triggered, because ptw_setl_slow
would always be called transitively from cpu_exec, resulting in a hang.
[1] https://gitlab.com/qemu-project/qemu/-/commit/6d03226b42247b68ab2f0b3663e0f624335a4055
[2] https://gitlab.com/qemu-project/qemu/-/commit/115ade42d50144c15b74368d32dc734ea277d853
[3] https://gitlab.com/qemu-project/qemu/-/issues/279
Fixes: https://gitlab.com/qemu-project/qemu/-/issues/2566
Signed-off-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-ID: <20241025175857.2554252-2-pierrick.bouvier@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
target/i386/tcg/sysemu/excp_helper.c | 5 +++++
1 file changed, 5 insertions(+)
diff --git a/target/i386/tcg/sysemu/excp_helper.c b/target/i386/tcg/sysemu/excp_helper.c
index 02d3486421..b1f40040f8 100644
--- a/target/i386/tcg/sysemu/excp_helper.c
+++ b/target/i386/tcg/sysemu/excp_helper.c
@@ -107,6 +107,10 @@ static bool ptw_setl_slow(const PTETranslate *in, uint32_t old, uint32_t new)
{
uint32_t cmp;
+ CPUState *cpu = env_cpu(in->env);
+ /* We are in cpu_exec, and start_exclusive can't be called directly.*/
+ g_assert(cpu->running);
+ cpu_exec_end(cpu);
/* Does x86 really perform a rmw cycle on mmio for ptw? */
start_exclusive();
cmp = cpu_ldl_mmuidx_ra(in->env, in->gaddr, in->ptw_idx, 0);
@@ -114,6 +118,7 @@ static bool ptw_setl_slow(const PTETranslate *in, uint32_t old, uint32_t new)
cpu_stl_mmuidx_ra(in->env, in->gaddr, new, in->ptw_idx, 0);
}
end_exclusive();
+ cpu_exec_start(cpu);
return cmp == old;
}
--
2.43.0
^ permalink raw reply related [flat|nested] 17+ messages in thread* [PULL 06/14] cpu: ensure we don't call start_exclusive from cpu_exec
2024-11-15 20:58 [PULL 00/14] tcg + linux-user patch queue Richard Henderson
` (4 preceding siblings ...)
2024-11-15 20:58 ` [PULL 05/14] target/i386: fix hang when using slow path for ptw_setl Richard Henderson
@ 2024-11-15 20:58 ` Richard Henderson
2024-11-15 20:58 ` [PULL 07/14] linux-user: Honor elf alignment when placing images Richard Henderson
` (8 subsequent siblings)
14 siblings, 0 replies; 17+ messages in thread
From: Richard Henderson @ 2024-11-15 20:58 UTC (permalink / raw)
To: qemu-devel; +Cc: Pierrick Bouvier, Philippe Mathieu-Daudé
From: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-ID: <20241025175857.2554252-3-pierrick.bouvier@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
cpu-common.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/cpu-common.c b/cpu-common.c
index 6b262233a3..0d607bbe49 100644
--- a/cpu-common.c
+++ b/cpu-common.c
@@ -194,6 +194,9 @@ void start_exclusive(void)
CPUState *other_cpu;
int running_cpus;
+ /* Ensure we are not running, or start_exclusive will be blocked. */
+ g_assert(!current_cpu->running);
+
if (current_cpu->exclusive_context_count) {
current_cpu->exclusive_context_count++;
return;
--
2.43.0
^ permalink raw reply related [flat|nested] 17+ messages in thread* [PULL 07/14] linux-user: Honor elf alignment when placing images
2024-11-15 20:58 [PULL 00/14] tcg + linux-user patch queue Richard Henderson
` (5 preceding siblings ...)
2024-11-15 20:58 ` [PULL 06/14] cpu: ensure we don't call start_exclusive from cpu_exec Richard Henderson
@ 2024-11-15 20:58 ` Richard Henderson
2024-11-15 20:58 ` [PULL 08/14] linux-user: Drop image_info.alignment Richard Henderson
` (7 subsequent siblings)
14 siblings, 0 replies; 17+ messages in thread
From: Richard Henderson @ 2024-11-15 20:58 UTC (permalink / raw)
To: qemu-devel
Most binaries don't actually depend on more than page alignment,
but any binary can request it. Not honoring this was a bug.
This became obvious when gdb reported
Failed to read a valid object file image from memory
when examining some vdso which are marked as needing more
than page alignment.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
linux-user/elfload.c | 35 ++++++++++++++++++++++++++++-------
1 file changed, 28 insertions(+), 7 deletions(-)
diff --git a/linux-user/elfload.c b/linux-user/elfload.c
index d6ad77d27d..90e79a01b4 100644
--- a/linux-user/elfload.c
+++ b/linux-user/elfload.c
@@ -3179,7 +3179,8 @@ static void load_elf_image(const char *image_name, const ImageSource *src,
char **pinterp_name)
{
g_autofree struct elf_phdr *phdr = NULL;
- abi_ulong load_addr, load_bias, loaddr, hiaddr, error;
+ abi_ulong load_addr, load_bias, loaddr, hiaddr, error, align;
+ size_t reserve_size, align_size;
int i, prot_exec;
Error *err = NULL;
@@ -3263,6 +3264,9 @@ static void load_elf_image(const char *image_name, const ImageSource *src,
load_addr = loaddr;
+ align = pow2ceil(info->alignment);
+ info->alignment = align;
+
if (pinterp_name != NULL) {
if (ehdr->e_type == ET_EXEC) {
/*
@@ -3271,8 +3275,6 @@ static void load_elf_image(const char *image_name, const ImageSource *src,
*/
probe_guest_base(image_name, loaddr, hiaddr);
} else {
- abi_ulong align;
-
/*
* The binary is dynamic, but we still need to
* select guest_base. In this case we pass a size.
