* [PATCH 2/9] ARM: traps: use get_kernel_nofault instead of set_fs()
2020-10-30 15:49 ` [PATCH 1/9] mm/maccess: fix unaligned copy_{from,to}_kernel_nofault Arnd Bergmann
@ 2020-10-30 15:49 ` Arnd Bergmann
2020-11-06 9:02 ` Linus Walleij
2020-10-30 15:49 ` [PATCH 3/9] ARM: oabi-compat: add epoll_pwait handler Arnd Bergmann
` (7 subsequent siblings)
8 siblings, 1 reply; 18+ messages in thread
From: Arnd Bergmann @ 2020-10-30 15:49 UTC (permalink / raw)
To: Russell King, Christoph Hellwig
Cc: linux-kernel, linux-arm-kernel, linux-arch, linux-mm, viro,
linus.walleij, arnd
From: Arnd Bergmann <arnd@arndb.de>
ARM uses set_fs() and __get_user() to allow the stack dumping code to
access possibly invalid pointers carefully. These can be changed to the
simpler get_kernel_nofault(), and allow the eventual removal of set_fs().
dump_instr() will print either kernel or user space pointers,
depending on how it was called. For dump_mem(), I assume we are only
interested in kernel pointers, and the only time that this is called
with user_mode(regs)==true is when the regs themselves are unreliable
as a result of the condition that caused the trap.
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
---
arch/arm/kernel/traps.c | 47 ++++++++++++++---------------------------
1 file changed, 16 insertions(+), 31 deletions(-)
diff --git a/arch/arm/kernel/traps.c b/arch/arm/kernel/traps.c
index 17d5a785df28..c3964a283b63 100644
--- a/arch/arm/kernel/traps.c
+++ b/arch/arm/kernel/traps.c
@@ -122,17 +122,8 @@ static void dump_mem(const char *lvl, const char *str, unsigned long bottom,
unsigned long top)
{
unsigned long first;
- mm_segment_t fs;
int i;
- /*
- * We need to switch to kernel mode so that we can use __get_user
- * to safely read from kernel space. Note that we now dump the
- * code first, just in case the backtrace kills us.
- */
- fs = get_fs();
- set_fs(KERNEL_DS);
-
printk("%s%s(0x%08lx to 0x%08lx)\n", lvl, str, bottom, top);
for (first = bottom & ~31; first < top; first += 32) {
@@ -145,7 +136,7 @@ static void dump_mem(const char *lvl, const char *str, unsigned long bottom,
for (p = first, i = 0; i < 8 && p < top; i++, p += 4) {
if (p >= bottom && p < top) {
unsigned long val;
- if (__get_user(val, (unsigned long *)p) == 0)
+ if (get_kernel_nofault(val, (unsigned long *)p))
sprintf(str + i * 9, " %08lx", val);
else
sprintf(str + i * 9, " ????????");
@@ -153,11 +144,9 @@ static void dump_mem(const char *lvl, const char *str, unsigned long bottom,
}
printk("%s%04lx:%s\n", lvl, first & 0xffff, str);
}
-
- set_fs(fs);
}
-static void __dump_instr(const char *lvl, struct pt_regs *regs)
+static void dump_instr(const char *lvl, struct pt_regs *regs)
{
unsigned long addr = instruction_pointer(regs);
const int thumb = thumb_mode(regs);
@@ -173,10 +162,20 @@ static void __dump_instr(const char *lvl, struct pt_regs *regs)
for (i = -4; i < 1 + !!thumb; i++) {
unsigned int val, bad;
- if (thumb)
- bad = get_user(val, &((u16 *)addr)[i]);
- else
- bad = get_user(val, &((u32 *)addr)[i]);
+ if (!user_mode(regs)) {
+ if (thumb) {
+ u16 val16;
+ bad = get_kernel_nofault(val16, &((u16 *)addr)[i]);
+ val = val16;
+ } else {
+ bad = get_kernel_nofault(val, &((u32 *)addr)[i]);
+ }
+ } else {
+ if (thumb)
+ bad = get_user(val, &((u16 *)addr)[i]);
+ else
+ bad = get_user(val, &((u32 *)addr)[i]);
+ }
if (!bad)
p += sprintf(p, i == 0 ? "(%0*x) " : "%0*x ",
@@ -189,20 +188,6 @@ static void __dump_instr(const char *lvl, struct pt_regs *regs)
printk("%sCode: %s\n", lvl, str);
}
-static void dump_instr(const char *lvl, struct pt_regs *regs)
-{
- mm_segment_t fs;
-
- if (!user_mode(regs)) {
- fs = get_fs();
- set_fs(KERNEL_DS);
- __dump_instr(lvl, regs);
- set_fs(fs);
- } else {
- __dump_instr(lvl, regs);
- }
-}
-
#ifdef CONFIG_ARM_UNWIND
static inline void dump_backtrace(struct pt_regs *regs, struct task_struct *tsk,
const char *loglvl)
--
2.27.0
^ permalink raw reply related [flat|nested] 18+ messages in thread* Re: [PATCH 2/9] ARM: traps: use get_kernel_nofault instead of set_fs()
2020-10-30 15:49 ` [PATCH 2/9] ARM: traps: use get_kernel_nofault instead of set_fs() Arnd Bergmann
@ 2020-11-06 9:02 ` Linus Walleij
0 siblings, 0 replies; 18+ messages in thread
From: Linus Walleij @ 2020-11-06 9:02 UTC (permalink / raw)
To: Arnd Bergmann
Cc: Russell King, Christoph Hellwig, linux-kernel, Linux ARM,
Linux-Arch, Linux Memory Management List, Al Viro, Arnd Bergmann
On Fri, Oct 30, 2020 at 4:49 PM Arnd Bergmann <arnd@kernel.org> wrote:
> From: Arnd Bergmann <arnd@arndb.de>
>
> ARM uses set_fs() and __get_user() to allow the stack dumping code to
> access possibly invalid pointers carefully. These can be changed to the
> simpler get_kernel_nofault(), and allow the eventual removal of set_fs().
>
> dump_instr() will print either kernel or user space pointers,
> depending on how it was called. For dump_mem(), I assume we are only
> interested in kernel pointers, and the only time that this is called
> with user_mode(regs)==true is when the regs themselves are unreliable
> as a result of the condition that caused the trap.
