* [PULL 1/3] linux-user/x86_64: Fix ELF_PLATFORM
2022-06-24 9:49 [PULL 0/3] Linux user for 7.1 patches Laurent Vivier
@ 2022-06-24 9:49 ` Laurent Vivier
2022-06-24 9:49 ` [PULL 2/3] linux-user: Add partial support for MADV_DONTNEED Laurent Vivier
` (2 subsequent siblings)
3 siblings, 0 replies; 5+ messages in thread
From: Laurent Vivier @ 2022-06-24 9:49 UTC (permalink / raw)
To: qemu-devel; +Cc: Laurent Vivier, Richard Henderson
From: Richard Henderson <richard.henderson@linaro.org>
We had been using the i686 platform string for x86_64.
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1041
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Message-Id: <20220603213801.64738-1-richard.henderson@linaro.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
---
linux-user/elfload.c | 30 +++++++++++++++++-------------
1 file changed, 17 insertions(+), 13 deletions(-)
diff --git a/linux-user/elfload.c b/linux-user/elfload.c
index f7eae357f4ae..163fc8a1eeab 100644
--- a/linux-user/elfload.c
+++ b/linux-user/elfload.c
@@ -130,19 +130,6 @@ typedef abi_int target_pid_t;
#ifdef TARGET_I386
-#define ELF_PLATFORM get_elf_platform()
-
-static const char *get_elf_platform(void)
-{
- static char elf_platform[] = "i386";
- int family = object_property_get_int(OBJECT(thread_cpu), "family", NULL);
- if (family > 6)
- family = 6;
- if (family >= 3)
- elf_platform[1] = '0' + family;
- return elf_platform;
-}
-
#define ELF_HWCAP get_elf_hwcap()
static uint32_t get_elf_hwcap(void)
@@ -158,6 +145,8 @@ static uint32_t get_elf_hwcap(void)
#define ELF_CLASS ELFCLASS64
#define ELF_ARCH EM_X86_64
+#define ELF_PLATFORM "x86_64"
+
static inline void init_thread(struct target_pt_regs *regs, struct image_info *infop)
{
regs->rax = 0;
@@ -221,6 +210,21 @@ static void elf_core_copy_regs(target_elf_gregset_t *regs, const CPUX86State *en
#define ELF_CLASS ELFCLASS32
#define ELF_ARCH EM_386
+#define ELF_PLATFORM get_elf_platform()
+
+static const char *get_elf_platform(void)
+{
+ static char elf_platform[] = "i386";
+ int family = object_property_get_int(OBJECT(thread_cpu), "family", NULL);
+ if (family > 6) {
+ family = 6;
+ }
+ if (family >= 3) {
+ elf_platform[1] = '0' + family;
+ }
+ return elf_platform;
+}
+
static inline void init_thread(struct target_pt_regs *regs,
struct image_info *infop)
{
--
2.36.1
^ permalink raw reply related [flat|nested] 5+ messages in thread* [PULL 2/3] linux-user: Add partial support for MADV_DONTNEED
2022-06-24 9:49 [PULL 0/3] Linux user for 7.1 patches Laurent Vivier
2022-06-24 9:49 ` [PULL 1/3] linux-user/x86_64: Fix ELF_PLATFORM Laurent Vivier
@ 2022-06-24 9:49 ` Laurent Vivier
2022-06-24 9:49 ` [PULL 3/3] linux-user: Adjust child_tidptr on set_tid_address() syscall Laurent Vivier
2022-06-24 17:50 ` [PULL 0/3] Linux user for 7.1 patches Richard Henderson
3 siblings, 0 replies; 5+ messages in thread
From: Laurent Vivier @ 2022-06-24 9:49 UTC (permalink / raw)
To: qemu-devel; +Cc: Laurent Vivier, Ilya Leoshkevich
From: Ilya Leoshkevich <iii@linux.ibm.com>
Currently QEMU ignores madvise(MADV_DONTNEED), which break apps that
rely on this for zeroing out memory [1]. Improve the situation by doing
a passthrough when the range in question is a host-page-aligned
anonymous mapping.
This is based on the patches from Simon Hausmann [2] and Chris Fallin
[3]. The structure is taken from Simon's patch. The PAGE_MAP_ANONYMOUS
bits are superseded by commit 26bab757d41b ("linux-user: Introduce
PAGE_ANON"). In the end the patch acts like the one from Chris: we
either pass-through the entire syscall, or do nothing, since doing this
only partially would not help the affected applications much. Finally,
add some extra checks to match the behavior of the Linux kernel [4].
