* [Qemu-devel] [PATCH 01/18] Don't translate pointer when in restore_sigcontext
2011-06-20 16:20 [Qemu-devel] [PATCH 00/18] pending linux-user patches riku.voipio
@ 2011-06-20 16:20 ` riku.voipio
2011-06-20 16:20 ` [Qemu-devel] [PATCH 02/18] linux-user: Fix the load of ELF files that have no "useful" symbol riku.voipio
` (16 subsequent siblings)
17 siblings, 0 replies; 23+ messages in thread
From: riku.voipio @ 2011-06-20 16:20 UTC (permalink / raw)
To: qemu-devel; +Cc: Mike McCormack
From: Mike McCormack <mj.mccormack@samsung.com>
Fixes crash in i386 when user emulation base address is non-zero.
21797 rt_sigreturn(8,1082124603,1,0,1082126048,1082126248)Exit reason and status: signal 11
Signed-off-by: Mike McCormack <mj.mccormack@samsung.com>
Signed-off-by: Riku Voipio <riku.voipio@iki.fi>
---
linux-user/signal.c | 4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/linux-user/signal.c b/linux-user/signal.c
index 11b25be..cb7138f 100644
--- a/linux-user/signal.c
+++ b/linux-user/signal.c
@@ -981,8 +981,8 @@ restore_sigcontext(CPUX86State *env, struct target_sigcontext *sc, int *peax)
env->regs[R_ECX] = tswapl(sc->ecx);
env->eip = tswapl(sc->eip);
- cpu_x86_load_seg(env, R_CS, lduw(&sc->cs) | 3);
- cpu_x86_load_seg(env, R_SS, lduw(&sc->ss) | 3);
+ cpu_x86_load_seg(env, R_CS, lduw_p(&sc->cs) | 3);
+ cpu_x86_load_seg(env, R_SS, lduw_p(&sc->ss) | 3);
tmpflags = tswapl(sc->eflags);
env->eflags = (env->eflags & ~0x40DD5) | (tmpflags & 0x40DD5);
--
1.7.4.1
^ permalink raw reply related [flat|nested] 23+ messages in thread* [Qemu-devel] [PATCH 02/18] linux-user: Fix the load of ELF files that have no "useful" symbol
2011-06-20 16:20 [Qemu-devel] [PATCH 00/18] pending linux-user patches riku.voipio
2011-06-20 16:20 ` [Qemu-devel] [PATCH 01/18] Don't translate pointer when in restore_sigcontext riku.voipio
@ 2011-06-20 16:20 ` riku.voipio
2011-06-21 7:07 ` cedric.vincent
2011-06-20 16:20 ` [Qemu-devel] [PATCH 03/18] linux-user: Handle images where lowest vaddr is not page aligned riku.voipio
` (15 subsequent siblings)
17 siblings, 1 reply; 23+ messages in thread
From: riku.voipio @ 2011-06-20 16:20 UTC (permalink / raw)
To: qemu-devel; +Cc: Cédric VINCENT
From: Cédric VINCENT <cedric.vincent@st.com>
This patch fixes a "double free()" due to "realloc(syms, 0)" in the
loader when the ELF file has no "useful" symbol, as with the following
example (compiled with "sh4-linux-gcc -nostdlib"):
.text
.align 1
.global _start
_start:
mov #1, r3
trapa #40 // syscall(__NR_exit)
nop
The bug appears when the log (option "-d") is enabled.
Signed-off-by: Cédric VINCENT <cedric.vincent@st.com>
Signed-off-by: Yves JANIN <yves.janin@st.com>
Signed-off-by: Riku Voipio <riku.voipio@iki.fi>
---
linux-user/elfload.c | 34 +++++++++++++++++++---------------
1 files changed, 19 insertions(+), 15 deletions(-)
diff --git a/linux-user/elfload.c b/linux-user/elfload.c
index dcfeb7a..a4aabd5 100644
--- a/linux-user/elfload.c
+++ b/linux-user/elfload.c
@@ -1643,9 +1643,9 @@ static void load_symbols(struct elfhdr *hdr, int fd, abi_ulong load_bias)
{
int i, shnum, nsyms, sym_idx = 0, str_idx = 0;
struct elf_shdr *shdr;
- char *strings;
- struct syminfo *s;
- struct elf_sym *syms, *new_syms;
+ char *strings = NULL;
+ struct syminfo *s = NULL;
+ struct elf_sym *new_syms, *syms = NULL;
shnum = hdr->e_shnum;
i = shnum * sizeof(struct elf_shdr);
@@ -1670,24 +1670,19 @@ static void load_symbols(struct elfhdr *hdr, int fd, abi_ulong load_bias)
/* Now know where the strtab and symtab are. Snarf them. */
s = malloc(sizeof(*s));
if (!s) {
- return;
+ goto give_up;
}
i = shdr[str_idx].sh_size;
s->disas_strtab = strings = malloc(i);
if (!strings || pread(fd, strings, i, shdr[str_idx].sh_offset) != i) {
- free(s);
- free(strings);
- return;
+ goto give_up;
}
i = shdr[sym_idx].sh_size;
syms = malloc(i);
if (!syms || pread(fd, syms, i, shdr[sym_idx].sh_offset) != i) {
- free(s);
- free(strings);
- free(syms);
- return;
+ goto give_up;
}
nsyms = i / sizeof(struct elf_sym);
@@ -1710,16 +1705,18 @@ static void load_symbols(struct elfhdr *hdr, int fd, abi_ulong load_bias)
}
}
+ /* No "useful" symbol. */
+ if (nsyms == 0) {
+ goto give_up;
+ }
+
/* Attempt to free the storage associated with the local symbols
that we threw away. Whether or not this has any effect on the
memory allocation depends on the malloc implementation and how
many symbols we managed to discard. */
new_syms = realloc(syms, nsyms * sizeof(*syms));
if (new_syms == NULL) {
- free(s);
- free(syms);
- free(strings);
- return;
+ goto give_up;
}
syms = new_syms;
@@ -1734,6 +1731,13 @@ static void load_symbols(struct elfhdr *hdr, int fd, abi_ulong load_bias)
s->lookup_symbol = lookup_symbolxx;
s->next = syminfos;
syminfos = s;
+
+ return;
+
+give_up:
+ free(s);
+ free(strings);
+ free(syms);
}
int load_elf_binary(struct linux_binprm * bprm, struct target_pt_regs * regs,
--
1.7.4.1
^ permalink raw reply related [flat|nested] 23+ messages in thread* [Qemu-devel] [PATCH 03/18] linux-user: Handle images where lowest vaddr is not page aligned
2011-06-20 16:20 [Qemu-devel] [PATCH 00/18] pending linux-user patches riku.voipio
2011-06-20 16:20 ` [Qemu-devel] [PATCH 01/18] Don't translate pointer when in restore_sigcontext riku.voipio
2011-06-20 16:20 ` [Qemu-devel] [PATCH 02/18] linux-user: Fix the load of ELF files that have no "useful" symbol riku.voipio
@ 2011-06-20 16:20 ` riku.voipio
2011-06-20 16:20 ` [Qemu-devel] [PATCH 04/18] linux-user: Don't use MAP_FIXED in do_brk() riku.voipio
` (14 subsequent siblings)
17 siblings, 0 replies; 23+ messages in thread
From: riku.voipio @ 2011-06-20 16:20 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell
From: Peter Maydell <peter.maydell@linaro.org>
Fix a bug in the linux-user ELF loader code where it was not correctly
handling images where the lowest vaddr to be loaded was not page aligned.
The problem was that the code to probe for a suitable guest base address
was changing the 'loaddr' variable (by rounding it to a page boundary),
which meant that the load bias would then be incorrectly calculated
unless loaddr happened to already be page-aligned.
Binaries generated by gcc with the default linker script do start with
a loadable segment at a page-aligned vaddr, so were unaffected. This
bug was noticed with a binary created by the Google Go toolchain for ARM.
We fix the bug by refactoring the "probe for guest base" code out into
its own self-contained function.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@iki.fi>
---
linux-user/elfload.c | 130 ++++++++++++++++++++++++++++----------------------
1 files changed, 73 insertions(+), 57 deletions(-)
diff --git a/linux-user/elfload.c b/linux-user/elfload.c
index a4aabd5..a13eb7b 100644
--- a/linux-user/elfload.c
+++ b/linux-user/elfload.c
@@ -1288,6 +1288,78 @@ static abi_ulong create_elf_tables(abi_ulong p, int argc, int envc,
return sp;
}
+static void probe_guest_base(const char *image_name,
+ abi_ulong loaddr, abi_ulong hiaddr)
+{
+ /* Probe for a suitable guest base address, if the user has not set
+ * it explicitly, and set guest_base appropriately.