@@ -3290,10 +3292,7 @@ static void load_elf_image(const char *image_name, const ImageSource *src,
* Since we do not have complete control over the guest
* address space, we prefer the kernel to choose some address
* rather than force the use of LOAD_ADDR via MAP_FIXED.
- * But without MAP_FIXED we cannot guarantee alignment,
- * only suggest it.
*/
- align = pow2ceil(info->alignment);
if (align) {
load_addr &= -align;
}
@@ -3317,13 +3316,35 @@ static void load_elf_image(const char *image_name, const ImageSource *src,
* In both cases, we will overwrite pages in this range with mappings
* from the executable.
*/
- load_addr = target_mmap(load_addr, (size_t)hiaddr - loaddr + 1, PROT_NONE,
+ reserve_size = (size_t)hiaddr - loaddr + 1;
+ align_size = reserve_size;
+
+ if (ehdr->e_type != ET_EXEC && align > qemu_real_host_page_size()) {
+ align_size += align - 1;
+ }
+
+ load_addr = target_mmap(load_addr, align_size, PROT_NONE,
MAP_PRIVATE | MAP_ANON | MAP_NORESERVE |
(ehdr->e_type == ET_EXEC ? MAP_FIXED_NOREPLACE : 0),
-1, 0);
if (load_addr == -1) {
goto exit_mmap;
}
+
+ if (align_size != reserve_size) {
+ abi_ulong align_addr = ROUND_UP(load_addr, align);
+ abi_ulong align_end = align_addr + reserve_size;
+ abi_ulong load_end = load_addr + align_size;
+
+ if (align_addr != load_addr) {
+ target_munmap(load_addr, align_addr - load_addr);
+ }
+ if (align_end != load_end) {
+ target_munmap(align_end, load_end - align_end);
+ }
+ load_addr = align_addr;
+ }
+
load_bias = load_addr - loaddr;
if (elf_is_fdpic(ehdr)) {
--
2.43.0
^ permalink raw reply related [flat|nested] 17+ messages in thread* [PULL 08/14] linux-user: Drop image_info.alignment
2024-11-15 20:58 [PULL 00/14] tcg + linux-user patch queue Richard Henderson
` (6 preceding siblings ...)
2024-11-15 20:58 ` [PULL 07/14] linux-user: Honor elf alignment when placing images Richard Henderson
@ 2024-11-15 20:58 ` Richard Henderson
2024-11-15 20:58 ` [PULL 09/14] linux-user/aarch64: Reduce vdso alignment to 4k Richard Henderson
` (6 subsequent siblings)
14 siblings, 0 replies; 17+ messages in thread
From: Richard Henderson @ 2024-11-15 20:58 UTC (permalink / raw)
To: qemu-devel; +Cc: Philippe Mathieu-Daudé
This field is write-only. Use only the function-local
variable within load_elf_image.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
linux-user/qemu.h | 1 -
linux-user/elfload.c | 7 +++----
2 files changed, 3 insertions(+), 5 deletions(-)
diff --git a/linux-user/qemu.h b/linux-user/qemu.h
index 895bdd722a..67bc81b149 100644
--- a/linux-user/qemu.h
+++ b/linux-user/qemu.h
@@ -44,7 +44,6 @@ struct image_info {
abi_ulong file_string;
uint32_t elf_flags;
int personality;
- abi_ulong alignment;
bool exec_stack;
/* Generic semihosting knows about these pointers. */
diff --git a/linux-user/elfload.c b/linux-user/elfload.c
index 90e79a01b4..ef9cffbe4a 100644
--- a/linux-user/elfload.c
+++ b/linux-user/elfload.c
@@ -3220,7 +3220,7 @@ static void load_elf_image(const char *image_name, const ImageSource *src,
* amount of memory to handle that. Locate the interpreter, if any.
*/
loaddr = -1, hiaddr = 0;
- info->alignment = 0;
+ align = 0;
info->exec_stack = EXSTACK_DEFAULT;
for (i = 0; i < ehdr->e_phnum; ++i) {
struct elf_phdr *eppnt = phdr + i;
@@ -3234,7 +3234,7 @@ static void load_elf_image(const char *image_name, const ImageSource *src,
hiaddr = a;
}
++info->nsegs;
- info->alignment |= eppnt->p_align;
+ align |= eppnt->p_align;
} else if (eppnt->p_type == PT_INTERP && pinterp_name) {
g_autofree char *interp_name = NULL;
@@ -3264,8 +3264,7 @@ static void load_elf_image(const char *image_name, const ImageSource *src,
load_addr = loaddr;
- align = pow2ceil(info->alignment);
- info->alignment = align;
+ align = pow2ceil(align);
if (pinterp_name != NULL) {
if (ehdr->e_type == ET_EXEC) {
--
2.43.0
^ permalink raw reply related [flat|nested] 17+ messages in thread* [PULL 09/14] linux-user/aarch64: Reduce vdso alignment to 4k
2024-11-15 20:58 [PULL 00/14] tcg + linux-user patch queue Richard Henderson
` (7 preceding siblings ...)