>
> Reviewed-by: Christoph Hellwig <hch@lst.de>
> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Not to mention the drastically improved readability of the code,
as ARM developers no more needs to cross-reference the
x86 FS segment register to figure out what this might be
doing.
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Yours,
Linus Walleij
^ permalink raw reply [flat|nested] 18+ messages in thread
* [PATCH 3/9] ARM: oabi-compat: add epoll_pwait handler
2020-10-30 15:49 ` [PATCH 1/9] mm/maccess: fix unaligned copy_{from,to}_kernel_nofault Arnd Bergmann
2020-10-30 15:49 ` [PATCH 2/9] ARM: traps: use get_kernel_nofault instead of set_fs() Arnd Bergmann
@ 2020-10-30 15:49 ` Arnd Bergmann
2020-10-30 15:49 ` [PATCH 4/9] ARM: syscall: always store thread_info->syscall Arnd Bergmann
` (6 subsequent siblings)
8 siblings, 0 replies; 18+ messages in thread
From: Arnd Bergmann @ 2020-10-30 15:49 UTC (permalink / raw)
To: Russell King, Christoph Hellwig
Cc: linux-kernel, linux-arm-kernel, linux-arch, linux-mm, viro,
linus.walleij, arnd, stable
From: Arnd Bergmann <arnd@arndb.de>
The epoll_wait() syscall has a special version for OABI compat
mode to convert the arguments to the EABI structure layout
of the kernel. However, the later epoll_pwait() syscall was
added in arch/arm in linux-2.6.32 without this conversion.
Use the same kind of handler for both.
Fixes: 369842658a36 ("ARM: 5677/1: ARM support for TIF_RESTORE_SIGMASK/pselect6/ppoll/epoll_pwait")
Cc: stable@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
---
arch/arm/kernel/sys_oabi-compat.c | 37 ++++++++++++++++++++++++++++---
arch/arm/tools/syscall.tbl | 2 +-
2 files changed, 35 insertions(+), 4 deletions(-)
diff --git a/arch/arm/kernel/sys_oabi-compat.c b/arch/arm/kernel/sys_oabi-compat.c
index 0203e545bbc8..a2b1ae01e5bf 100644
--- a/arch/arm/kernel/sys_oabi-compat.c
+++ b/arch/arm/kernel/sys_oabi-compat.c
@@ -264,9 +264,8 @@ asmlinkage long sys_oabi_epoll_ctl(int epfd, int op, int fd,
return do_epoll_ctl(epfd, op, fd, &kernel, false);
}
-asmlinkage long sys_oabi_epoll_wait(int epfd,
- struct oabi_epoll_event __user *events,
- int maxevents, int timeout)
+static long do_oabi_epoll_wait(int epfd, struct oabi_epoll_event __user *events,
+ int maxevents, int timeout)
{
struct epoll_event *kbuf;
struct oabi_epoll_event e;
@@ -299,6 +298,38 @@ asmlinkage long sys_oabi_epoll_wait(int epfd,
return err ? -EFAULT : ret;
}
+SYSCALL_DEFINE4(oabi_epoll_wait, int, epfd,
+ struct oabi_epoll_event __user *, events,
+ int, maxevents, int, timeout)
+{
+ return do_oabi_epoll_wait(epfd, events, maxevents, timeout);
+}
+
+/*
+ * Implement the event wait interface for the eventpoll file. It is the kernel
+ * part of the user space epoll_pwait(2).
+ */
+SYSCALL_DEFINE6(oabi_epoll_pwait, int, epfd,
+ struct oabi_epoll_event __user *, events, int, maxevents,
+ int, timeout, const sigset_t __user *, sigmask,
+ size_t, sigsetsize)
+{
+ int error;
+
+ /*
+ * If the caller wants a certain signal mask to be set during the wait,
+ * we apply it here.
+ */
+ error = set_user_sigmask(sigmask, sigsetsize);
+ if (error)
+ return error;
+
+ error = do_oabi_epoll_wait(epfd, events, maxevents, timeout);
+ restore_saved_sigmask_unless(error == -EINTR);
+
+ return error;
+}
+
struct oabi_sembuf {
unsigned short sem_num;
short sem_op;
diff --git a/arch/arm/tools/syscall.tbl b/arch/arm/tools/syscall.tbl
index d056a548358e..330cf0aa04c4 100644
--- a/arch/arm/tools/syscall.tbl
+++ b/arch/arm/tools/syscall.tbl
@@ -360,7 +360,7 @@
343 common vmsplice sys_vmsplice
344 common move_pages sys_move_pages
345 common getcpu sys_getcpu
-346 common epoll_pwait sys_epoll_pwait
+346 common epoll_pwait sys_epoll_pwait sys_oabi_epoll_pwait
347 common kexec_load sys_kexec_load
348 common utimensat sys_utimensat_time32
349 common signalfd sys_signalfd
--
2.27.0
^ permalink raw reply related [flat|nested] 18+ messages in thread* [PATCH 4/9] ARM: syscall: always store thread_info->syscall
2020-10-30 15:49 ` [PATCH 1/9] mm/maccess: fix unaligned copy_{from,to}_kernel_nofault Arnd Bergmann
2020-10-30 15:49 ` [PATCH 2/9] ARM: traps: use get_kernel_nofault instead of set_fs() Arnd Bergmann
2020-10-30 15:49 ` [PATCH 3/9] ARM: oabi-compat: add epoll_pwait handler Arnd Bergmann
@ 2020-10-30 15:49 ` Arnd Bergmann
2020-10-30 16:53 ` Russell King - ARM Linux admin
2020-10-30 15:49 ` [PATCH 5/9] ARM: oabi-compat: rework epoll_wait/epoll_pwait emulation Arnd Bergmann
` (5 subsequent siblings)
8 siblings, 1 reply; 18+ messages in thread
From: Arnd Bergmann @ 2020-10-30 15:49 UTC (permalink / raw)
To: Russell King, Christoph Hellwig
Cc: linux-kernel, linux-arm-kernel, linux-arch, linux-mm, viro,
linus.walleij, arnd
From: Arnd Bergmann <arnd@arndb.de>
The system call number is used in a a couple of places, in particular
ptrace, seccomp and /proc/<pid>/syscall.