[1] https://gitlab.com/qemu-project/qemu/-/issues/326
[2] https://patchew.org/QEMU/20180827084037.25316-1-simon.hausmann@qt.io/
[3] https://github.com/bytecodealliance/wasmtime/blob/v0.37.0/ci/qemu-madvise.patch
[4] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/mm/madvise.c?h=v5.19-rc3#n1368
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Message-Id: <20220621144205.158452-1-iii@linux.ibm.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
---
linux-user/mmap.c | 64 +++++++++++++++++++++++++++++++++++++
linux-user/syscall.c | 8 ++---
linux-user/user-internals.h | 1 +
linux-user/user-mmap.h | 1 +
4 files changed, 68 insertions(+), 6 deletions(-)
diff --git a/linux-user/mmap.c b/linux-user/mmap.c
index 48e13737962a..4e7a6be6ee19 100644
--- a/linux-user/mmap.c
+++ b/linux-user/mmap.c
@@ -835,3 +835,67 @@ abi_long target_mremap(abi_ulong old_addr, abi_ulong old_size,
mmap_unlock();
return new_addr;
}
+
+static bool can_passthrough_madv_dontneed(abi_ulong start, abi_ulong end)
+{
+ ulong addr;
+
+ if ((start | end) & ~qemu_host_page_mask) {
+ return false;
+ }
+
+ for (addr = start; addr < end; addr += TARGET_PAGE_SIZE) {
+ if (!(page_get_flags(addr) & PAGE_ANON)) {
+ return false;
+ }
+ }
+
+ return true;
+}
+
+abi_long target_madvise(abi_ulong start, abi_ulong len_in, int advice)
+{
+ abi_ulong len, end;
+ int ret = 0;
+
+ if (start & ~TARGET_PAGE_MASK) {
+ return -TARGET_EINVAL;
+ }
+ len = TARGET_PAGE_ALIGN(len_in);
+
+ if (len_in && !len) {
+ return -TARGET_EINVAL;
+ }
+
+ end = start + len;
+ if (end < start) {
+ return -TARGET_EINVAL;
+ }
+
+ if (end == start) {
+ return 0;
+ }
+
+ if (!guest_range_valid_untagged(start, len)) {
+ return -TARGET_EINVAL;
+ }
+
+ /*
+ * A straight passthrough may not be safe because qemu sometimes turns
+ * private file-backed mappings into anonymous mappings.
+ *
+ * This is a hint, so ignoring and returning success is ok.
+ *
+ * This breaks MADV_DONTNEED, completely implementing which is quite
+ * complicated. However, there is one low-hanging fruit: host-page-aligned
+ * anonymous mappings. In this case passthrough is safe, so do it.
+ */
+ mmap_lock();
+ if ((advice & MADV_DONTNEED) &&
+ can_passthrough_madv_dontneed(start, end)) {
+ ret = get_errno(madvise(g2h_untagged(start), len, MADV_DONTNEED));
+ }
+ mmap_unlock();
+
+ return ret;
+}
diff --git a/linux-user/syscall.c b/linux-user/syscall.c
index f55cdebee5ab..8f68f255c0d9 100644
--- a/linux-user/syscall.c
+++ b/linux-user/syscall.c
@@ -538,7 +538,7 @@ static inline int target_to_host_errno(int target_errno)
}
}
-static inline abi_long get_errno(abi_long ret)
+abi_long get_errno(abi_long ret)
{
if (ret == -1)
return -host_to_target_errno(errno);
@@ -11807,11 +11807,7 @@ static abi_long do_syscall1(CPUArchState *cpu_env, int num, abi_long arg1,
#ifdef TARGET_NR_madvise
case TARGET_NR_madvise:
- /* A straight passthrough may not be safe because qemu sometimes
- turns private file-backed mappings into anonymous mappings.
- This will break MADV_DONTNEED.
- This is a hint, so ignoring and returning success is ok. */
- return 0;
+ return target_madvise(arg1, arg2, arg3);
#endif
#ifdef TARGET_NR_fcntl64
case TARGET_NR_fcntl64:
diff --git a/linux-user/user-internals.h b/linux-user/user-internals.h
index 6175ce53dbd6..0280e76addda 100644
--- a/linux-user/user-internals.h
+++ b/linux-user/user-internals.h
@@ -65,6 +65,7 @@ abi_long do_syscall(CPUArchState *cpu_env, int num, abi_long arg1,
abi_long arg8);
extern __thread CPUState *thread_cpu;
G_NORETURN void cpu_loop(CPUArchState *env);
+abi_long get_errno(abi_long ret);
const char *target_strerror(int err);
int get_osversion(void);
void init_qemu_uname_release(void);
diff --git a/linux-user/user-mmap.h b/linux-user/user-mmap.h
index d1dec99c0249..480ce1c11487 100644
--- a/linux-user/user-mmap.h
+++ b/linux-user/user-mmap.h
@@ -25,6 +25,7 @@ int target_munmap(abi_ulong start, abi_ulong len);
abi_long target_mremap(abi_ulong old_addr, abi_ulong old_size,
abi_ulong new_size, unsigned long flags,
abi_ulong new_addr);
+abi_long target_madvise(abi_ulong start, abi_ulong len_in, int advice);
extern unsigned long last_brk;
extern abi_ulong mmap_next_start;
abi_ulong mmap_find_vma(abi_ulong, abi_ulong, abi_ulong);
--
2.36.1
^ permalink raw reply related [flat|nested] 5+ messages in thread* [PULL 3/3] linux-user: Adjust child_tidptr on set_tid_address() syscall
2022-06-24 9:49 [PULL 0/3] Linux user for 7.1 patches Laurent Vivier
2022-06-24 9:49 ` [PULL 1/3] linux-user/x86_64: Fix ELF_PLATFORM Laurent Vivier
2022-06-24 9:49 ` [PULL 2/3] linux-user: Add partial support for MADV_DONTNEED Laurent Vivier
@ 2022-06-24 9:49 ` Laurent Vivier
2022-06-24 17:50 ` [PULL 0/3] Linux user for 7.1 patches Richard Henderson
3 siblings, 0 replies; 5+ messages in thread
From: Laurent Vivier @ 2022-06-24 9:49 UTC (permalink / raw)
To: qemu-devel; +Cc: Laurent Vivier, Helge Deller, Richard Henderson
From: Helge Deller <deller@gmx.de>
Keep track of the new child tidptr given by a set_tid_address() syscall.