2024-11-15 20:58 ` [PULL 08/14] linux-user: Drop image_info.alignment Richard Henderson
@ 2024-11-15 20:58 ` Richard Henderson
2024-11-15 20:58 ` [PULL 10/14] linux-user/arm: " Richard Henderson
` (5 subsequent siblings)
14 siblings, 0 replies; 17+ messages in thread
From: Richard Henderson @ 2024-11-15 20:58 UTC (permalink / raw)
To: qemu-devel; +Cc: Philippe Mathieu-Daudé
Reduce vdso alignment to minimum page size.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
linux-user/aarch64/Makefile.vdso | 5 +++--
linux-user/aarch64/vdso-be.so | Bin 3224 -> 3224 bytes
linux-user/aarch64/vdso-le.so | Bin 3224 -> 3224 bytes
3 files changed, 3 insertions(+), 2 deletions(-)
diff --git a/linux-user/aarch64/Makefile.vdso b/linux-user/aarch64/Makefile.vdso
index 599958116b..c33a679c0f 100644
--- a/linux-user/aarch64/Makefile.vdso
+++ b/linux-user/aarch64/Makefile.vdso
@@ -5,8 +5,9 @@ VPATH += $(SUBDIR)
all: $(SUBDIR)/vdso-be.so $(SUBDIR)/vdso-le.so
-LDFLAGS = -nostdlib -shared -Wl,-h,linux-vdso.so.1 -Wl,--build-id=sha1 \
- -Wl,--hash-style=both -Wl,-T,$(SUBDIR)/vdso.ld
+LDFLAGS = -nostdlib -shared -Wl,-h,linux-vdso.so.1 \
+ -Wl,--build-id=sha1 -Wl,--hash-style=both \
+ -Wl,-z,max-page-size=4096 -Wl,-T,$(SUBDIR)/vdso.ld
$(SUBDIR)/vdso-be.so: vdso.S vdso.ld
$(CC) -o $@ $(LDFLAGS) -mbig-endian $<
diff --git a/linux-user/aarch64/vdso-be.so b/linux-user/aarch64/vdso-be.so
index 808206ade824b09d786f6cc34f7cddf80b63130e..d43c3b19cdf6588757f2039f2308a8bce21aed9c 100755
GIT binary patch
delta 50
zcmV-20L}lH8JHQ6tpWfLk+0bl<zcO@sx8Bk2RT5)R`<=5QB@Vnvz!6X1qOFSFiLP%
IvvCI}1uUl&oB#j-
delta 50
zcmV-20L}lH8JHQ6tpWi6k+0blA5kqX4aFpe)BQ2Nwg4c8+P|Ifvz!6X1qO9#Ib$<7
IvvCI}1^-?Xy#N3J
diff --git a/linux-user/aarch64/vdso-le.so b/linux-user/aarch64/vdso-le.so
index 941aaf29931193300de1f62097867c282a7e0c74..aaedc9d85e568d5dadc9c394e9c4459f1f0c8c88 100755
GIT binary patch
delta 49
zcmV-10M7rI8JHQ6sR9rHk*nPlH6|OK!ah4~>4vUoZ1rE*XtsXbvz!6Y1qE4ha#lpM
HaR(;_22&J+
delta 49
zcmV-10M7rI8JHQ6sR952k*nPl^~i=KtbcHFtl42<9TUsw2PpfHvz!6Y1qD@cX=ik^
HaR(;_8o3m&
--
2.43.0
^ permalink raw reply related [flat|nested] 17+ messages in thread* [PULL 10/14] linux-user/arm: Reduce vdso alignment to 4k
2024-11-15 20:58 [PULL 00/14] tcg + linux-user patch queue Richard Henderson
` (8 preceding siblings ...)
2024-11-15 20:58 ` [PULL 09/14] linux-user/aarch64: Reduce vdso alignment to 4k Richard Henderson
@ 2024-11-15 20:58 ` Richard Henderson
2024-11-15 20:58 ` [PULL 11/14] linux-user/loongarch64: " Richard Henderson
` (4 subsequent siblings)
14 siblings, 0 replies; 17+ messages in thread
From: Richard Henderson @ 2024-11-15 20:58 UTC (permalink / raw)
To: qemu-devel; +Cc: Philippe Mathieu-Daudé
Reduce vdso alignment to minimum page size.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
linux-user/arm/Makefile.vdso | 2 +-
linux-user/arm/vdso-be.so | Bin 2648 -> 2648 bytes
linux-user/arm/vdso-le.so | Bin 2648 -> 2648 bytes
3 files changed, 1 insertion(+), 1 deletion(-)
diff --git a/linux-user/arm/Makefile.vdso b/linux-user/arm/Makefile.vdso
index 2d098a5748..8a24b0e534 100644
--- a/linux-user/arm/Makefile.vdso
+++ b/linux-user/arm/Makefile.vdso
@@ -6,7 +6,7 @@ VPATH += $(SUBDIR)
all: $(SUBDIR)/vdso-be.so $(SUBDIR)/vdso-le.so
# Adding -use-blx disables unneeded interworking without actually using blx.
-LDFLAGS = -nostdlib -shared -Wl,-use-blx \
+LDFLAGS = -nostdlib -shared -Wl,-use-blx -Wl,-z,max-page-size=4096 \
-Wl,-h,linux-vdso.so.1 -Wl,--build-id=sha1 \
-Wl,--hash-style=both -Wl,-T,$(SUBDIR)/vdso.ld
diff --git a/linux-user/arm/vdso-be.so b/linux-user/arm/vdso-be.so
index 69cafbb956e283e2975bac59a10491c0cbafca57..bed02804a4bd367eb9fd8ca54d0c980103c02245 100755
GIT binary patch
delta 49
zcmV-10M7r|6xbAyaRLAkk#lhrGQ`BrU>NUTo0WUr&~YvSTwestvG{WZ25D?qS41(h
HaR%oD3DgvN
delta 49
zcmV-10M7r|6xbAyaRLDVk#lhrweyTc_Z*p@&&2@VLR1?$m|vtIvG{WZ23l}Oc}8xt
HaR%oD4;B=<
diff --git a/linux-user/arm/vdso-le.so b/linux-user/arm/vdso-le.so
index ad05a1251875ac0c76685e1f9190a7307a8444d1..38d3d51047372391b3125c3f9f6ea5401f04bba1 100755
GIT binary patch
delta 49
zcmV-10M7r|6xbAyaRLwkk#lhr8{q*880U1i=t7z4dQ|b*DDU}VvG{WZ25CxDZDwh+
HaR%oD5d#(3
delta 49
zcmV-10M7r|6xbAyaRLAVk#lhrkUMV7Jk`NO!6O#urC$6IB@6+uvG{WZ26R_8MpaO=
HaR%oD?$r}}
--
2.43.0
^ permalink raw reply related [flat|nested] 17+ messages in thread* [PULL 11/14] linux-user/loongarch64: Reduce vdso alignment to 4k
2024-11-15 20:58 [PULL 00/14] tcg + linux-user patch queue Richard Henderson
` (9 preceding siblings ...)