The last one apparently never worked reliably on ARM for tasks
that are not currently getting traced.
Storing the syscall number in the normal entry path makes it work,
as well as allowing us to see if the current system call is for
OABI compat mode, which is the next thing I want to hook into.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
---
arch/arm/include/asm/syscall.h | 5 ++++-
arch/arm/kernel/asm-offsets.c | 1 +
arch/arm/kernel/entry-common.S | 8 ++++++--
arch/arm/kernel/ptrace.c | 9 +++++----
4 files changed, 16 insertions(+), 7 deletions(-)
diff --git a/arch/arm/include/asm/syscall.h b/arch/arm/include/asm/syscall.h
index fd02761ba06c..89898497edd6 100644
--- a/arch/arm/include/asm/syscall.h
+++ b/arch/arm/include/asm/syscall.h
@@ -22,7 +22,10 @@ extern const unsigned long sys_call_table[];
static inline int syscall_get_nr(struct task_struct *task,
struct pt_regs *regs)
{
- return task_thread_info(task)->syscall;
+ if (IS_ENABLED(CONFIG_AEABI) && !IS_ENABLED(CONFIG_OABI_COMPAT))
+ return task_thread_info(task)->syscall;
+
+ return task_thread_info(task)->syscall & ~__NR_OABI_SYSCALL_BASE;
}
static inline void syscall_rollback(struct task_struct *task,
diff --git a/arch/arm/kernel/asm-offsets.c b/arch/arm/kernel/asm-offsets.c
index a1570c8bab25..97af6735172b 100644
--- a/arch/arm/kernel/asm-offsets.c
+++ b/arch/arm/kernel/asm-offsets.c
@@ -46,6 +46,7 @@ int main(void)
DEFINE(TI_CPU, offsetof(struct thread_info, cpu));
DEFINE(TI_CPU_DOMAIN, offsetof(struct thread_info, cpu_domain));
DEFINE(TI_CPU_SAVE, offsetof(struct thread_info, cpu_context));
+ DEFINE(TI_SYSCALL, offsetof(struct thread_info, syscall));
DEFINE(TI_USED_CP, offsetof(struct thread_info, used_cp));
DEFINE(TI_TP_VALUE, offsetof(struct thread_info, tp_value));
DEFINE(TI_FPSTATE, offsetof(struct thread_info, fpstate));
diff --git a/arch/arm/kernel/entry-common.S b/arch/arm/kernel/entry-common.S
index 271cb8a1eba1..9a76467bbb47 100644
--- a/arch/arm/kernel/entry-common.S
+++ b/arch/arm/kernel/entry-common.S
@@ -223,6 +223,7 @@ ENTRY(vector_swi)
/* saved_psr and saved_pc are now dead */
uaccess_disable tbl
+ get_thread_info tsk
adr tbl, sys_call_table @ load syscall table pointer
@@ -234,13 +235,17 @@ ENTRY(vector_swi)
* get the old ABI syscall table address.
*/
bics r10, r10, #0xff000000
+ strne r10, [tsk, #TI_SYSCALL]
+ streq scno, [tsk, #TI_SYSCALL]
eorne scno, r10, #__NR_OABI_SYSCALL_BASE
ldrne tbl, =sys_oabi_call_table
#elif !defined(CONFIG_AEABI)
bic scno, scno, #0xff000000 @ mask off SWI op-code
+ str scno, [tsk, #TI_SYSCALL]
eor scno, scno, #__NR_SYSCALL_BASE @ check OS number
+#else
+ str scno, [tsk, #TI_SYSCALL]
#endif
- get_thread_info tsk
/*
* Reload the registers that may have been corrupted on entry to
* the syscall assembly (by tracing or context tracking.)
@@ -285,7 +290,6 @@ ENDPROC(vector_swi)
* context switches, and waiting for our parent to respond.
*/
__sys_trace:
- mov r1, scno
add r0, sp, #S_OFF
bl syscall_trace_enter
mov scno, r0
diff --git a/arch/arm/kernel/ptrace.c b/arch/arm/kernel/ptrace.c
index 2771e682220b..683edb8b627d 100644
--- a/arch/arm/kernel/ptrace.c
+++ b/arch/arm/kernel/ptrace.c
@@ -25,6 +25,7 @@
#include <linux/tracehook.h>
#include <linux/unistd.h>
+#include <asm/syscall.h>
#include <asm/traps.h>
#define CREATE_TRACE_POINTS
@@ -885,9 +886,9 @@ static void tracehook_report_syscall(struct pt_regs *regs,
regs->ARM_ip = ip;
}
-asmlinkage int syscall_trace_enter(struct pt_regs *regs, int scno)
+asmlinkage int syscall_trace_enter(struct pt_regs *regs)
{
- current_thread_info()->syscall = scno;
+ int scno;
if (test_thread_flag(TIF_SYSCALL_TRACE))
tracehook_report_syscall(regs, PTRACE_SYSCALL_ENTER);
@@ -898,11 +899,11 @@ asmlinkage int syscall_trace_enter(struct pt_regs *regs, int scno)
return -1;
#else
/* XXX: remove this once OABI gets fixed */
- secure_computing_strict(current_thread_info()->syscall);
+ secure_computing_strict(syscall_get_nr(current, regs));
#endif
/* Tracer or seccomp may have changed syscall. */
- scno = current_thread_info()->syscall;
+ scno = syscall_get_nr(current, regs);
if (test_thread_flag(TIF_SYSCALL_TRACEPOINT))
trace_sys_enter(regs, scno);
--
2.27.0
^ permalink raw reply related [flat|nested] 18+ messages in thread* Re: [PATCH 4/9] ARM: syscall: always store thread_info->syscall
2020-10-30 15:49 ` [PATCH 4/9] ARM: syscall: always store thread_info->syscall Arnd Bergmann
@ 2020-10-30 16:53 ` Russell King - ARM Linux admin
2020-10-30 21:28 ` Arnd Bergmann
0 siblings, 1 reply; 18+ messages in thread
From: Russell King - ARM Linux admin @ 2020-10-30 16:53 UTC (permalink / raw)
To: Arnd Bergmann
Cc: Christoph Hellwig, linux-kernel, linux-arm-kernel, linux-arch,
linux-mm, viro, linus.walleij, arnd
On Fri, Oct 30, 2020 at 04:49:14PM +0100, Arnd Bergmann wrote:
> From: Arnd Bergmann <arnd@arndb.de>
>
> The system call number is used in a a couple of places, in particular
> ptrace, seccomp and /proc/<pid>/syscall.