+ * In case of error we will print a suitable message and exit.
+ */
+#if defined(CONFIG_USE_GUEST_BASE)
+ const char *errmsg;
+ if (!have_guest_base && !reserved_va) {
+ unsigned long host_start, real_start, host_size;
+
+ /* Round addresses to page boundaries. */
+ loaddr &= qemu_host_page_mask;
+ hiaddr = HOST_PAGE_ALIGN(hiaddr);
+
+ if (loaddr < mmap_min_addr) {
+ host_start = HOST_PAGE_ALIGN(mmap_min_addr);
+ } else {
+ host_start = loaddr;
+ if (host_start != loaddr) {
+ errmsg = "Address overflow loading ELF binary";
+ goto exit_errmsg;
+ }
+ }
+ host_size = hiaddr - loaddr;
+ while (1) {
+ /* Do not use mmap_find_vma here because that is limited to the
+ guest address space. We are going to make the
+ guest address space fit whatever we're given. */
+ real_start = (unsigned long)
+ mmap((void *)host_start, host_size, PROT_NONE,
+ MAP_ANONYMOUS | MAP_PRIVATE | MAP_NORESERVE, -1, 0);
+ if (real_start == (unsigned long)-1) {
+ goto exit_perror;
+ }
+ if (real_start == host_start) {
+ break;
+ }
+ /* That address didn't work. Unmap and try a different one.
+ The address the host picked because is typically right at
+ the top of the host address space and leaves the guest with
+ no usable address space. Resort to a linear search. We
+ already compensated for mmap_min_addr, so this should not
+ happen often. Probably means we got unlucky and host
+ address space randomization put a shared library somewhere
+ inconvenient. */
+ munmap((void *)real_start, host_size);
+ host_start += qemu_host_page_size;
+ if (host_start == loaddr) {
+ /* Theoretically possible if host doesn't have any suitably
+ aligned areas. Normally the first mmap will fail. */
+ errmsg = "Unable to find space for application";
+ goto exit_errmsg;
+ }
+ }
+ qemu_log("Relocating guest address space from 0x"
+ TARGET_ABI_FMT_lx " to 0x%lx\n",
+ loaddr, real_start);
+ guest_base = real_start - loaddr;
+ }
+ return;
+
+exit_perror:
+ errmsg = strerror(errno);
+exit_errmsg:
+ fprintf(stderr, "%s: %s\n", image_name, errmsg);
+ exit(-1);
+#endif
+}
+
+
/* Load an ELF image into the address space.
IMAGE_NAME is the filename of the image, to use in error messages.
@@ -1373,63 +1445,7 @@ static void load_elf_image(const char *image_name, int image_fd,
/* This is the main executable. Make sure that the low
address does not conflict with MMAP_MIN_ADDR or the
QEMU application itself. */
-#if defined(CONFIG_USE_GUEST_BASE)
- /*
- * In case where user has not explicitly set the guest_base, we
- * probe here that should we set it automatically.
- */
- if (!have_guest_base && !reserved_va) {
- unsigned long host_start, real_start, host_size;
-
- /* Round addresses to page boundaries. */
- loaddr &= qemu_host_page_mask;
- hiaddr = HOST_PAGE_ALIGN(hiaddr);
-
- if (loaddr < mmap_min_addr) {
- host_start = HOST_PAGE_ALIGN(mmap_min_addr);
- } else {
- host_start = loaddr;
- if (host_start != loaddr) {
- errmsg = "Address overflow loading ELF binary";
- goto exit_errmsg;
- }
- }
- host_size = hiaddr - loaddr;
- while (1) {
- /* Do not use mmap_find_vma here because that is limited to the
- guest address space. We are going to make the
- guest address space fit whatever we're given. */
- real_start = (unsigned long)
- mmap((void *)host_start, host_size, PROT_NONE,
- MAP_ANONYMOUS | MAP_PRIVATE | MAP_NORESERVE, -1, 0);
- if (real_start == (unsigned long)-1) {
- goto exit_perror;
- }
- if (real_start == host_start) {
- break;
- }
- /* That address didn't work. Unmap and try a different one.
- The address the host picked because is typically right at
- the top of the host address space and leaves the guest with
- no usable address space. Resort to a linear search. We
- already compensated for mmap_min_addr, so this should not
- happen often. Probably means we got unlucky and host
- address space randomization put a shared library somewhere
- inconvenient. */
- munmap((void *)real_start, host_size);
- host_start += qemu_host_page_size;
- if (host_start == loaddr) {
- /* Theoretically possible if host doesn't have any suitably
- aligned areas. Normally the first mmap will fail. */
- errmsg = "Unable to find space for application";
- goto exit_errmsg;
- }
- }
- qemu_log("Relocating guest address space from 0x"
- TARGET_ABI_FMT_lx " to 0x%lx\n", loaddr, real_start);
- guest_base = real_start - loaddr;
- }
-#endif
+ probe_guest_base(image_name, loaddr, hiaddr);
}
load_bias = load_addr - loaddr;
--
1.7.4.1
^ permalink raw reply related [flat|nested] 23+ messages in thread* [Qemu-devel] [PATCH 04/18] linux-user: Don't use MAP_FIXED in do_brk()
2011-06-20 16:20 [Qemu-devel] [PATCH 00/18] pending linux-user patches riku.voipio
` (2 preceding siblings ...)
2011-06-20 16:20 ` [Qemu-devel] [PATCH 03/18] linux-user: Handle images where lowest vaddr is not page aligned riku.voipio
@ 2011-06-20 16:20 ` riku.voipio
2011-06-20 16:20 ` [Qemu-devel] [PATCH 05/18] arm-semi.c: Use correct check for failure of do_brk() riku.voipio
` (13 subsequent siblings)
17 siblings, 0 replies; 23+ messages in thread
From: riku.voipio @ 2011-06-20 16:20 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell
From: Peter Maydell <peter.maydell@linaro.org>
Since mmap() with MAP_FIXED will map over the top of existing mappings,
it's a bad idea to use it to implement brk(), because brk() with a
large size is likely to overwrite important things like qemu itself
or the host libc. So we drop MAP_FIXED and handle "mapped but at
different address" as an error case instead.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@iki.fi>
---
linux-user/syscall.c | 29 ++++++++++++++++++++---------
1 files changed, 20 insertions(+), 9 deletions(-)
diff --git a/linux-user/syscall.c b/linux-user/syscall.c
index 5cb27c7..b975730 100644
--- a/linux-user/syscall.c
+++ b/linux-user/syscall.c
@@ -735,23 +735,34 @@ abi_long do_brk(abi_ulong new_brk)
return target_brk;
}
- /* We need to allocate more memory after the brk... */
+ /* We need to allocate more memory after the brk... Note that
+ * we don't use MAP_FIXED because that will map over the top of
+ * any existing mapping (like the one with the host libc or qemu
+ * itself); instead we treat "mapped but at wrong address" as
+ * a failure and unmap again.
+ */
new_alloc_size = HOST_PAGE_ALIGN(new_brk - brk_page + 1);
mapped_addr = get_errno(target_mmap(brk_page, new_alloc_size,
PROT_READ|PROT_WRITE,
- MAP_ANON|MAP_FIXED|MAP_PRIVATE, 0, 0));
+ MAP_ANON|MAP_PRIVATE, 0, 0));
+
+ if (mapped_addr == brk_page) {
+ target_brk = new_brk;
+ return target_brk;
+ } else if (mapped_addr != -1) {
+ /* Mapped but at wrong address, meaning there wasn't actually
+ * enough space for this brk.
+ */
+ target_munmap(mapped_addr, new_alloc_size);
+ mapped_addr = -1;
+ }
#if defined(TARGET_ALPHA)
/* We (partially) emulate OSF/1 on Alpha, which requires we
return a proper errno, not an unchanged brk value. */
- if (is_error(mapped_addr)) {
- return -TARGET_ENOMEM;
- }
+ return -TARGET_ENOMEM;
#endif
-
- if (!is_error(mapped_addr)) {
- target_brk = new_brk;
- }
+ /* For everything else, return the previous break. */
return target_brk;
}
--
1.7.4.1
^ permalink raw reply related [flat|nested] 23+ messages in thread* [Qemu-devel] [PATCH 05/18] arm-semi.c: Use correct check for failure of do_brk()
2011-06-20 16:20 [Qemu-devel] [PATCH 00/18] pending linux-user patches riku.voipio
` (3 preceding siblings ...)