2024-11-15 20:58 ` [PULL 10/14] linux-user/arm: " Richard Henderson
@ 2024-11-15 20:58 ` Richard Henderson
2024-11-15 20:58 ` [PULL 12/14] linux-user/ppc: " Richard Henderson
` (3 subsequent siblings)
14 siblings, 0 replies; 17+ messages in thread
From: Richard Henderson @ 2024-11-15 20:58 UTC (permalink / raw)
To: qemu-devel; +Cc: Philippe Mathieu-Daudé
Reduce vdso alignment to minimum page size.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
linux-user/loongarch64/Makefile.vdso | 3 ++-
linux-user/loongarch64/vdso.so | Bin 3560 -> 3560 bytes
2 files changed, 2 insertions(+), 1 deletion(-)
diff --git a/linux-user/loongarch64/Makefile.vdso b/linux-user/loongarch64/Makefile.vdso
index 369de13344..1d760b1e47 100644
--- a/linux-user/loongarch64/Makefile.vdso
+++ b/linux-user/loongarch64/Makefile.vdso
@@ -8,4 +8,5 @@ all: $(SUBDIR)/vdso.so
$(SUBDIR)/vdso.so: vdso.S vdso.ld vdso-asmoffset.h
$(CC) -o $@ -nostdlib -shared -fpic -Wl,-h,linux-vdso.so.1 \
-Wl,--build-id=sha1 -Wl,--hash-style=both \
- -Wl,--no-warn-rwx-segments -Wl,-T,$(SUBDIR)/vdso.ld $<
+ -Wl,--no-warn-rwx-segments -Wl,-z,max-page-size=4096 \
+ -Wl,-T,$(SUBDIR)/vdso.ld $<
diff --git a/linux-user/loongarch64/vdso.so b/linux-user/loongarch64/vdso.so
index bfaa26f2bfe1aaa01d9a349b8b030ef6323e1f8e..7c2de6c50e706164225e82f652d4becc04c71ff0 100755
GIT binary patch
delta 37
tcmaDM{X%-eN=AW+tM-YA3hb&jk8@2<QS14Vzx}<pii5+a%{7cUJOC|W51jx2
delta 37
tcmaDM{X%-eN=Ao?tM-WmX?HPm)y|eS*?us!At};x;l^g2%{7cUJOC+;4-)_Y
--
2.43.0
^ permalink raw reply related [flat|nested] 17+ messages in thread* [PULL 12/14] linux-user/ppc: Reduce vdso alignment to 4k
2024-11-15 20:58 [PULL 00/14] tcg + linux-user patch queue Richard Henderson
` (10 preceding siblings ...)
2024-11-15 20:58 ` [PULL 11/14] linux-user/loongarch64: " Richard Henderson
@ 2024-11-15 20:58 ` Richard Henderson
2024-11-15 20:58 ` [PULL 13/14] linux-user/arm: Select vdso for be8 and be32 modes Richard Henderson
` (2 subsequent siblings)
14 siblings, 0 replies; 17+ messages in thread
From: Richard Henderson @ 2024-11-15 20:58 UTC (permalink / raw)
To: qemu-devel; +Cc: Philippe Mathieu-Daudé
Reduce vdso alignment to minimum page size.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
linux-user/ppc/Makefile.vdso | 6 ++++--
linux-user/ppc/vdso-32.so | Bin 3020 -> 3020 bytes
linux-user/ppc/vdso-64.so | Bin 3896 -> 3896 bytes
linux-user/ppc/vdso-64le.so | Bin 3896 -> 3896 bytes
4 files changed, 4 insertions(+), 2 deletions(-)
diff --git a/linux-user/ppc/Makefile.vdso b/linux-user/ppc/Makefile.vdso
index 3ca3c6b83e..e2b8facbb5 100644
--- a/linux-user/ppc/Makefile.vdso
+++ b/linux-user/ppc/Makefile.vdso
@@ -6,9 +6,11 @@ VPATH += $(SUBDIR)
all: $(SUBDIR)/vdso-32.so $(SUBDIR)/vdso-64.so $(SUBDIR)/vdso-64le.so
LDFLAGS32 = -nostdlib -shared -Wl,-T,$(SUBDIR)/vdso-32.ld \
- -Wl,-h,linux-vdso32.so.1 -Wl,--hash-style=both -Wl,--build-id=sha1
+ -Wl,-h,linux-vdso32.so.1 -Wl,--hash-style=both \
+ -Wl,--build-id=sha1 -Wl,-z,max-page-size=4096
LDFLAGS64 = -nostdlib -shared -Wl,-T,$(SUBDIR)/vdso-64.ld \
- -Wl,-h,linux-vdso64.so.1 -Wl,--hash-style=both -Wl,--build-id=sha1
+ -Wl,-h,linux-vdso64.so.1 -Wl,--hash-style=both \
+ -Wl,--build-id=sha1 -Wl,-z,max-page-size=4096
$(SUBDIR)/vdso-32.so: vdso.S vdso-32.ld vdso-asmoffset.h
$(CC) -o $@ $(LDFLAGS32) -m32 $<
diff --git a/linux-user/ppc/vdso-32.so b/linux-user/ppc/vdso-32.so
index b19baafb0d38e15b4a24def5c44a6d684714be45..0dc55e0dddff618b954dbb939335e99956daf64a 100755
GIT binary patch
delta 42
zcmV+_0M-A@7t9xsCINtvCrSbU5Rr3n6lq1YQP%3b&XBjV4sl%JXqmbIL$UbO3Tx34
A(f|Me
delta 42
xcmX>jenxzP8Y9C*buT6$SzIX6c=(tjbAN``v6&a|RP{JtzgDQRW#f-4TmVmR5>WsE
diff --git a/linux-user/ppc/vdso-64.so b/linux-user/ppc/vdso-64.so