>
> The last one apparently never worked reliably on ARM for tasks
> that are not currently getting traced.
>
> Storing the syscall number in the normal entry path makes it work,
> as well as allowing us to see if the current system call is for
> OABI compat mode, which is the next thing I want to hook into.
I'm not sure this patch is correct.
Tracing the existing code for OABI:
asmlinkage int syscall_trace_enter(struct pt_regs *regs, int scno)
{
current_thread_info()->syscall = scno;
/* Legacy ABI only. */
USER( ldr scno, [saved_pc, #-4] ) @ get SWI instruction
bic scno, scno, #0xff000000 @ mask off SWI op-code
eor scno, scno, #__NR_SYSCALL_BASE @ check OS number
tst r10, #_TIF_SYSCALL_WORK @ are we tracing syscalls?
bne __sys_trace
__sys_trace:
mov r1, scno
add r0, sp, #S_OFF
bl syscall_trace_enter
So, thread_info->syscall does not include __NR_SYSCALL_BASE. The
reason for this is the code that makes use of that via syscall_get_nr().
kernel/trace/trace_syscalls.c:
syscall_nr = trace_get_syscall_nr(current, regs);
if (syscall_nr < 0 || syscall_nr >= NR_syscalls)
return;
and NR_syscalls is the number of syscalls, which doesn't include the
__NR_SYSCALL_BASE offset.
So, I think this patch actually breaks OABI.
--
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTP is here! 40Mbps down 10Mbps up. Decent connectivity at last!
^ permalink raw reply [flat|nested] 18+ messages in thread* Re: [PATCH 4/9] ARM: syscall: always store thread_info->syscall
2020-10-30 16:53 ` Russell King - ARM Linux admin
@ 2020-10-30 21:28 ` Arnd Bergmann
0 siblings, 0 replies; 18+ messages in thread
From: Arnd Bergmann @ 2020-10-30 21:28 UTC (permalink / raw)
To: Russell King - ARM Linux admin
Cc: Christoph Hellwig, linux-kernel, Linux ARM, linux-arch, Linux-MM,
Al Viro, Linus Walleij, Arnd Bergmann
On Fri, Oct 30, 2020 at 5:53 PM Russell King - ARM Linux admin
<linux@armlinux.org.uk> wrote:
>
> On Fri, Oct 30, 2020 at 04:49:14PM +0100, Arnd Bergmann wrote:
> > From: Arnd Bergmann <arnd@arndb.de>
> >
> > The system call number is used in a a couple of places, in particular
> > ptrace, seccomp and /proc/<pid>/syscall.
> >
> > The last one apparently never worked reliably on ARM for tasks
> > that are not currently getting traced.
> >
> > Storing the syscall number in the normal entry path makes it work,
> > as well as allowing us to see if the current system call is for
> > OABI compat mode, which is the next thing I want to hook into.
>
> I'm not sure this patch is correct.
I'm not following where you still see a mismatch, I was hoping I
had fixed them all after your previous review :(
The thread_info->syscall entry should now consistently contain
__NR_SYSCALL_BASE on an EABI kernel, and all users of
that should consistently mask it out.
> Tracing the existing code for OABI:
>
> asmlinkage int syscall_trace_enter(struct pt_regs *regs, int scno)
> {
> current_thread_info()->syscall = scno;
This no longer stores to current_thread_info()->syscall but instead
reads the number from syscall_get_nr().
> /* Legacy ABI only. */
> USER( ldr scno, [saved_pc, #-4] ) @ get SWI instruction
> bic scno, scno, #0xff000000 @ mask off SWI op-code
> eor scno, scno, #__NR_SYSCALL_BASE @ check OS number
> tst r10, #_TIF_SYSCALL_WORK @ are we tracing syscalls?
> bne __sys_trace
>
> __sys_trace:
> mov r1, scno
> add r0, sp, #S_OFF
> bl syscall_trace_enter
>
> So, thread_info->syscall does not include __NR_SYSCALL_BASE. The
> reason for this is the code that makes use of that via syscall_get_nr().
> kernel/trace/trace_syscalls.c:
On both CONFIG_OABI_COMPAT and on !CONFIG_AEABI kernels,
I store the value before masking out __NR_SYSCALL_BASE
after my change. For EABI-only kernels there is no need for
the mask.
> syscall_nr = trace_get_syscall_nr(current, regs);
> if (syscall_nr < 0 || syscall_nr >= NR_syscalls)
> return;
>
> and NR_syscalls is the number of syscalls, which doesn't include the
> __NR_SYSCALL_BASE offset.
>
> So, I think this patch actually breaks OABI.
The value returned from trace_get_syscall_nr() is always in
the 0...NR_syscalls range without the __NR_SYSCALL_BASE
for a valid syscall. because of the added
static inline int syscall_get_nr(struct task_struct *task,
struct pt_regs *regs)
{
- return task_thread_info(task)->syscall;
+ return task_thread_info(task)->syscall & ~__NR_OABI_SYSCALL_BASE;
}
(plus the corresponding logic for OABI_COMPAT.
Which of the above do you think I got wrong?
Arnd
^ permalink raw reply [flat|nested] 18+ messages in thread
* [PATCH 5/9] ARM: oabi-compat: rework epoll_wait/epoll_pwait emulation
2020-10-30 15:49 ` [PATCH 1/9] mm/maccess: fix unaligned copy_{from,to}_kernel_nofault Arnd Bergmann
` (2 preceding siblings ...)