2011-06-20 16:20 ` [Qemu-devel] [PATCH 04/18] linux-user: Don't use MAP_FIXED in do_brk() riku.voipio
@ 2011-06-20 16:20 ` riku.voipio
2011-06-20 16:20 ` [Qemu-devel] [PATCH 06/18] m68k-semi.c: " riku.voipio
` (12 subsequent siblings)
17 siblings, 0 replies; 23+ messages in thread
From: riku.voipio @ 2011-06-20 16:20 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell
From: Peter Maydell <peter.maydell@linaro.org>
In the ARM semihosting implementation of SYS_HEAPINFO, use the correct
check for whether do_brk() has failed -- it does not return -1 but the
previous value of the break limit.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@iki.fi>
---
arm-semi.c | 5 +++--
1 files changed, 3 insertions(+), 2 deletions(-)
diff --git a/arm-semi.c b/arm-semi.c
index e9e6f89..5a62d03 100644
--- a/arm-semi.c
+++ b/arm-semi.c
@@ -440,15 +440,16 @@ uint32_t do_arm_semihosting(CPUState *env)
/* Some C libraries assume the heap immediately follows .bss, so
allocate it using sbrk. */
if (!ts->heap_limit) {
- long ret;
+ abi_ulong ret;
ts->heap_base = do_brk(0);
limit = ts->heap_base + ARM_ANGEL_HEAP_SIZE;
/* Try a big heap, and reduce the size if that fails. */
for (;;) {
ret = do_brk(limit);
- if (ret != -1)
+ if (ret >= limit) {
break;
+ }
limit = (ts->heap_base >> 1) + (limit >> 1);
}
ts->heap_limit = limit;
--
1.7.4.1
^ permalink raw reply related [flat|nested] 23+ messages in thread* [Qemu-devel] [PATCH 06/18] m68k-semi.c: Use correct check for failure of do_brk()
2011-06-20 16:20 [Qemu-devel] [PATCH 00/18] pending linux-user patches riku.voipio
` (4 preceding siblings ...)
2011-06-20 16:20 ` [Qemu-devel] [PATCH 05/18] arm-semi.c: Use correct check for failure of do_brk() riku.voipio
@ 2011-06-20 16:20 ` riku.voipio
2011-06-20 16:20 ` [Qemu-devel] [PATCH 07/18] linux-user: Fix the computation of the requested heap size riku.voipio
` (11 subsequent siblings)
17 siblings, 0 replies; 23+ messages in thread
From: riku.voipio @ 2011-06-20 16:20 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell
From: Peter Maydell <peter.maydell@linaro.org>
In the m68k semihosting implementation of HOSTED_INIT_SIM, use the correct
check for whether do_brk() has failed -- it does not return -1 but the
previous value of the break limit.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@iki.fi>
---
m68k-semi.c | 5 +++--
1 files changed, 3 insertions(+), 2 deletions(-)
diff --git a/m68k-semi.c b/m68k-semi.c
index 0371089..7fde10e 100644
--- a/m68k-semi.c
+++ b/m68k-semi.c
@@ -370,7 +370,7 @@ void do_m68k_semihosting(CPUM68KState *env, int nr)
TaskState *ts = env->opaque;
/* Allocate the heap using sbrk. */
if (!ts->heap_limit) {
- long ret;
+ abi_ulong ret;
uint32_t size;
uint32_t base;
@@ -379,8 +379,9 @@ void do_m68k_semihosting(CPUM68KState *env, int nr)
/* Try a big heap, and reduce the size if that fails. */
for (;;) {
ret = do_brk(base + size);
- if (ret != -1)
+ if (ret >= (base + size)) {
break;
+ }
size >>= 1;
}
ts->heap_limit = base + size;
--
1.7.4.1
^ permalink raw reply related [flat|nested] 23+ messages in thread* [Qemu-devel] [PATCH 07/18] linux-user: Fix the computation of the requested heap size
2011-06-20 16:20 [Qemu-devel] [PATCH 00/18] pending linux-user patches riku.voipio
` (5 preceding siblings ...)
2011-06-20 16:20 ` [Qemu-devel] [PATCH 06/18] m68k-semi.c: " riku.voipio
@ 2011-06-20 16:20 ` riku.voipio
2011-06-21 7:11 ` cedric.vincent
2011-06-20 16:20 ` [Qemu-devel] [PATCH 08/18] linux-user: add pselect6 syscall support riku.voipio
` (10 subsequent siblings)
17 siblings, 1 reply; 23+ messages in thread
From: riku.voipio @ 2011-06-20 16:20 UTC (permalink / raw)
To: qemu-devel; +Cc: Cédric VINCENT
From: Cédric VINCENT <cedric.vincent@st.com>
There were two remaining bugs in the previous implementation of
do_brk():
1. the value of "new_alloc_size" was one page too large when the
requested brk was aligned on a host page boundary.
2. no new pages should be (re-)allocated when the requested brk is
in the range of the pages that were already allocated
previsouly (for the same purpose). Technically these pages are
never unmapped in the current implementation.
The problem/fix can be reproduced/validated with the following test
case:
#include <unistd.h> /* syscall(2), */
#include <sys/syscall.h> /* SYS_brk, */
#include <stdio.h> /* puts(3), */
#include <stdlib.h> /* exit(3), EXIT_*, */
int main()
{
int current_brk = 0;
int new_brk;
int failure = 0;
void test(int increment) {
static int test_number = 0;
test_number++;
new_brk = syscall(SYS_brk, current_brk + increment);
if (new_brk == current_brk) {
printf("test %d fails\n", test_number);
failure++;
}
current_brk = new_brk;
}
/* Initialization. */
test(0);
/* Does QEMU overlap host pages? */
test(HOST_PAGE_SIZE);
test(HOST_PAGE_SIZE);
/* Does QEMU allocate the same host page twice? */
test(-HOST_PAGE_SIZE);
test(HOST_PAGE_SIZE);
if (!failure) {
printf("success\n");
exit(EXIT_SUCCESS);
}
else {
exit(EXIT_FAILURE);
}
}
Signed-off-by: Cédric VINCENT <cedric.vincent@st.com>
Reviewed-by: Christophe Guillon <christophe.guillon@st.com>
Signed-off-by: Riku Voipio <riku.voipio@iki.fi>
---
linux-user/syscall.c | 11 ++++++-----
1 files changed, 6 insertions(+), 5 deletions(-)
diff --git a/linux-user/syscall.c b/linux-user/syscall.c
index b975730..be27f53 100644
--- a/linux-user/syscall.c
+++ b/linux-user/syscall.c
@@ -709,16 +709,17 @@ char *target_strerror(int err)
static abi_ulong target_brk;
static abi_ulong target_original_brk;
+static abi_ulong brk_page;
void target_set_brk(abi_ulong new_brk)
{
target_original_brk = target_brk = HOST_PAGE_ALIGN(new_brk);
+ brk_page = HOST_PAGE_ALIGN(target_brk);
}
/* do_brk() must return target values and target errnos. */
abi_long do_brk(abi_ulong new_brk)
{
- abi_ulong brk_page;
abi_long mapped_addr;
int new_alloc_size;
@@ -727,9 +728,8 @@ abi_long do_brk(abi_ulong new_brk)
if (new_brk < target_original_brk)
return target_brk;
- brk_page = HOST_PAGE_ALIGN(target_brk);
-
- /* If the new brk is less than this, set it and we're done... */
+ /* If the new brk is less than the highest page reserved to the
+ * target heap allocation, set it and we're done... */
if (new_brk < brk_page) {
target_brk = new_brk;
return target_brk;
@@ -741,13 +741,14 @@ abi_long do_brk(abi_ulong new_brk)
* itself); instead we treat "mapped but at wrong address" as
* a failure and unmap again.