index 913c831b3819fc09912b9b31f7fbe9ee311ae12f..ac1ab2582e4675979ffca3ce90dce17df579ab2a 100755
GIT binary patch
delta 38
wcmV+>0NMYz9=INmtpWfLk*~QFqejR%tq=sGFts+qF9Cf{%e>^#vwQ)(4KYR#WB>pF
delta 38
wcmV+>0NMYz9=INmtpWi6k*~QF<@TF=qRS8+wHM`Qf0n_&>m=ZivwQ)(4PLhrs{jB1
diff --git a/linux-user/ppc/vdso-64le.so b/linux-user/ppc/vdso-64le.so
index 258a03b807c4eca23547d978c16d1ad5ebd08bc5..424abb4290b7d3100e9dede2f3059483608ba703 100755
GIT binary patch
delta 38
wcmV+>0NMYz9=INmsR9rHk*mEFj|Vri9^_Z(nV0Nw;)4VN>*aZovwQ)(4Mvj@kpKVy
delta 38
wcmV+>0NMYz9=INmsR952k*mEF`$9%&`;)~jB!C2C?=itIoQ(CfvwQ)(4MbHDMgRZ+
--
2.43.0
^ permalink raw reply related [flat|nested] 17+ messages in thread* [PULL 13/14] linux-user/arm: Select vdso for be8 and be32 modes
2024-11-15 20:58 [PULL 00/14] tcg + linux-user patch queue Richard Henderson
` (11 preceding siblings ...)
2024-11-15 20:58 ` [PULL 12/14] linux-user/ppc: " Richard Henderson
@ 2024-11-15 20:58 ` Richard Henderson
2024-11-15 20:58 ` [PULL 14/14] tcg: Allow top bit of SIMD_DATA_BITS to be set in simd_desc() Richard Henderson
2024-11-16 10:39 ` [PULL 00/14] tcg + linux-user patch queue Peter Maydell
14 siblings, 0 replies; 17+ messages in thread
From: Richard Henderson @ 2024-11-15 20:58 UTC (permalink / raw)
To: qemu-devel; +Cc: Philippe Mathieu-Daudé
In be8 mode, instructions are little-endian.
In be32 mode, instructions are big-endian.
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/2333
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
linux-user/elfload.c | 31 +++++++++++++++++----
linux-user/arm/Makefile.vdso | 9 ++++--
linux-user/arm/meson.build | 13 +++++++--
linux-user/arm/vdso-be32.so | Bin 0 -> 2648 bytes
linux-user/arm/{vdso-be.so => vdso-be8.so} | Bin 2648 -> 2648 bytes
5 files changed, 41 insertions(+), 12 deletions(-)
create mode 100755 linux-user/arm/vdso-be32.so
rename linux-user/arm/{vdso-be.so => vdso-be8.so} (95%)
diff --git a/linux-user/elfload.c b/linux-user/elfload.c
index ef9cffbe4a..471a384b22 100644
--- a/linux-user/elfload.c
+++ b/linux-user/elfload.c
@@ -659,6 +659,23 @@ static const char *get_elf_platform(void)
#undef END
}
+#if TARGET_BIG_ENDIAN
+#include "elf.h"
+#include "vdso-be8.c.inc"
+#include "vdso-be32.c.inc"
+
+static const VdsoImageInfo *vdso_image_info(uint32_t elf_flags)
+{
+ return (EF_ARM_EABI_VERSION(elf_flags) >= EF_ARM_EABI_VER4
+ && (elf_flags & EF_ARM_BE8)
+ ? &vdso_be8_image_info
+ : &vdso_be32_image_info);
+}
+#define vdso_image_info vdso_image_info
+#else
+# define VDSO_HEADER "vdso-le.c.inc"
+#endif
+
#else
/* 64 bit ARM definitions */
@@ -958,14 +975,14 @@ const char *elf_hwcap2_str(uint32_t bit)
#undef GET_FEATURE_ID
-#endif /* not TARGET_AARCH64 */
-
#if TARGET_BIG_ENDIAN
# define VDSO_HEADER "vdso-be.c.inc"
#else
# define VDSO_HEADER "vdso-le.c.inc"
#endif
+#endif /* not TARGET_AARCH64 */
+
#endif /* TARGET_ARM */
#ifdef TARGET_SPARC
@@ -3524,12 +3541,14 @@ static void load_elf_interp(const char *filename, struct image_info *info,
load_elf_image(filename, &src, info, &ehdr, NULL);
}
+#ifndef vdso_image_info
#ifdef VDSO_HEADER
#include VDSO_HEADER
-#define vdso_image_info() &vdso_image_info
+#define vdso_image_info(flags) &vdso_image_info
#else
-#define vdso_image_info() NULL
-#endif
+#define vdso_image_info(flags) NULL
+#endif /* VDSO_HEADER */
+#endif /* vdso_image_info */
static void load_elf_vdso(struct image_info *info, const VdsoImageInfo *vdso)
{
@@ -3860,7 +3879,7 @@ int load_elf_binary(struct linux_binprm *bprm, struct image_info *info)
* Load a vdso if available, which will amongst other things contain the
* signal trampolines. Otherwise, allocate a separate page for them.