2020-10-30 15:49 ` [PATCH 4/9] ARM: syscall: always store thread_info->syscall Arnd Bergmann
@ 2020-10-30 15:49 ` Arnd Bergmann
2020-10-30 15:49 ` [PATCH 6/9] ARM: oabi-compat: rework sys_semtimedop emulation Arnd Bergmann
` (4 subsequent siblings)
8 siblings, 0 replies; 18+ messages in thread
From: Arnd Bergmann @ 2020-10-30 15:49 UTC (permalink / raw)
To: Russell King, Christoph Hellwig
Cc: linux-kernel, linux-arm-kernel, linux-arch, linux-mm, viro,
linus.walleij, arnd
From: Arnd Bergmann <arnd@arndb.de>
The epoll_wait() system call wrapper is one of the remaining users of
the set_fs() infrasturcture for Arm. Changing it to not require set_fs()
is rather complex unfortunately.
The approach I'm taking here is to allow architectures to override
the code that copies the output to user space, and let the oabi-compat
implementation check whether it is getting called from an EABI or OABI
system call based on the thread_info->syscall value.
The in_oabi_syscall() check here mirrors the in_compat_syscall() and
in_x32_syscall() helpers for 32-bit compat implementations on other
architectures.
Overall, the amount of code goes down, at least with the newly added
sys_oabi_epoll_pwait() helper getting removed again. The downside
is added complexity in the source code for the native implementation.
There should be no difference in runtime performance except for Arm
kernels with CONFIG_OABI_COMPAT enabled that now have to go through
an external function call to check which of the two variants to use.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
---
arch/arm/include/asm/syscall.h | 11 +++++
arch/arm/kernel/sys_oabi-compat.c | 75 +++++++------------------------
arch/arm/tools/syscall.tbl | 4 +-
fs/eventpoll.c | 5 +--
include/linux/eventpoll.h | 18 ++++++++
5 files changed, 49 insertions(+), 64 deletions(-)
diff --git a/arch/arm/include/asm/syscall.h b/arch/arm/include/asm/syscall.h
index 89898497edd6..9efb7b3384e5 100644
--- a/arch/arm/include/asm/syscall.h
+++ b/arch/arm/include/asm/syscall.h
@@ -28,6 +28,17 @@ static inline int syscall_get_nr(struct task_struct *task,
return task_thread_info(task)->syscall & ~__NR_OABI_SYSCALL_BASE;
}
+static inline bool __in_oabi_syscall(struct task_struct *task)
+{
+ return IS_ENABLED(CONFIG_OABI_COMPAT) &&
+ (task_thread_info(task)->syscall & __NR_OABI_SYSCALL_BASE);
+}
+
+static inline bool in_oabi_syscall(void)
+{
+ return __in_oabi_syscall(current);
+}
+
static inline void syscall_rollback(struct task_struct *task,
struct pt_regs *regs)
{
diff --git a/arch/arm/kernel/sys_oabi-compat.c b/arch/arm/kernel/sys_oabi-compat.c
index a2b1ae01e5bf..f9d8e5be6ba0 100644
--- a/arch/arm/kernel/sys_oabi-compat.c
+++ b/arch/arm/kernel/sys_oabi-compat.c
@@ -83,6 +83,8 @@
#include <linux/uaccess.h>
#include <linux/slab.h>
+#include <asm/syscall.h>
+
struct oldabi_stat64 {
unsigned long long st_dev;
unsigned int __pad1;
@@ -264,70 +266,25 @@ asmlinkage long sys_oabi_epoll_ctl(int epfd, int op, int fd,
return do_epoll_ctl(epfd, op, fd, &kernel, false);
}
-static long do_oabi_epoll_wait(int epfd, struct oabi_epoll_event __user *events,
- int maxevents, int timeout)
+struct epoll_event __user *
+epoll_put_uevent(__poll_t revents, __u64 data,
+ struct epoll_event __user *uevent)
{
- struct epoll_event *kbuf;
- struct oabi_epoll_event e;
- mm_segment_t fs;
- long ret, err, i;
+ if (in_oabi_syscall()) {
+ struct oabi_epoll_event __user *oevent = (void __user *)uevent;
- if (maxevents <= 0 ||
- maxevents > (INT_MAX/sizeof(*kbuf)) ||
- maxevents > (INT_MAX/sizeof(*events)))
- return -EINVAL;
- if (!access_ok(events, sizeof(*events) * maxevents))
- return -EFAULT;
- kbuf = kmalloc_array(maxevents, sizeof(*kbuf), GFP_KERNEL);
- if (!kbuf)
- return -ENOMEM;
- fs = get_fs();
- set_fs(KERNEL_DS);
- ret = sys_epoll_wait(epfd, kbuf, maxevents, timeout);
- set_fs(fs);
- err = 0;
- for (i = 0; i < ret; i++) {
- e.events = kbuf[i].events;
- e.data = kbuf[i].data;
- err = __copy_to_user(events, &e, sizeof(e));
- if (err)
- break;
- events++;
- }
- kfree(kbuf);
- return err ? -EFAULT : ret;
-}
+ if (__put_user(revents, &oevent->events) ||
+ __put_user(data, &oevent->data))
+ return NULL;
-SYSCALL_DEFINE4(oabi_epoll_wait, int, epfd,
- struct oabi_epoll_event __user *, events,
- int, maxevents, int, timeout)
-{
- return do_oabi_epoll_wait(epfd, events, maxevents, timeout);
-}
-
-/*
- * Implement the event wait interface for the eventpoll file. It is the kernel
- * part of the user space epoll_pwait(2).
- */
-SYSCALL_DEFINE6(oabi_epoll_pwait, int, epfd,
- struct oabi_epoll_event __user *, events, int, maxevents,
- int, timeout, const sigset_t __user *, sigmask,
- size_t, sigsetsize)
-{
- int error;
-
- /*
- * If the caller wants a certain signal mask to be set during the wait,
- * we apply it here.