Do not call the host set_tid_address() syscall because we are emulating
the behaviour of writing to child_tidptr in the exit() path.
Signed-off-by: Helge Deller<deller@gmx.de>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Message-Id: <YpH+2sw1PCRqx/te@p100>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
---
linux-user/syscall.c | 12 +++++++-----
1 file changed, 7 insertions(+), 5 deletions(-)
diff --git a/linux-user/syscall.c b/linux-user/syscall.c
index 8f68f255c0d9..669add74c11a 100644
--- a/linux-user/syscall.c
+++ b/linux-user/syscall.c
@@ -320,9 +320,6 @@ _syscall3(int,sys_syslog,int,type,char*,bufp,int,len)
#ifdef __NR_exit_group
_syscall1(int,exit_group,int,error_code)
#endif
-#if defined(TARGET_NR_set_tid_address) && defined(__NR_set_tid_address)
-_syscall1(int,set_tid_address,int *,tidptr)
-#endif
#if defined(__NR_futex)
_syscall6(int,sys_futex,int *,uaddr,int,op,int,val,
const struct timespec *,timeout,int *,uaddr2,int,val3)
@@ -12196,9 +12193,14 @@ static abi_long do_syscall1(CPUArchState *cpu_env, int num, abi_long arg1,
}
#endif
-#if defined(TARGET_NR_set_tid_address) && defined(__NR_set_tid_address)
+#if defined(TARGET_NR_set_tid_address)
case TARGET_NR_set_tid_address:
- return get_errno(set_tid_address((int *)g2h(cpu, arg1)));
+ {
+ TaskState *ts = cpu->opaque;
+ ts->child_tidptr = arg1;
+ /* do not call host set_tid_address() syscall, instead return tid() */
+ return get_errno(sys_gettid());
+ }
#endif
case TARGET_NR_tkill:
--
2.36.1
^ permalink raw reply related [flat|nested] 5+ messages in thread* Re: [PULL 0/3] Linux user for 7.1 patches
2022-06-24 9:49 [PULL 0/3] Linux user for 7.1 patches Laurent Vivier
` (2 preceding siblings ...)
2022-06-24 9:49 ` [PULL 3/3] linux-user: Adjust child_tidptr on set_tid_address() syscall Laurent Vivier
@ 2022-06-24 17:50 ` Richard Henderson
3 siblings, 0 replies; 5+ messages in thread
From: Richard Henderson @ 2022-06-24 17:50 UTC (permalink / raw)
To: Laurent Vivier, qemu-devel
On 6/24/22 02:49, Laurent Vivier wrote:
> The following changes since commit c8b2d413761af732a0798d8df45ce968732083fe:
>
> Merge tag 'bsd-user-syscall-2022q2-pull-request' of ssh://github.com/qemu-bsd-user/qemu-bsd-user into staging (2022-06-19 13:56:13 -0700)
>
> are available in the Git repository at:
>
> https://gitlab.com/laurent_vivier/qemu.git tags/linux-user-for-7.1-pull-request
>
> for you to fetch changes up to 9a7f682c26acae5bc8bfd1f7c774070da54f1625:
>
> linux-user: Adjust child_tidptr on set_tid_address() syscall (2022-06-24 10:00:01 +0200)
>
> ----------------------------------------------------------------
> linux-user pull request 20220624
Applied, thanks. Please update https://wiki.qemu.org/ChangeLog/7.1 as appropriate.
r~
>
> ----------------------------------------------------------------
>
> Helge Deller (1):
> linux-user: Adjust child_tidptr on set_tid_address() syscall
>
> Ilya Leoshkevich (1):
> linux-user: Add partial support for MADV_DONTNEED
>
> Richard Henderson (1):
> linux-user/x86_64: Fix ELF_PLATFORM
>
> linux-user/elfload.c | 30 +++++++++--------
> linux-user/mmap.c | 64 +++++++++++++++++++++++++++++++++++++
> linux-user/syscall.c | 20 ++++++------
> linux-user/user-internals.h | 1 +
> linux-user/user-mmap.h | 1 +
> 5 files changed, 92 insertions(+), 24 deletions(-)
>
^ permalink raw reply [flat|nested] 5+ messages in thread