*/
- const VdsoImageInfo *vdso = vdso_image_info();
+ const VdsoImageInfo *vdso = vdso_image_info(info->elf_flags);
if (vdso) {
load_elf_vdso(&vdso_info, vdso);
info->vdso = vdso_info.load_bias;
diff --git a/linux-user/arm/Makefile.vdso b/linux-user/arm/Makefile.vdso
index 8a24b0e534..ede489e236 100644
--- a/linux-user/arm/Makefile.vdso
+++ b/linux-user/arm/Makefile.vdso
@@ -3,15 +3,18 @@ include $(BUILD_DIR)/tests/tcg/arm-linux-user/config-target.mak
SUBDIR = $(SRC_PATH)/linux-user/arm
VPATH += $(SUBDIR)
-all: $(SUBDIR)/vdso-be.so $(SUBDIR)/vdso-le.so
+all: $(SUBDIR)/vdso-be8.so $(SUBDIR)/vdso-be32.so $(SUBDIR)/vdso-le.so
# Adding -use-blx disables unneeded interworking without actually using blx.
LDFLAGS = -nostdlib -shared -Wl,-use-blx -Wl,-z,max-page-size=4096 \
-Wl,-h,linux-vdso.so.1 -Wl,--build-id=sha1 \
-Wl,--hash-style=both -Wl,-T,$(SUBDIR)/vdso.ld
-$(SUBDIR)/vdso-be.so: vdso.S vdso.ld vdso-asmoffset.h
- $(CC) -o $@ $(LDFLAGS) -mbig-endian $<
+$(SUBDIR)/vdso-be8.so: vdso.S vdso.ld vdso-asmoffset.h
+ $(CC) -o $@ $(LDFLAGS) -mbig-endian -mbe8 $<
+
+$(SUBDIR)/vdso-be32.so: vdso.S vdso.ld vdso-asmoffset.h
+ $(CC) -o $@ $(LDFLAGS) -mbig-endian -mbe32 $<
$(SUBDIR)/vdso-le.so: vdso.S vdso.ld vdso-asmoffset.h
$(CC) -o $@ $(LDFLAGS) -mlittle-endian $<
diff --git a/linux-user/arm/meson.build b/linux-user/arm/meson.build
index c4bb9af5b8..348ffb810d 100644
--- a/linux-user/arm/meson.build
+++ b/linux-user/arm/meson.build
@@ -10,10 +10,17 @@ syscall_nr_generators += {
# is always true as far as source_set.apply() is concerned. Always build
# both header files and include the right one via #if.
-vdso_be_inc = gen_vdso.process('vdso-be.so',
- extra_args: ['-s', 'sigreturn_codes'])
+vdso_be8_inc = gen_vdso.process('vdso-be8.so',
+ extra_args: ['-s', 'sigreturn_codes',
+ '-p', 'vdso_be8'])
+
+vdso_be32_inc = gen_vdso.process('vdso-be32.so',
+ extra_args: ['-s', 'sigreturn_codes',
+ '-p', 'vdso_be32'])
vdso_le_inc = gen_vdso.process('vdso-le.so',
extra_args: ['-s', 'sigreturn_codes'])
-linux_user_ss.add(when: 'TARGET_ARM', if_true: [vdso_be_inc, vdso_le_inc])
+linux_user_ss.add(when: 'TARGET_ARM', if_true: [
+ vdso_be8_inc, vdso_be32_inc, vdso_le_inc
+])
diff --git a/linux-user/arm/vdso-be32.so b/linux-user/arm/vdso-be32.so
new file mode 100755
index 0000000000000000000000000000000000000000..b896d3d545ebf91942038831a9535b023137a86b
GIT binary patch
literal 2648
zcmbtVO-x)>6h3eM80atz#?V$w#FU7s?MrPY6@OwJ7zS#9;uIpr#K$o6U?zom&CCPA
z_(K)jYPEI2#x#u>*BWAK(oG}X7}JGu(U>kYb%7gWj2bmA)bac7dzTK##(3a+=YHp$
zd+&Sqo^$4vzQK$UFl1ALazDn}kQ8ZH3rj#GDIF4quT>K8M*#RXv5GMf@}cL0QbMa9
zq(RhlE+~K0VF_bG%`dp$jfe!`*N`-Pg!23P2DC;e#zejUPViv=@R<Cvd*J4^Txa)J
z@7#0Z+Iu5y`)ZH;3||}=^X3>EPVMjCuphrI44sC?HyR7jO@IUJ0uoNG_9y)7(0_pk
zpnrn8jPd`K6kebEMDGxZK_3LSp__c(2L5av#~Fv7s^g4R(keB6AJ11y5Ch&BHHGIy
z!MCcr=6z^k4DXK!hxdkb4CjQ`+OJFT(1mRCl`nhCy&vqo@3Z~q#IoiKRmaNDIQbLS
zwC%d3S$jw0k>tP36}zf!{!C}8kXw|QQh9FSAlEhV|DX&W9UdFE4x2~t7C7+DKLhWH
zzvdr5eIIt-`P~5TRezZg{AliXF#0I!oQ(G1qZmH{ASaD+Xuz0+J_$VoeHMBidKLO*
zXxy{00z?7Y4#38I5#xod1YV3A?XhN3?z0Er`8P=_9z7E{&$Fcc5b*sD_5s+w-oZWr
z+ZRBzZk;4J_dJeGTe#`OQsPw0tMS*Gmt(K*I^A@p@oe;shBqVU!f%D%4xSHO;N}qS
z{}^x@Xp{c<x4yjt{BvL#h)E(Ie(Be6B$8-spxzf^+DHfR1Hf~MdNC0{CZc2D68U7Q
ztM*XzmRwy@qm|83qdyv^&aCUduIFzF$6X(K`r_5=vCO;uiHqNOI)ycN=j_~M3AZfL
z98Qi(B+%%;I(OV6H^`8{+Rr;64^YFrz$Z9!9Aez$6U46vW5egnd5G!<^s7SvPtyvH
zoS}Gi500Ere4S75Mr!!={_G);AkUwE_7X^Fgir7owH9zDlKuv~oT_KUc=7Ntz6CSY
ze2o3|FoO8;Xd7M+FueKg@SK9J1Bb!+59s&9nD6(vJ`X>B6gDM(KY0%G>-$p$oVBF;