- */
- error = set_user_sigmask(sigmask, sigsetsize);
- if (error)
- return error;
+ return (void __user *)(oevent+1);
+ }
- error = do_oabi_epoll_wait(epfd, events, maxevents, timeout);
- restore_saved_sigmask_unless(error == -EINTR);
+ if (__put_user(revents, &uevent->events) ||
+ __put_user(data, &uevent->data))
+ return NULL;
- return error;
+ return uevent+1;
}
struct oabi_sembuf {
diff --git a/arch/arm/tools/syscall.tbl b/arch/arm/tools/syscall.tbl
index 330cf0aa04c4..2dd99f86e8a5 100644
--- a/arch/arm/tools/syscall.tbl
+++ b/arch/arm/tools/syscall.tbl
@@ -266,7 +266,7 @@
249 common lookup_dcookie sys_lookup_dcookie
250 common epoll_create sys_epoll_create
251 common epoll_ctl sys_epoll_ctl sys_oabi_epoll_ctl
-252 common epoll_wait sys_epoll_wait sys_oabi_epoll_wait
+252 common epoll_wait sys_epoll_wait
253 common remap_file_pages sys_remap_file_pages
# 254 for set_thread_area
# 255 for get_thread_area
@@ -360,7 +360,7 @@
343 common vmsplice sys_vmsplice
344 common move_pages sys_move_pages
345 common getcpu sys_getcpu
-346 common epoll_pwait sys_epoll_pwait sys_oabi_epoll_pwait
+346 common epoll_pwait sys_epoll_pwait
347 common kexec_load sys_kexec_load
348 common utimensat sys_utimensat_time32
349 common signalfd sys_signalfd
diff --git a/fs/eventpoll.c b/fs/eventpoll.c
index 4df61129566d..a21e8dbbe567 100644
--- a/fs/eventpoll.c
+++ b/fs/eventpoll.c
@@ -1743,8 +1743,8 @@ static __poll_t ep_send_events_proc(struct eventpoll *ep, struct list_head *head
if (!revents)
continue;
- if (__put_user(revents, &uevent->events) ||
- __put_user(epi->event.data, &uevent->data)) {
+ uevent = epoll_put_uevent(revents, epi->event.data, uevent);
+ if (!uevent) {
list_add(&epi->rdllink, head);
ep_pm_stay_awake(epi);
if (!esed->res)
@@ -1752,7 +1752,6 @@ static __poll_t ep_send_events_proc(struct eventpoll *ep, struct list_head *head
return 0;
}
esed->res++;
- uevent++;
if (epi->event.events & EPOLLONESHOT)
epi->event.events &= EP_PRIVATE_BITS;
else if (!(epi->event.events & EPOLLET)) {
diff --git a/include/linux/eventpoll.h b/include/linux/eventpoll.h
index 8f000fada5a4..d1f13147ed69 100644
--- a/include/linux/eventpoll.h
+++ b/include/linux/eventpoll.h
@@ -77,4 +77,22 @@ static inline void eventpoll_release(struct file *file) {}
#endif
+#if defined(CONFIG_ARM) && defined(CONFIG_OABI_COMPAT)
+/* ARM OABI has an incompatible struct layout and needs a special handler */
+extern struct epoll_event __user *
+epoll_put_uevent(__poll_t revents, __u64 data,
+ struct epoll_event __user *uevent);
+#else
+static inline struct epoll_event __user *
+epoll_put_uevent(__poll_t revents, __u64 data,
+ struct epoll_event __user *uevent)
+{
+ if (__put_user(revents, &uevent->events) ||
+ __put_user(data, &uevent->data))
+ return NULL;
+
+ return uevent+1;
+}
+#endif
+
#endif /* #ifndef _LINUX_EVENTPOLL_H */
--
2.27.0
^ permalink raw reply related [flat|nested] 18+ messages in thread* [PATCH 6/9] ARM: oabi-compat: rework sys_semtimedop emulation
2020-10-30 15:49 ` [PATCH 1/9] mm/maccess: fix unaligned copy_{from,to}_kernel_nofault Arnd Bergmann
` (3 preceding siblings ...)
2020-10-30 15:49 ` [PATCH 5/9] ARM: oabi-compat: rework epoll_wait/epoll_pwait emulation Arnd Bergmann
@ 2020-10-30 15:49 ` Arnd Bergmann
2020-10-30 15:49 ` [PATCH 7/9] ARM: oabi-compat: rework fcntl64() emulation Arnd Bergmann
` (3 subsequent siblings)
8 siblings, 0 replies; 18+ messages in thread
From: Arnd Bergmann @ 2020-10-30 15:49 UTC (permalink / raw)
To: Russell King, Christoph Hellwig
Cc: linux-kernel, linux-arm-kernel, linux-arch, linux-mm, viro,
linus.walleij, arnd
From: Arnd Bergmann <arnd@arndb.de>
sys_oabi_semtimedop() is one of the last users of set_fs() on Arm. To
remove this one, expose the internal code of the actual implementation
that operates on a kernel pointer and call it directly after copying.
There should be no measurable impact on the normal execution of this
function, and it makes the overly long function a little shorter, which
may help readability.
While reworking the oabi version, make it behave a little more like
the native one, using kvmalloc_array() and restructure the code
flow in a similar way.
The naming of __do_semtimedop() is not very good, I hope someone can
come up with a better name.