*/
- new_alloc_size = HOST_PAGE_ALIGN(new_brk - brk_page + 1);
+ new_alloc_size = HOST_PAGE_ALIGN(new_brk - brk_page);
mapped_addr = get_errno(target_mmap(brk_page, new_alloc_size,
PROT_READ|PROT_WRITE,
MAP_ANON|MAP_PRIVATE, 0, 0));
if (mapped_addr == brk_page) {
target_brk = new_brk;
+ brk_page = HOST_PAGE_ALIGN(target_brk);
return target_brk;
} else if (mapped_addr != -1) {
/* Mapped but at wrong address, meaning there wasn't actually
--
1.7.4.1
^ permalink raw reply related [flat|nested] 23+ messages in thread* Re: [Qemu-devel] [PATCH 07/18] linux-user: Fix the computation of the requested heap size
2011-06-20 16:20 ` [Qemu-devel] [PATCH 07/18] linux-user: Fix the computation of the requested heap size riku.voipio
@ 2011-06-21 7:11 ` cedric.vincent
2011-06-21 17:50 ` Riku Voipio
0 siblings, 1 reply; 23+ messages in thread
From: cedric.vincent @ 2011-06-21 7:11 UTC (permalink / raw)
To: riku.voipio@iki.fi; +Cc: qemu-devel@nongnu.org
On Mon, Jun 20, 2011 at 06:20:12PM +0200, riku.voipio@iki.fi wrote:
> linux-user/syscall.c | 11 ++++++-----
> 1 files changed, 6 insertions(+), 5 deletions(-)
This is not the latest version of the patch (that one introduced some
regressions):
http://patchwork.ozlabs.org/patch/100492/
^ permalink raw reply [flat|nested] 23+ messages in thread* Re: [Qemu-devel] [PATCH 07/18] linux-user: Fix the computation of the requested heap size
2011-06-21 7:11 ` cedric.vincent
@ 2011-06-21 17:50 ` Riku Voipio
0 siblings, 0 replies; 23+ messages in thread
From: Riku Voipio @ 2011-06-21 17:50 UTC (permalink / raw)
To: cedric.vincent, qemu-devel@nongnu.org
On Tue, Jun 21, 2011 at 09:11:00AM +0200, cedric.vincent@st.com wrote:
> On Mon, Jun 20, 2011 at 06:20:12PM +0200, riku.voipio@iki.fi wrote:
> > linux-user/syscall.c | 11 ++++++-----
> > 1 files changed, 6 insertions(+), 5 deletions(-)
> This is not the latest version of the patch (that one introduced some
> regressions):
>
> http://patchwork.ozlabs.org/patch/100492/
Thanks. replaced the older patch with that one. Linux-user-for-upstream branch
has been refreshed with this patch updated and the commit message changes
you requested.
Riku
^ permalink raw reply [flat|nested] 23+ messages in thread
* [Qemu-devel] [PATCH 08/18] linux-user: add pselect6 syscall support
2011-06-20 16:20 [Qemu-devel] [PATCH 00/18] pending linux-user patches riku.voipio
` (6 preceding siblings ...)
2011-06-20 16:20 ` [Qemu-devel] [PATCH 07/18] linux-user: Fix the computation of the requested heap size riku.voipio
@ 2011-06-20 16:20 ` riku.voipio
2011-06-20 16:20 ` [Qemu-devel] [PATCH 09/18] linux-user: Define AT_RANDOM to support target stack protection mechanism riku.voipio
` (9 subsequent siblings)
17 siblings, 0 replies; 23+ messages in thread
From: riku.voipio @ 2011-06-20 16:20 UTC (permalink / raw)
To: qemu-devel; +Cc: Mike Frysinger
From: Mike Frysinger <vapier@gentoo.org>
Some architectures (like Blackfin) only implement pselect6 (and skip
select/newselect). So add support for it.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Riku Voipio <riku.voipio@iki.fi>
---
linux-user/syscall.c | 149 +++++++++++++++++++++++++++++++++++++++++++------
1 files changed, 130 insertions(+), 19 deletions(-)
diff --git a/linux-user/syscall.c b/linux-user/syscall.c
index be27f53..362cc63 100644
--- a/linux-user/syscall.c
+++ b/linux-user/syscall.c
@@ -550,6 +550,15 @@ _syscall5(int, sys_ppoll, struct pollfd *, fds, nfds_t, nfds,
size_t, sigsetsize)
#endif
+#if defined(TARGET_NR_pselect6)
+#ifndef __NR_pselect6
+# define __NR_pselect6 -1
+#endif
+#define __NR_sys_pselect6 __NR_pselect6
+_syscall6(int, sys_pselect6, int, nfds, fd_set *, readfds, fd_set *, writefds,
+ fd_set *, exceptfds, struct timespec *, timeout, void *, sig);
+#endif
+
extern int personality(int);
extern int flock(int, int);
extern int setfsuid(int);
@@ -799,6 +808,20 @@ static inline abi_long copy_from_user_fdset(fd_set *fds,
return 0;
}
+static inline abi_ulong copy_from_user_fdset_ptr(fd_set *fds, fd_set **fds_ptr,
+ abi_ulong target_fds_addr,
+ int n)
+{
+ if (target_fds_addr) {
+ if (copy_from_user_fdset(fds, target_fds_addr, n))
+ return -TARGET_EFAULT;
+ *fds_ptr = fds;
+ } else {
+ *fds_ptr = NULL;
+ }
+ return 0;
+}
+
static inline abi_long copy_to_user_fdset(abi_ulong target_fds_addr,
const fd_set *fds,
int n)
@@ -964,6 +987,7 @@ static inline abi_long copy_to_user_mq_attr(abi_ulong target_mq_attr_addr,
}
#endif
+#if defined(TARGET_NR_select) || defined(TARGET_NR__newselect)
/* do_select() must return target values and target errnos. */
static abi_long do_select(int n,
abi_ulong rfd_addr, abi_ulong wfd_addr,
@@ -974,26 +998,17 @@ static abi_long do_select(int n,
struct timeval tv, *tv_ptr;
abi_long ret;
- if (rfd_addr) {
- if (copy_from_user_fdset(&rfds, rfd_addr, n))
- return -TARGET_EFAULT;
- rfds_ptr = &rfds;
- } else {
- rfds_ptr = NULL;
+ ret = copy_from_user_fdset_ptr(&rfds, &rfds_ptr, rfd_addr, n);
+ if (ret) {
+ return ret;
}
- if (wfd_addr) {
- if (copy_from_user_fdset(&wfds, wfd_addr, n))
- return -TARGET_EFAULT;
- wfds_ptr = &wfds;
- } else {
- wfds_ptr = NULL;
+ ret = copy_from_user_fdset_ptr(&wfds, &wfds_ptr, wfd_addr, n);
+ if (ret) {
+ return ret;
}
- if (efd_addr) {
- if (copy_from_user_fdset(&efds, efd_addr, n))
- return -TARGET_EFAULT;
- efds_ptr = &efds;
- } else {
- efds_ptr = NULL;
+ ret = copy_from_user_fdset_ptr(&efds, &efds_ptr, efd_addr, n);
+ if (ret) {
+ return ret;
}
if (target_tv_addr) {
@@ -1020,6 +1035,7 @@ static abi_long do_select(int n,
return ret;
}
+#endif
static abi_long do_pipe2(int host_pipe[], int flags)
{
@@ -5581,7 +5597,102 @@ abi_long do_syscall(void *cpu_env, int num, abi_long arg1,
#endif
#ifdef TARGET_NR_pselect6
case TARGET_NR_pselect6:
- goto unimplemented_nowarn;
+ {
+ abi_long rfd_addr, wfd_addr, efd_addr, n, ts_addr;
+ fd_set rfds, wfds, efds;
+ fd_set *rfds_ptr, *wfds_ptr, *efds_ptr;
+ struct timespec ts, *ts_ptr;
+
+ /*
+ * The 6th arg is actually two args smashed together,
+ * so we cannot use the C library.
+ */
+ sigset_t set;
+ struct {
+ sigset_t *set;
+ size_t size;
+ } sig, *sig_ptr;
+
+ abi_ulong arg_sigset, arg_sigsize, *arg7;
+ target_sigset_t *target_sigset;
+
+ n = arg1;
+ rfd_addr = arg2;
+ wfd_addr = arg3;
+ efd_addr = arg4;
+ ts_addr = arg5;
+
+ ret = copy_from_user_fdset_ptr(&rfds, &rfds_ptr, rfd_addr, n);
+ if (ret) {
+ goto fail;
+ }
+ ret = copy_from_user_fdset_ptr(&wfds, &wfds_ptr, wfd_addr, n);
+ if (ret) {
+ goto fail;
+ }
+ ret = copy_from_user_fdset_ptr(&efds, &efds_ptr, efd_addr, n);
+ if (ret) {
+ goto fail;
+ }
+
+ /*
+ * This takes a timespec, and not a timeval, so we cannot
+ * use the do_select() helper ...