z6Y^?Hf4T&IyFcagBjZl~;aSs>PFFe$Vx^xMP7NLHm1=3aV!LydvXyrVsvDqbTYUpo
zCYu`Svj);xvHAy(^`r)^W0}nHzENv5)ic;<)wKk_0c~j@+nZ8dquJ7^cIZi}e~DRL
zoON@P&~C-kMLjAzt}SL^v7DPN<<$VO&^b3J=5%?^Eas|3&+b+{%e-BwmYnhyVim-+
zi`G;nH)~tP0wOgN)3p~|F;m$gGv~UM(&U_rqxnyQ`&N0%+-F@7oq!fU=jC1)nBPb{
zl`qWMCBV;dq?4E5`=Oa<SdVq2YH0$Tam-c5u^%4xIo4x+sT!-#Z^jX<1=bnY+wdR<
z!1*2&9oLQDSoUwku^v@TCc(GI_4qc%r9G(MZJ3qldU_J;)nH-VWe=(z^m?$<Cr|6&
zc+hnZ(hWPmmwvt<WeVWhTn}fbY99M<ijH$(@#EMB2xI%Xa1wuO7>7Gyy{nqnbxuFe
zn6iBG_8EYNR`J*aXyF<Ze@~Af?g~t6VCWzC9e@Ggn(SHn3^0ro@%R|fqStGMp)RiB
MF7Oo{1g+x!0gGmCQ~&?~
literal 0
HcmV?d00001
diff --git a/linux-user/arm/vdso-be.so b/linux-user/arm/vdso-be8.so
similarity index 95%
rename from linux-user/arm/vdso-be.so
rename to linux-user/arm/vdso-be8.so
index bed02804a4bd367eb9fd8ca54d0c980103c02245..784b7bdb2a9308671bbc4c1e38ab4eac568ea6f7 100755
GIT binary patch
delta 19
acmca1azkW8HXECHj$?pF(B?w6=ga^|Lk8ji
delta 19
acmca1azkW8HXB=JPDHqi;pRfN=ga^~jt30@
--
2.43.0
^ permalink raw reply related [flat|nested] 17+ messages in thread* [PULL 14/14] tcg: Allow top bit of SIMD_DATA_BITS to be set in simd_desc()
2024-11-15 20:58 [PULL 00/14] tcg + linux-user patch queue Richard Henderson
` (12 preceding siblings ...)
2024-11-15 20:58 ` [PULL 13/14] linux-user/arm: Select vdso for be8 and be32 modes Richard Henderson
@ 2024-11-15 20:58 ` Richard Henderson
2024-11-16 10:39 ` [PULL 00/14] tcg + linux-user patch queue Peter Maydell
14 siblings, 0 replies; 17+ messages in thread
From: Richard Henderson @ 2024-11-15 20:58 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell, qemu-stable
From: Peter Maydell <peter.maydell@linaro.org>
In simd_desc() we create a SIMD descriptor from various pieces
including an arbitrary data value from the caller. We try to
sanitize these to make sure everything will fit: the 'data' value
needs to fit in the SIMD_DATA_BITS (== 22) sized field. However we
do that sanitizing with:
tcg_debug_assert(data == sextract32(data, 0, SIMD_DATA_BITS));
This works for the case where the data is supposed to be considered
as a signed integer (which can then be returned via simd_data()).
However, some callers want to treat the data value as unsigned.
Specifically, for the Arm SVE operations, make_svemte_desc()
assembles a data value as a collection of fields, and it needs to use
all 22 bits. Currently if MTE is enabled then its MTEDESC SIZEM1
field may have the most significant bit set, and then it will trip
this assertion.
Loosen the assertion so that we only check that the data value will
fit into the field in some way, either as a signed or as an unsigned
value. This means we will fail to detect some kinds of bug in the
callers, but we won't spuriously assert for intentional use of the
data field as unsigned.
Cc: qemu-stable@nongnu.org
Fixes: db432672dc50e ("tcg: Add generic vector expanders")
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/2601
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-ID: <20241115172515.1229393-1-peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
tcg/tcg-op-gvec.c | 15 ++++++++++++++-
1 file changed, 14 insertions(+), 1 deletion(-)
diff --git a/tcg/tcg-op-gvec.c b/tcg/tcg-op-gvec.c
index 78ee1ced80..97e4df221a 100644
--- a/tcg/tcg-op-gvec.c
+++ b/tcg/tcg-op-gvec.c
@@ -88,7 +88,20 @@ uint32_t simd_desc(uint32_t oprsz, uint32_t maxsz, int32_t data)
uint32_t desc = 0;
check_size_align(oprsz, maxsz, 0);
- tcg_debug_assert(data == sextract32(data, 0, SIMD_DATA_BITS));
+
+ /*
+ * We want to check that 'data' will fit into SIMD_DATA_BITS.