One regression was spotted by kernel test robot <rong.a.chen@intel.com>
and fixed before the first mailing list submission.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
---
arch/arm/kernel/sys_oabi-compat.c | 38 ++++++++------
include/linux/syscalls.h | 3 ++
ipc/sem.c | 84 +++++++++++++++++++------------
3 files changed, 77 insertions(+), 48 deletions(-)
diff --git a/arch/arm/kernel/sys_oabi-compat.c b/arch/arm/kernel/sys_oabi-compat.c
index f9d8e5be6ba0..c3e63b73b6ae 100644
--- a/arch/arm/kernel/sys_oabi-compat.c
+++ b/arch/arm/kernel/sys_oabi-compat.c
@@ -80,6 +80,7 @@
#include <linux/socket.h>
#include <linux/net.h>
#include <linux/ipc.h>
+#include <linux/ipc_namespace.h>
#include <linux/uaccess.h>
#include <linux/slab.h>
@@ -294,46 +295,51 @@ struct oabi_sembuf {
unsigned short __pad;
};
+#define sc_semopm sem_ctls[2]
+
asmlinkage long sys_oabi_semtimedop(int semid,
struct oabi_sembuf __user *tsops,
unsigned nsops,
const struct old_timespec32 __user *timeout)
{
+ struct ipc_namespace *ns;
struct sembuf *sops;
- struct old_timespec32 local_timeout;
long err;
int i;
+ ns = current->nsproxy->ipc_ns;
+ if (nsops > ns->sc_semopm)
+ return -E2BIG;
if (nsops < 1 || nsops > SEMOPM)
return -EINVAL;
- if (!access_ok(tsops, sizeof(*tsops) * nsops))
- return -EFAULT;
- sops = kmalloc_array(nsops, sizeof(*sops), GFP_KERNEL);
+ sops = kvmalloc_array(nsops, sizeof(*sops), GFP_KERNEL);
if (!sops)
return -ENOMEM;
err = 0;
for (i = 0; i < nsops; i++) {
struct oabi_sembuf osb;
- err |= __copy_from_user(&osb, tsops, sizeof(osb));
+ err |= copy_from_user(&osb, tsops, sizeof(osb));
sops[i].sem_num = osb.sem_num;
sops[i].sem_op = osb.sem_op;
sops[i].sem_flg = osb.sem_flg;
tsops++;
}
- if (timeout) {
- /* copy this as well before changing domain protection */
- err |= copy_from_user(&local_timeout, timeout, sizeof(*timeout));
- timeout = &local_timeout;
- }
if (err) {
err = -EFAULT;
- } else {
- mm_segment_t fs = get_fs();
- set_fs(KERNEL_DS);
- err = sys_semtimedop_time32(semid, sops, nsops, timeout);
- set_fs(fs);
+ goto out;
+ }
+
+ if (timeout) {
+ struct timespec64 ts;
+ err = get_old_timespec32(&ts, timeout);
+ if (err)
+ goto out;
+ err = __do_semtimedop(semid, sops, nsops, &ts, ns);
+ goto out;
}
- kfree(sops);
+ err = __do_semtimedop(semid, sops, nsops, NULL, ns);
+out:
+ kvfree(sops);
return err;
}
diff --git a/include/linux/syscalls.h b/include/linux/syscalls.h
index 37bea07c12f2..fc7340fa4702 100644
--- a/include/linux/syscalls.h
+++ b/include/linux/syscalls.h
@@ -1342,6 +1342,9 @@ long ksys_old_shmctl(int shmid, int cmd, struct shmid_ds __user *buf);
long compat_ksys_semtimedop(int semid, struct sembuf __user *tsems,
unsigned int nsops,
const struct old_timespec32 __user *timeout);
+long __do_semtimedop(int semid, struct sembuf *tsems, unsigned int nsops,
+ const struct timespec64 *timeout,
+ struct ipc_namespace *ns);
int __sys_getsockopt(int fd, int level, int optname, char __user *optval,
int __user *optlen);
diff --git a/ipc/sem.c b/ipc/sem.c
index f6c30a85dadf..c7725031c2ba 100644
--- a/ipc/sem.c
+++ b/ipc/sem.c
@@ -1978,46 +1978,34 @@ static struct sem_undo *find_alloc_undo(struct ipc_namespace *ns, int semid)
return un;
}
-static long do_semtimedop(int semid, struct sembuf __user *tsops,
- unsigned nsops, const struct timespec64 *timeout)
+long __do_semtimedop(int semid, struct sembuf *sops,
+ unsigned nsops, const struct timespec64 *timeout,
+ struct ipc_namespace *ns)
{
int error = -EINVAL;
struct sem_array *sma;
- struct sembuf fast_sops[SEMOPM_FAST];
- struct sembuf *sops = fast_sops, *sop;
+ struct sembuf *sop;
struct sem_undo *un;
int max, locknum;
bool undos = false, alter = false, dupsop = false;
struct sem_queue queue;
unsigned long dup = 0, jiffies_left = 0;
- struct ipc_namespace *ns;
-
- ns = current->nsproxy->ipc_ns;
if (nsops < 1 || semid < 0)
return -EINVAL;
if (nsops > ns->sc_semopm)
return -E2BIG;
- if (nsops > SEMOPM_FAST) {
- sops = kvmalloc_array(nsops, sizeof(*sops), GFP_KERNEL);
- if (sops == NULL)
- return -ENOMEM;
- }
-
- if (copy_from_user(sops, tsops, nsops * sizeof(*tsops))) {
- error = -EFAULT;
- goto out_free;
- }
if (timeout) {
if (timeout->tv_sec < 0 || timeout->tv_nsec < 0 ||
timeout->tv_nsec >= 1000000000L) {
error = -EINVAL;
- goto out_free;
+ goto out;
}
jiffies_left = timespec64_to_jiffies(timeout);
}
+
max = 0;
for (sop = sops; sop < sops + nsops; sop++) {
unsigned long mask = 1ULL << ((sop->sem_num) % BITS_PER_LONG);
@@ -2046,7 +2034,7 @@ static long do_semtimedop(int semid, struct sembuf __user *tsops,
un = find_alloc_undo(ns, semid);
if (IS_ERR(un)) {
error = PTR_ERR(un);
- goto out_free;
+ goto out;
}
} else {
un = NULL;
@@ -2057,25 +2045,25 @@ static long do_semtimedop(int semid, struct sembuf __user *tsops,
if (IS_ERR(sma)) {
rcu_read_unlock();
error = PTR_ERR(sma);
- goto out_free;
+ goto out;
}
error = -EFBIG;
if (max >= sma->sem_nsems) {
rcu_read_unlock();
- goto out_free;
+ goto out;
}
error = -EACCES;
if (ipcperms(ns, &sma->sem_perm, alter ? S_IWUGO : S_IRUGO)) {
rcu_read_unlock();
- goto out_free;
+ goto out;
}
error = security_sem_semop(&sma->sem_perm, sops, nsops, alter);
if (error) {
rcu_read_unlock();
- goto out_free;
+ goto out;
}
error = -EIDRM;
@@ -2089,7 +2077,7 @@ static long do_semtimedop(int semid, struct sembuf __user *tsops,
* entangled here and why it's RMID race safe on comments at sem_lock()
*/
if (!ipc_valid_object(&sma->sem_perm))
- goto out_unlock_free;
+ goto out_unlock;
/*
* semid identifiers are not unique - find_alloc_undo may have
* allocated an undo structure, it was invalidated by an RMID
@@ -2098,7 +2086,7 @@ static long do_semtimedop(int semid, struct sembuf __user *tsops,
* "un" itself is guaranteed by rcu.