+ */
+ if (ts_addr) {
+ if (target_to_host_timespec(&ts, ts_addr)) {
+ goto efault;
+ }
+ ts_ptr = &ts;
+ } else {
+ ts_ptr = NULL;
+ }
+
+ /* Extract the two packed args for the sigset */
+ if (arg6) {
+ sig_ptr = &sig;
+ sig.size = _NSIG / 8;
+
+ arg7 = lock_user(VERIFY_READ, arg6, sizeof(*arg7) * 2, 1);
+ if (!arg7) {
+ goto efault;
+ }
+ arg_sigset = tswapl(arg7[0]);
+ arg_sigsize = tswapl(arg7[1]);
+ unlock_user(arg7, arg6, 0);
+
+ if (arg_sigset) {
+ sig.set = &set;
+ target_sigset = lock_user(VERIFY_READ, arg_sigset,
+ sizeof(*target_sigset), 1);
+ if (!target_sigset) {
+ goto efault;
+ }
+ target_to_host_sigset(&set, target_sigset);
+ unlock_user(target_sigset, arg_sigset, 0);
+ } else {
+ sig.set = NULL;
+ }
+ } else {
+ sig_ptr = NULL;
+ }
+
+ ret = get_errno(sys_pselect6(n, rfds_ptr, wfds_ptr, efds_ptr,
+ ts_ptr, sig_ptr));
+
+ if (!is_error(ret)) {
+ if (rfd_addr && copy_to_user_fdset(rfd_addr, &rfds, n))
+ goto efault;
+ if (wfd_addr && copy_to_user_fdset(wfd_addr, &wfds, n))
+ goto efault;
+ if (efd_addr && copy_to_user_fdset(efd_addr, &efds, n))
+ goto efault;
+
+ if (ts_addr && host_to_target_timespec(ts_addr, &ts))
+ goto efault;
+ }
+ }
+ break;
#endif
case TARGET_NR_symlink:
{
--
1.7.4.1
^ permalink raw reply related [flat|nested] 23+ messages in thread* [Qemu-devel] [PATCH 09/18] linux-user: Define AT_RANDOM to support target stack protection mechanism.
2011-06-20 16:20 [Qemu-devel] [PATCH 00/18] pending linux-user patches riku.voipio
` (7 preceding siblings ...)
2011-06-20 16:20 ` [Qemu-devel] [PATCH 08/18] linux-user: add pselect6 syscall support riku.voipio
@ 2011-06-20 16:20 ` riku.voipio
2011-06-21 7:05 ` cedric.vincent
2011-06-20 16:20 ` [Qemu-devel] [PATCH 10/18] linuxload: id_change was a write only variable riku.voipio
` (8 subsequent siblings)
17 siblings, 1 reply; 23+ messages in thread
From: riku.voipio @ 2011-06-20 16:20 UTC (permalink / raw)
To: qemu-devel; +Cc: Laurent ALFONSI
From: Laurent ALFONSI <laurent.alfonsi@st.com>
Note that the support for the command-line argument requires:
1. add the new field "uint8_t rand_bytes[16]" to "struct
image_info" since only the variable "info" lives both in
main() and in create_elf_tables()
2. write a dedicated parser to convert the command-line to fill
rand_bytes[]
These two steps aren't really hard to achieve but I finally think they
are a little bit overkill regarding the purpose of these 16 bytes.
Maybe we could always fill the 16 bytes pointed to by AT_RANDOM with
zero if we really want to get reproducibility.
Regards,
Cédric.
8<----8<----8<----8<----8<----8<----8<----8<----8<----8<----8<----8<----
The dynamic linker from the GNU C library v2.10+ uses the ELF
auxiliary vector AT_RANDOM [1] as a pointer to 16 bytes with random
values to initialize the stack protection mechanism. Technically the
emulated GNU dynamic linker crashes due to a NULL pointer
derefencement if it is built with stack protection enabled and if
AT_RANDOM is not defined by the QEMU ELF loader.
[1] This ELF auxiliary vector was introduced in Linux v2.6.29.
This patch can be tested with the code above:
#include <elf.h> /* Elf*_auxv_t, AT_RANDOM, */
#include <stdio.h> /* printf(3), */
#include <stdlib.h> /* exit(3), EXIT_*, */
#include <stdint.h> /* uint8_t, */
#include <string.h> /* memcpy(3), */
#if defined(__LP64__) || defined(__ILP64__) || defined(__LLP64__)
# define Elf_auxv_t Elf64_auxv_t
#else
# define Elf_auxv_t Elf32_auxv_t
#endif
main(int argc, char* argv[], char* envp[])
{
Elf_auxv_t *auxv;
/* *envp = NULL marks end of envp. */
while (*envp++ != NULL);
/* auxv->a_type = AT_NULL marks the end of auxv. */
for (auxv = (Elf_auxv_t *)envp; auxv->a_type != AT_NULL; auxv++) {
if (auxv->a_type == AT_RANDOM) {
int i;
uint8_t rand_bytes[16];
printf("AT_RANDOM is: 0x%x\n", auxv->a_un.a_val);
memcpy(rand_bytes, (const uint8_t *)auxv->a_un.a_val, sizeof(rand_bytes));
printf("it points to: ");
for (i = 0; i < 16; i++) {
printf("0x%02x ", rand_bytes[i]);
}
printf("\n");
exit(EXIT_SUCCESS);
}
}
exit(EXIT_FAILURE);
}
Changes introduced in v2 and v3:
* Fix typos + thinko (AT_RANDOM is used for stack canary, not for
ASLR)
* AT_RANDOM points to 16 random bytes stored inside the user
stack.
* Add a small test program.
Signed-off-by: Cédric VINCENT <cedric.vincent@st.com>
Signed-off-by: Laurent ALFONSI <laurent.alfonsi@st.com>
Signed-off-by: Riku Voipio <riku.voipio@iki.fi>
---
linux-user/elfload.c | 21 ++++++++++++++++++++-
1 files changed, 20 insertions(+), 1 deletions(-)
diff --git a/linux-user/elfload.c b/linux-user/elfload.c
index a13eb7b..b2746f2 100644
--- a/linux-user/elfload.c
+++ b/linux-user/elfload.c
@@ -927,7 +927,7 @@ struct exec
#define TARGET_ELF_PAGESTART(_v) ((_v) & ~(unsigned long)(TARGET_ELF_EXEC_PAGESIZE-1))
#define TARGET_ELF_PAGEOFFSET(_v) ((_v) & (TARGET_ELF_EXEC_PAGESIZE-1))
-#define DLINFO_ITEMS 12
+#define DLINFO_ITEMS 13
static inline void memcpy_fromfs(void * to, const void * from, unsigned long n)
{
@@ -1202,6 +1202,9 @@ static abi_ulong create_elf_tables(abi_ulong p, int argc, int envc,
{
abi_ulong sp;
int size;
+ int i;
+ abi_ulong u_rand_bytes;
+ uint8_t k_rand_bytes[16];
abi_ulong u_platform;
const char *k_platform;
const int n = sizeof(elf_addr_t);
@@ -1231,6 +1234,20 @@ static abi_ulong create_elf_tables(abi_ulong p, int argc, int envc,
/* FIXME - check return value of memcpy_to_target() for failure */
memcpy_to_target(sp, k_platform, len);
}
+
+ /*
+ * Generate 16 random bytes for userspace PRNG seeding (not
+ * cryptically secure but it's not the aim of QEMU).
+ */
+ srand((unsigned int) time(NULL));
+ for (i = 0; i < 16; i++) {
+ k_rand_bytes[i] = rand();
+ }
+ sp -= 16;
+ u_rand_bytes = sp;
+ /* FIXME - check return value of memcpy_to_target() for failure */
+ memcpy_to_target(sp, k_rand_bytes, 16);
+
/*
* Force 16 byte _final_ alignment here for generality.
*/
@@ -1271,6 +1288,8 @@ static abi_ulong create_elf_tables(abi_ulong p, int argc, int envc,
NEW_AUX_ENT(AT_EGID, (abi_ulong) getegid());
NEW_AUX_ENT(AT_HWCAP, (abi_ulong) ELF_HWCAP);
NEW_AUX_ENT(AT_CLKTCK, (abi_ulong) sysconf(_SC_CLK_TCK));
+ NEW_AUX_ENT(AT_RANDOM, (abi_ulong) u_rand_bytes);
+
if (k_platform)
NEW_AUX_ENT(AT_PLATFORM, u_platform);
#ifdef ARCH_DLINFO
--
1.7.4.1
^ permalink raw reply related [flat|nested] 23+ messages in thread* [Qemu-devel] [PATCH 10/18] linuxload: id_change was a write only variable
2011-06-20 16:20 [Qemu-devel] [PATCH 00/18] pending linux-user patches riku.voipio
` (8 preceding siblings ...)