+ * However, some callers want to treat the data as a signed
+ * value (which they can later get back with simd_data())
+ * and some want to treat it as an unsigned value.
+ * So here we assert only that the data will fit into the
+ * field in at least one way. This means that some invalid
+ * values from the caller will not be detected, e.g. if the
+ * caller wants to handle the value as a signed integer but
+ * incorrectly passes us 1 << (SIMD_DATA_BITS - 1).
+ */
+ tcg_debug_assert(data == sextract32(data, 0, SIMD_DATA_BITS) ||
+ data == extract32(data, 0, SIMD_DATA_BITS));
oprsz = (oprsz / 8) - 1;
maxsz = (maxsz / 8) - 1;
--
2.43.0
^ permalink raw reply related [flat|nested] 17+ messages in thread* Re: [PULL 00/14] tcg + linux-user patch queue
2024-11-15 20:58 [PULL 00/14] tcg + linux-user patch queue Richard Henderson
` (13 preceding siblings ...)
2024-11-15 20:58 ` [PULL 14/14] tcg: Allow top bit of SIMD_DATA_BITS to be set in simd_desc() Richard Henderson
@ 2024-11-16 10:39 ` Peter Maydell
2024-11-16 16:38 ` Richard Henderson
14 siblings, 1 reply; 17+ messages in thread
From: Peter Maydell @ 2024-11-16 10:39 UTC (permalink / raw)
To: Richard Henderson; +Cc: qemu-devel
On Fri, 15 Nov 2024 at 20:59, Richard Henderson
<richard.henderson@linaro.org> wrote:
>
> The following changes since commit f0a5a31c33a8109061c2493e475c8a2f4d022432:
>
> Update version for v9.2.0-rc0 release (2024-11-13 21:44:45 +0000)
>
> are available in the Git repository at:
>
> https://gitlab.com/rth7680/qemu.git tags/pull-tcg-20241115
>
> for you to fetch changes up to a020e0a807a6e34890d1f2ca02e49f814fd0b64f:
>
> tcg: Allow top bit of SIMD_DATA_BITS to be set in simd_desc() (2024-11-15 12:51:27 -0800)
>
> ----------------------------------------------------------------
> cpu: ensure we don't call start_exclusive from cpu_exec
> tcg: Allow top bit of SIMD_DATA_BITS to be set in simd_desc()
> accel/tcg: Fix user-only probe_access_internal plugin check
> linux-user: Fix setreuid and setregid to use direct syscalls
> linux-user: Tolerate CONFIG_LSM_MMAP_MIN_ADDR
> linux-user: Honor elf alignment when placing images
> linux-user/*: Reduce vdso alignment to 4k
> linux-user/arm: Select vdso for be8 and be32 modes
This asserts on the user-mode tcg tests:
https://gitlab.com/qemu-project/qemu/-/jobs/8388145379
https://gitlab.com/qemu-project/qemu/-/jobs/8388145391
TEST catch-syscalls-with-libinline.so on aarch64
**
ERROR:../target/arm/tcg/sve_helper.c:6321:sve_stN_r: code should not be reached
timeout: the monitored command dumped core
Aborted
make[1]: *** [Makefile:185:
run-plugin-catch-syscalls-with-libinline.so] Error 134
Does this need "target/arm: Drop user-only special case in sve_stN_r"
now? (I'd picked that up into target-arm.next but feel free to slot it
in here if something in this series depends on it.)
thanks
-- PMM
^ permalink raw reply [flat|nested] 17+ messages in thread* Re: [PULL 00/14] tcg + linux-user patch queue
2024-11-16 10:39 ` [PULL 00/14] tcg + linux-user patch queue Peter Maydell
@ 2024-11-16 16:38 ` Richard Henderson
0 siblings, 0 replies; 17+ messages in thread
From: Richard Henderson @ 2024-11-16 16:38 UTC (permalink / raw)
To: Peter Maydell; +Cc: qemu-devel
On 11/16/24 02:39, Peter Maydell wrote:
>> cpu: ensure we don't call start_exclusive from cpu_exec
>> tcg: Allow top bit of SIMD_DATA_BITS to be set in simd_desc()
>> accel/tcg: Fix user-only probe_access_internal plugin check
>> linux-user: Fix setreuid and setregid to use direct syscalls
>> linux-user: Tolerate CONFIG_LSM_MMAP_MIN_ADDR
>> linux-user: Honor elf alignment when placing images
>> linux-user/*: Reduce vdso alignment to 4k
>> linux-user/arm: Select vdso for be8 and be32 modes
>
> This asserts on the user-mode tcg tests:
>
> https://gitlab.com/qemu-project/qemu/-/jobs/8388145379
> https://gitlab.com/qemu-project/qemu/-/jobs/8388145391
>
> TEST catch-syscalls-with-libinline.so on aarch64
> **
> ERROR:../target/arm/tcg/sve_helper.c:6321:sve_stN_r: code should not be reached
> timeout: the monitored command dumped core
> Aborted
> make[1]: *** [Makefile:185:
> run-plugin-catch-syscalls-with-libinline.so] Error 134
>
> Does this need "target/arm: Drop user-only special case in sve_stN_r"
> now? (I'd picked that up into target-arm.next but feel free to slot it
> in here if something in this series depends on it.)
Yes it does. I should have double-checked that was merged, not just queued.
I'll re-spin with that sorted before "Fix user-only probe_access_internal plugin check",
which is of course what exposes the target/arm issue.
r~
^ permalink raw reply [flat|nested] 17+ messages in thread