2011-06-20 16:20 ` [Qemu-devel] [PATCH 09/18] linux-user: Define AT_RANDOM to support target stack protection mechanism riku.voipio
@ 2011-06-20 16:20 ` riku.voipio
2011-06-20 16:20 ` [Qemu-devel] [PATCH 11/18] syscall: really return ret code riku.voipio
` (7 subsequent siblings)
17 siblings, 0 replies; 23+ messages in thread
From: riku.voipio @ 2011-06-20 16:20 UTC (permalink / raw)
To: qemu-devel; +Cc: Juan Quintela
From: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@iki.fi>
---
linux-user/linuxload.c | 25 +------------------------
1 files changed, 1 insertions(+), 24 deletions(-)
diff --git a/linux-user/linuxload.c b/linux-user/linuxload.c
index ac8c486..62ebc7e 100644
--- a/linux-user/linuxload.c
+++ b/linux-user/linuxload.c
@@ -26,22 +26,6 @@ abi_long memcpy_to_target(abi_ulong dest, const void *src,
return 0;
}
-static int in_group_p(gid_t g)
-{
- /* return TRUE if we're in the specified group, FALSE otherwise */
- int ngroup;
- int i;
- gid_t grouplist[NGROUPS];
-
- ngroup = getgroups(NGROUPS, grouplist);
- for(i = 0; i < ngroup; i++) {
- if(grouplist[i] == g) {
- return 1;
- }
- }
- return 0;
-}
-
static int count(char ** vec)
{
int i;
@@ -57,7 +41,7 @@ static int prepare_binprm(struct linux_binprm *bprm)
{
struct stat st;
int mode;
- int retval, id_change;
+ int retval;
if(fstat(bprm->fd, &st) < 0) {
return(-errno);
@@ -73,14 +57,10 @@ static int prepare_binprm(struct linux_binprm *bprm)
bprm->e_uid = geteuid();
bprm->e_gid = getegid();
- id_change = 0;
/* Set-uid? */
if(mode & S_ISUID) {
bprm->e_uid = st.st_uid;
- if(bprm->e_uid != geteuid()) {
- id_change = 1;
- }
}
/* Set-gid? */
@@ -91,9 +71,6 @@ static int prepare_binprm(struct linux_binprm *bprm)
*/
if ((mode & (S_ISGID | S_IXGRP)) == (S_ISGID | S_IXGRP)) {
bprm->e_gid = st.st_gid;
- if (!in_group_p(bprm->e_gid)) {
- id_change = 1;
- }
}
retval = read(bprm->fd, bprm->buf, BPRM_BUF_SIZE);
--
1.7.4.1
^ permalink raw reply related [flat|nested] 23+ messages in thread* [Qemu-devel] [PATCH 11/18] syscall: really return ret code
2011-06-20 16:20 [Qemu-devel] [PATCH 00/18] pending linux-user patches riku.voipio
` (9 preceding siblings ...)
2011-06-20 16:20 ` [Qemu-devel] [PATCH 10/18] linuxload: id_change was a write only variable riku.voipio
@ 2011-06-20 16:20 ` riku.voipio
2011-06-20 16:20 ` [Qemu-devel] [PATCH 12/18] linux-user: syscall should use sanitized arg1 riku.voipio
` (6 subsequent siblings)
17 siblings, 0 replies; 23+ messages in thread
From: riku.voipio @ 2011-06-20 16:20 UTC (permalink / raw)
To: qemu-devel; +Cc: Juan Quintela
From: Juan Quintela <quintela@redhat.com>
We assign ret with the error code, but then return 0 unconditionally.
Signed-off-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@iki.fi>
---
linux-user/syscall.c | 8 ++++----
1 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/linux-user/syscall.c b/linux-user/syscall.c
index 362cc63..57d9233 100644
--- a/linux-user/syscall.c
+++ b/linux-user/syscall.c
@@ -3779,10 +3779,10 @@ static abi_long do_get_thread_area(CPUX86State *env, abi_ulong ptr)
#ifndef TARGET_ABI32
static abi_long do_arch_prctl(CPUX86State *env, int code, abi_ulong addr)
{
- abi_long ret;
+ abi_long ret = 0;
abi_ulong val;
int idx;
-
+
switch(code) {
case TARGET_ARCH_SET_GS:
case TARGET_ARCH_SET_FS:
@@ -3801,13 +3801,13 @@ static abi_long do_arch_prctl(CPUX86State *env, int code, abi_ulong addr)
idx = R_FS;
val = env->segs[idx].base;
if (put_user(val, addr, abi_ulong))
- return -TARGET_EFAULT;
+ ret = -TARGET_EFAULT;
break;
default:
ret = -TARGET_EINVAL;
break;
}
- return 0;
+ return ret;
}
#endif
--
1.7.4.1
^ permalink raw reply related [flat|nested] 23+ messages in thread* [Qemu-devel] [PATCH 12/18] linux-user: syscall should use sanitized arg1
2011-06-20 16:20 [Qemu-devel] [PATCH 00/18] pending linux-user patches riku.voipio
` (10 preceding siblings ...)
2011-06-20 16:20 ` [Qemu-devel] [PATCH 11/18] syscall: really return ret code riku.voipio
@ 2011-06-20 16:20 ` riku.voipio
2011-06-20 16:20 ` [Qemu-devel] [PATCH 13/18] flatload: end_code was only used in a debug message riku.voipio
` (5 subsequent siblings)
17 siblings, 0 replies; 23+ messages in thread
From: riku.voipio @ 2011-06-20 16:20 UTC (permalink / raw)
To: qemu-devel; +Cc: Juan Quintela
From: Juan Quintela <quintela@redhat.com>
Looking at the other architectures, we should be using "how" not "arg1".
Signed-off-by: Juan Quintela <quintela@redhat.com>
[peter.maydell@linaro.org: remove unnecessary initialisation of how]
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@iki.fi>
---
linux-user/syscall.c | 4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/linux-user/syscall.c b/linux-user/syscall.c
index 57d9233..1c0503f 100644
--- a/linux-user/syscall.c
+++ b/linux-user/syscall.c
@@ -7181,7 +7181,7 @@ abi_long do_syscall(void *cpu_env, int num, abi_long arg1,
case TARGET_NR_osf_sigprocmask:
{
abi_ulong mask;
- int how = arg1;
+ int how;
sigset_t set, oldset;
switch(arg1) {
@@ -7200,7 +7200,7 @@ abi_long do_syscall(void *cpu_env, int num, abi_long arg1,
}
mask = arg2;
target_to_host_old_sigset(&set, &mask);
- sigprocmask(arg1, &set, &oldset);
+ sigprocmask(how, &set, &oldset);
host_to_target_old_sigset(&mask, &oldset);
ret = mask;
}
--
1.7.4.1
^ permalink raw reply related [flat|nested] 23+ messages in thread* [Qemu-devel] [PATCH 13/18] flatload: end_code was only used in a debug message
2011-06-20 16:20 [Qemu-devel] [PATCH 00/18] pending linux-user patches riku.voipio
` (11 preceding siblings ...)
2011-06-20 16:20 ` [Qemu-devel] [PATCH 12/18] linux-user: syscall should use sanitized arg1 riku.voipio
@ 2011-06-20 16:20 ` riku.voipio
2011-06-20 16:20 ` [Qemu-devel] [PATCH 14/18] flatload: memp was a write-only variable riku.voipio
` (4 subsequent siblings)
17 siblings, 0 replies; 23+ messages in thread
From: riku.voipio @ 2011-06-20 16:20 UTC (permalink / raw)
To: qemu-devel; +Cc: Juan Quintela
From: Juan Quintela <quintela@redhat.com>
Just unfold its definition in only use.
Signed-off-by: Juan Quintela <quintela@redhat.com>
[peter.maydell@linaro.org: fixed typo in the debug code,
added parentheses to fix precedence issue]
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@iki.fi>
---
linux-user/flatload.c | 5 ++---
1 files changed, 2 insertions(+), 3 deletions(-)
diff --git a/linux-user/flatload.c b/linux-user/flatload.c
index cd7af7c..6fb78f5 100644
--- a/linux-user/flatload.c
+++ b/linux-user/flatload.c
@@ -384,7 +384,7 @@ static int load_flat_file(struct linux_binprm * bprm,
abi_ulong reloc = 0, rp;
int i, rev, relocs = 0;
abi_ulong fpos;
- abi_ulong start_code, end_code;
+ abi_ulong start_code;
abi_ulong indx_len;
hdr = ((struct flat_hdr *) bprm->buf); /* exec-header */
@@ -552,11 +552,10 @@ static int load_flat_file(struct linux_binprm * bprm,
/* The main program needs a little extra setup in the task structure */
start_code = textpos + sizeof (struct flat_hdr);
- end_code = textpos + text_len;
DBG_FLT("%s %s: TEXT=%x-%x DATA=%x-%x BSS=%x-%x\n",
id ? "Lib" : "Load", bprm->filename,
- (int) start_code, (int) end_code,
+ (int) start_code, (int) (textpos + text_len),
(int) datapos,
(int) (datapos + data_len),
(int) (datapos + data_len),
--
1.7.4.1
^ permalink raw reply related [flat|nested] 23+ messages in thread* [Qemu-devel] [PATCH 14/18] flatload: memp was a write-only variable
2011-06-20 16:20 [Qemu-devel] [PATCH 00/18] pending linux-user patches riku.voipio
` (12 preceding siblings ...)
2011-06-20 16:20 ` [Qemu-devel] [PATCH 13/18] flatload: end_code was only used in a debug message riku.voipio
@ 2011-06-20 16:20 ` riku.voipio
2011-06-20 16:20 ` [Qemu-devel] [PATCH 15/18] linux-user: Bump do_syscall() up to 8 syscall arguments riku.voipio
` (3 subsequent siblings)
17 siblings, 0 replies; 23+ messages in thread
From: riku.voipio @ 2011-06-20 16:20 UTC (permalink / raw)
To: qemu-devel; +Cc: Juan Quintela
From: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@iki.fi>
---
linux-user/flatload.c | 3 ---
1 files changed, 0 insertions(+), 3 deletions(-)
diff --git a/linux-user/flatload.c b/linux-user/flatload.c
index 6fb78f5..1062da3 100644
--- a/linux-user/flatload.c
+++ b/linux-user/flatload.c
@@ -379,7 +379,6 @@ static int load_flat_file(struct linux_binprm * bprm,
abi_long result;
abi_ulong realdatastart = 0;
abi_ulong text_len, data_len, bss_len, stack_len, flags;
- abi_ulong memp = 0; /* for finding the brk area */
abi_ulong extra;
abi_ulong reloc = 0, rp;
int i, rev, relocs = 0;
@@ -491,7 +490,6 @@ static int load_flat_file(struct linux_binprm * bprm,
}
reloc = datapos + (ntohl(hdr->reloc_start) - text_len);
- memp = realdatastart;
} else {
@@ -506,7 +504,6 @@ static int load_flat_file(struct linux_binprm * bprm,
realdatastart = textpos + ntohl(hdr->data_start);
datapos = realdatastart + indx_len;
reloc = (textpos + ntohl(hdr->reloc_start) + indx_len);
- memp = textpos;
#ifdef CONFIG_BINFMT_ZFLAT
#error code needs checking
--
1.7.4.1
^ permalink raw reply related [flat|nested] 23+ messages in thread* [Qemu-devel] [PATCH 15/18] linux-user: Bump do_syscall() up to 8 syscall arguments
2011-06-20 16:20 [Qemu-devel] [PATCH 00/18] pending linux-user patches riku.voipio
` (13 preceding siblings ...)
2011-06-20 16:20 ` [Qemu-devel] [PATCH 14/18] flatload: memp was a write-only variable riku.voipio
@ 2011-06-20 16:20 ` riku.voipio
2011-06-20 16:20 ` [Qemu-devel] [PATCH 16/18] linux-user/signal.c: Remove only-ever-set variable fpu_save_addr riku.voipio
` (2 subsequent siblings)
17 siblings, 0 replies; 23+ messages in thread
From: riku.voipio @ 2011-06-20 16:20 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell
From: Peter Maydell <peter.maydell@linaro.org>
On 32 bit MIPS a few syscalls have 7 arguments, and so to call
them via NR_syscall the guest needs to be able to pass 8 arguments
to do_syscall(). Raise the number of arguments do_syscall() takes
accordingly.
This fixes some gcc 4.6 compiler warnings about arg7 and arg8
variables being set and never used.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@iki.fi>
---
linux-user/main.c | 37 ++++++++++++++++++++++++-------------
linux-user/qemu.h | 3 ++-
linux-user/syscall.c | 8 +++++---
3 files changed, 31 insertions(+), 17 deletions(-)
diff --git a/linux-user/main.c b/linux-user/main.c
index 71dd253..1293450 100644
--- a/linux-user/main.c
+++ b/linux-user/main.c
@@ -319,7 +319,8 @@ void cpu_loop(CPUX86State *env)
env->regs[R_EDX],
env->regs[R_ESI],
env->regs[R_EDI],
- env->regs[R_EBP]);
+ env->regs[R_EBP],
+ 0, 0);
break;
#ifndef TARGET_ABI32
case EXCP_SYSCALL:
@@ -331,7 +332,8 @@ void cpu_loop(CPUX86State *env)
env->regs[R_EDX],
env->regs[10],
env->regs[8],
- env->regs[9]);
+ env->regs[9],
+ 0, 0);
env->eip = env->exception_next_eip;
break;
#endif
@@ -735,7 +737,8 @@ void cpu_loop(CPUARMState *env)
env->regs[2],
env->regs[3],
env->regs[4],
- env->regs[5]);
+ env->regs[5],
+ 0, 0);
}
} else {
goto error;
@@ -831,7 +834,8 @@ void cpu_loop(CPUState *env)
env->regs[2],
env->regs[3],
env->regs[4],
- env->regs[5]);
+ env->regs[5],
+ 0, 0);
}
} else {
goto error;
@@ -1018,7 +1022,8 @@ void cpu_loop (CPUSPARCState *env)
ret = do_syscall (env, env->gregs[1],
env->regwptr[0], env->regwptr[1],
env->regwptr[2], env->regwptr[3],
- env->regwptr[4], env->regwptr[5]);
+ env->regwptr[4], env->regwptr[5],
+ 0, 0);
if ((abi_ulong)ret >= (abi_ulong)(-515)) {
#if defined(TARGET_SPARC64) && !defined(TARGET_ABI32)
env->xcc |= PSR_CARRY;
@@ -1611,7 +1616,7 @@ void cpu_loop(CPUPPCState *env)
env->crf[0] &= ~0x1;
ret = do_syscall(env, env->gpr[0], env->gpr[3], env->gpr[4],
env->gpr[5], env->gpr[6], env->gpr[7],
- env->gpr[8]);
+ env->gpr[8], 0, 0);
if (ret == (uint32_t)(-TARGET_QEMU_ESIGRETURN)) {
/* Returning from a successful sigreturn syscall.
Avoid corrupting register state. */
@@ -2072,7 +2077,7 @@ void cpu_loop(CPUMIPSState *env)
env->active_tc.gpr[5],
env->active_tc.gpr[6],
env->active_tc.gpr[7],
- arg5, arg6/*, arg7, arg8*/);
+ arg5, arg6, arg7, arg8);
}
if (ret == -TARGET_QEMU_ESIGRETURN) {
/* Returning from a successful sigreturn syscall.
@@ -2160,7 +2165,8 @@ void cpu_loop (CPUState *env)
env->gregs[6],
env->gregs[7],
env->gregs[0],
- env->gregs[1]);
+ env->gregs[1],
+ 0, 0);
env->gregs[0] = ret;
break;
case EXCP_INTERRUPT:
@@ -2229,7 +2235,8 @@ void cpu_loop (CPUState *env)
env->regs[12],
env->regs[13],
env->pregs[7],
- env->pregs[11]);
+ env->pregs[11],
+ 0, 0);
env->regs[10] = ret;
break;
case EXCP_DEBUG:
@@ -2288,7 +2295,8 @@ void cpu_loop (CPUState *env)
env->regs[7],
env->regs[8],
env->regs[9],
- env->regs[10]);
+ env->regs[10],
+ 0, 0);
env->regs[3] = ret;
env->sregs[SR_PC] = env->regs[14];
break;
@@ -2398,7 +2406,8 @@ void cpu_loop(CPUM68KState *env)
env->dregs[3],
env->dregs[4],
env->dregs[5],
- env->aregs[0]);
+ env->aregs[0],
+ 0, 0);
}
break;
case EXCP_INTERRUPT:
@@ -2576,7 +2585,8 @@ void cpu_loop (CPUState *env)
sysret = do_syscall(env, trapnr,
env->ir[IR_A0], env->ir[IR_A1],
env->ir[IR_A2], env->ir[IR_A3],
- env->ir[IR_A4], env->ir[IR_A5]);
+ env->ir[IR_A4], env->ir[IR_A5],
+ 0, 0);
if (trapnr == TARGET_NR_sigreturn
|| trapnr == TARGET_NR_rt_sigreturn) {
break;
@@ -2707,7 +2717,8 @@ void cpu_loop(CPUS390XState *env)
env->regs[4],
env->regs[5],
env->regs[6],
- env->regs[7]);
+ env->regs[7],
+ 0, 0);
}
break;
case EXCP_ADDR:
diff --git a/linux-user/qemu.h b/linux-user/qemu.h
index 237386c..627c8b3 100644
--- a/linux-user/qemu.h
+++ b/linux-user/qemu.h
@@ -192,7 +192,8 @@ abi_long do_brk(abi_ulong new_brk);
void syscall_init(void);
abi_long do_syscall(void *cpu_env, int num, abi_long arg1,
abi_long arg2, abi_long arg3, abi_long arg4,
- abi_long arg5, abi_long arg6);
+ abi_long arg5, abi_long arg6, abi_long arg7,
+ abi_long arg8);
void gemu_log(const char *fmt, ...) GCC_FMT_ATTR(1, 2);
extern THREAD CPUState *thread_env;
void cpu_loop(CPUState *env);
diff --git a/linux-user/syscall.c b/linux-user/syscall.c
index 1c0503f..aa11a2c 100644
--- a/linux-user/syscall.c
+++ b/linux-user/syscall.c
@@ -4512,7 +4512,8 @@ int get_osversion(void)
All errnos that do_syscall() returns must be -TARGET_<errcode>. */
abi_long do_syscall(void *cpu_env, int num, abi_long arg1,
abi_long arg2, abi_long arg3, abi_long arg4,
- abi_long arg5, abi_long arg6)
+ abi_long arg5, abi_long arg6, abi_long arg7,
+ abi_long arg8)
{
abi_long ret;
struct stat st;
@@ -6152,8 +6153,9 @@ abi_long do_syscall(void *cpu_env, int num, abi_long arg1,
#endif
#ifdef TARGET_NR_syscall
case TARGET_NR_syscall:
- ret = do_syscall(cpu_env,arg1 & 0xffff,arg2,arg3,arg4,arg5,arg6,0);
- break;
+ ret = do_syscall(cpu_env, arg1 & 0xffff, arg2, arg3, arg4, arg5,
+ arg6, arg7, arg8, 0);
+ break;
#endif
case TARGET_NR_wait4:
{
--
1.7.4.1
^ permalink raw reply related [flat|nested] 23+ messages in thread* [Qemu-devel] [PATCH 16/18] linux-user/signal.c: Remove only-ever-set variable fpu_save_addr
2011-06-20 16:20 [Qemu-devel] [PATCH 00/18] pending linux-user patches riku.voipio
` (14 preceding siblings ...)
2011-06-20 16:20 ` [Qemu-devel] [PATCH 15/18] linux-user: Bump do_syscall() up to 8 syscall arguments riku.voipio
@ 2011-06-20 16:20 ` riku.voipio
2011-06-20 16:20 ` [Qemu-devel] [PATCH 17/18] linux-user/signal.c: Remove unused fenab riku.voipio
2011-06-20 16:20 ` [Qemu-devel] [PATCH 18/18] linux-user: Fix sync_file_range on 32bit mips riku.voipio
17 siblings, 0 replies; 23+ messages in thread
From: riku.voipio @ 2011-06-20 16:20 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell
From: Peter Maydell <peter.maydell@linaro.org>
Move the access of fpu_save into the commented out skeleton code for
restoring FPU registers on SPARC sigreturn, thus silencing a gcc
4.6 "variable set but never used" warning.
(This doesn't affect the calculation of 'err' because in fact
__get_user() can never fail.)
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@iki.fi>
---
linux-user/signal.c | 10 +++++-----
1 files changed, 5 insertions(+), 5 deletions(-)
diff --git a/linux-user/signal.c b/linux-user/signal.c
index cb7138f..4edd974 100644
--- a/linux-user/signal.c
+++ b/linux-user/signal.c
@@ -2080,7 +2080,6 @@ long do_sigreturn(CPUState *env)
uint32_t up_psr, pc, npc;
target_sigset_t set;
sigset_t host_set;
- abi_ulong fpu_save_addr;
int err, i;
sf_addr = env->regwptr[UREG_FP];
@@ -2120,10 +2119,11 @@ long do_sigreturn(CPUState *env)
err |= __get_user(env->regwptr[i + UREG_I0], &sf->info.si_regs.u_regs[i+8]);
}
- err |= __get_user(fpu_save_addr, &sf->fpu_save);
-
- //if (fpu_save)
- // err |= restore_fpu_state(env, fpu_save);
+ /* FIXME: implement FPU save/restore:
+ * __get_user(fpu_save, &sf->fpu_save);
+ * if (fpu_save)
+ * err |= restore_fpu_state(env, fpu_save);
+ */
/* This is pretty much atomic, no amount locking would prevent
* the races which exist anyways.
--
1.7.4.1
^ permalink raw reply related [flat|nested] 23+ messages in thread* [Qemu-devel] [PATCH 17/18] linux-user/signal.c: Remove unused fenab
2011-06-20 16:20 [Qemu-devel] [PATCH 00/18] pending linux-user patches riku.voipio
` (15 preceding siblings ...)
2011-06-20 16:20 ` [Qemu-devel] [PATCH 16/18] linux-user/signal.c: Remove only-ever-set variable fpu_save_addr riku.voipio
@ 2011-06-20 16:20 ` riku.voipio
2011-06-20 16:20 ` [Qemu-devel] [PATCH 18/18] linux-user: Fix sync_file_range on 32bit mips riku.voipio
17 siblings, 0 replies; 23+ messages in thread
From: riku.voipio @ 2011-06-20 16:20 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell
From: Peter Maydell <peter.maydell@linaro.org>
Remove fenab as it is only written, never used. Add a FIXME
comment about the discrepancy between our behaviour and that
of the Linux kernel for this routine.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@iki.fi>
---
linux-user/signal.c | 7 +++++--
1 files changed, 5 insertions(+), 2 deletions(-)
diff --git a/linux-user/signal.c b/linux-user/signal.c
index 4edd974..7d168e1 100644
--- a/linux-user/signal.c
+++ b/linux-user/signal.c
@@ -2228,7 +2228,6 @@ void sparc64_set_context(CPUSPARCState *env)
target_mc_gregset_t *grp;
abi_ulong pc, npc, tstate;
abi_ulong fp, i7, w_addr;
- unsigned char fenab;
int err;
unsigned int i;
@@ -2293,7 +2292,11 @@ void sparc64_set_context(CPUSPARCState *env)
if (put_user(i7, w_addr + offsetof(struct target_reg_window, ins[7]),
abi_ulong) != 0)
goto do_sigsegv;
- err |= __get_user(fenab, &(ucp->tuc_mcontext.mc_fpregs.mcfpu_enab));
+ /* FIXME this does not match how the kernel handles the FPU in
+ * its sparc64_set_context implementation. In particular the FPU
+ * is only restored if fenab is non-zero in:
+ * __get_user(fenab, &(ucp->tuc_mcontext.mc_fpregs.mcfpu_enab));
+ */
err |= __get_user(env->fprs, &(ucp->tuc_mcontext.mc_fpregs.mcfpu_fprs));
{
uint32_t *src, *dst;
--
1.7.4.1
^ permalink raw reply related [flat|nested] 23+ messages in thread* [Qemu-devel] [PATCH 18/18] linux-user: Fix sync_file_range on 32bit mips
2011-06-20 16:20 [Qemu-devel] [PATCH 00/18] pending linux-user patches riku.voipio
` (16 preceding siblings ...)
2011-06-20 16:20 ` [Qemu-devel] [PATCH 17/18] linux-user/signal.c: Remove unused fenab riku.voipio
@ 2011-06-20 16:20 ` riku.voipio
17 siblings, 0 replies; 23+ messages in thread
From: riku.voipio @ 2011-06-20 16:20 UTC (permalink / raw)
To: qemu-devel; +Cc: Riku Voipio
From: Riku Voipio <riku.voipio@iki.fi>
As noticed while looking at "Bump do_syscall() up to 8 syscall arguments"
patch, sync_file_range uses a pad argument on 32bit mips. Deal with it
by reading the correct arguments when on mips.
Signed-off-by: Riku Voipio <riku.voipio@iki.fi>
---
linux-user/syscall.c | 5 +++++
1 files changed, 5 insertions(+), 0 deletions(-)
diff --git a/linux-user/syscall.c b/linux-user/syscall.c
index aa11a2c..beb482c 100644
--- a/linux-user/syscall.c
+++ b/linux-user/syscall.c
@@ -7842,8 +7842,13 @@ abi_long do_syscall(void *cpu_env, int num, abi_long arg1,
#if defined(TARGET_NR_sync_file_range)
case TARGET_NR_sync_file_range:
#if TARGET_ABI_BITS == 32
+#if defined(TARGET_MIPS)
+ ret = get_errno(sync_file_range(arg1, target_offset64(arg3, arg4),
+ target_offset64(arg5, arg6), arg7));
+#else
ret = get_errno(sync_file_range(arg1, target_offset64(arg2, arg3),
target_offset64(arg4, arg5), arg6));
+#endif /* !TARGET_MIPS */
#else
ret = get_errno(sync_file_range(arg1, arg2, arg3, arg4));
#endif
--
1.7.4.1
^ permalink raw reply related [flat|nested] 23+ messages in thread