*/
if (un && un->semid == -1)
- goto out_unlock_free;
+ goto out_unlock;
queue.sops = sops;
queue.nsops = nsops;
@@ -2124,10 +2112,10 @@ static long do_semtimedop(int semid, struct sembuf __user *tsops,
rcu_read_unlock();
wake_up_q(&wake_q);
- goto out_free;
+ goto out;
}
if (error < 0) /* non-blocking error path */
- goto out_unlock_free;
+ goto out_unlock;
/*
* We need to sleep on this operation, so we put the current
@@ -2192,14 +2180,14 @@ static long do_semtimedop(int semid, struct sembuf __user *tsops,
if (error != -EINTR) {
/* see SEM_BARRIER_2 for purpose/pairing */
smp_acquire__after_ctrl_dep();
- goto out_free;
+ goto out;
}
rcu_read_lock();
locknum = sem_lock(sma, sops, nsops);
if (!ipc_valid_object(&sma->sem_perm))
- goto out_unlock_free;
+ goto out_unlock;
/*
* No necessity for any barrier: We are protect by sem_lock()
@@ -2211,7 +2199,7 @@ static long do_semtimedop(int semid, struct sembuf __user *tsops,
* Leave without unlink_queue(), but with sem_unlock().
*/
if (error != -EINTR)
- goto out_unlock_free;
+ goto out_unlock;
/*
* If an interrupt occurred we have to clean up the queue.
@@ -2222,13 +2210,45 @@ static long do_semtimedop(int semid, struct sembuf __user *tsops,
unlink_queue(sma, &queue);
-out_unlock_free:
+out_unlock:
sem_unlock(sma, locknum);
rcu_read_unlock();
+out:
+ return error;
+}
+
+static long do_semtimedop(int semid, struct sembuf __user *tsops,
+ unsigned nsops, const struct timespec64 *timeout)
+{
+ struct sembuf fast_sops[SEMOPM_FAST];
+ struct sembuf *sops = fast_sops;
+ struct ipc_namespace *ns;
+ int ret;
+
+ ns = current->nsproxy->ipc_ns;
+ if (nsops > ns->sc_semopm)
+ return -E2BIG;
+ if (nsops < 1)
+ return -EINVAL;
+
+ if (nsops > SEMOPM_FAST) {
+ sops = kvmalloc_array(nsops, sizeof(*sops), GFP_KERNEL);
+ if (sops == NULL)
+ return -ENOMEM;
+ }
+
+ if (copy_from_user(sops, tsops, nsops * sizeof(*tsops))) {
+ ret = -EFAULT;
+ goto out_free;
+ }
+
+ ret = __do_semtimedop(semid, sops, nsops, timeout, ns);
+
out_free:
if (sops != fast_sops)
kvfree(sops);
- return error;
+
+ return ret;
}
long ksys_semtimedop(int semid, struct sembuf __user *tsops,
--
2.27.0
^ permalink raw reply related [flat|nested] 18+ messages in thread* [PATCH 7/9] ARM: oabi-compat: rework fcntl64() emulation
2020-10-30 15:49 ` [PATCH 1/9] mm/maccess: fix unaligned copy_{from,to}_kernel_nofault Arnd Bergmann
` (4 preceding siblings ...)
2020-10-30 15:49 ` [PATCH 6/9] ARM: oabi-compat: rework sys_semtimedop emulation Arnd Bergmann
@ 2020-10-30 15:49 ` Arnd Bergmann
2020-10-30 15:49 ` [PATCH 8/9] ARM: uaccess: add __{get,put}_kernel_nofault Arnd Bergmann
` (2 subsequent siblings)
8 siblings, 0 replies; 18+ messages in thread
From: Arnd Bergmann @ 2020-10-30 15:49 UTC (permalink / raw)
To: Russell King, Christoph Hellwig
Cc: linux-kernel, linux-arm-kernel, linux-arch, linux-mm, viro,
linus.walleij, arnd
From: Arnd Bergmann <arnd@arndb.de>
This is one of the last users of get_fs(), and this is fairly easy to
change, since the infrastructure for it is already there.
The replacement here is essentially a copy of the existing fcntl64()
syscall entry function.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
---
arch/arm/kernel/sys_oabi-compat.c | 93 ++++++++++++++++++++-----------
1 file changed, 60 insertions(+), 33 deletions(-)
diff --git a/arch/arm/kernel/sys_oabi-compat.c b/arch/arm/kernel/sys_oabi-compat.c
index c3e63b73b6ae..3449e163ea88 100644
--- a/arch/arm/kernel/sys_oabi-compat.c
+++ b/arch/arm/kernel/sys_oabi-compat.c
@@ -194,56 +194,83 @@ struct oabi_flock64 {
pid_t l_pid;
} __attribute__ ((packed,aligned(4)));
-static long do_locks(unsigned int fd, unsigned int cmd,
- unsigned long arg)
+static int get_oabi_flock(struct flock64 *kernel, struct oabi_flock64 __user *arg)
{
- struct flock64 kernel;
struct oabi_flock64 user;
- mm_segment_t fs;
- long ret;
if (copy_from_user(&user, (struct oabi_flock64 __user *)arg,
sizeof(user)))
return -EFAULT;
- kernel.l_type = user.l_type;
- kernel.l_whence = user.l_whence;
- kernel.l_start = user.l_start;
- kernel.l_len = user.l_len;
- kernel.l_pid = user.l_pid;
-
- fs = get_fs();
- set_fs(KERNEL_DS);
- ret = sys_fcntl64(fd, cmd, (unsigned long)&kernel);
- set_fs(fs);
-
- if (!ret && (cmd == F_GETLK64 || cmd == F_OFD_GETLK)) {
- user.l_type = kernel.l_type;
- user.l_whence = kernel.l_whence;
- user.l_start = kernel.l_start;
- user.l_len = kernel.l_len;
- user.l_pid = kernel.l_pid;
- if (copy_to_user((struct oabi_flock64 __user *)arg,
- &user, sizeof(user)))
- ret = -EFAULT;
- }
- return ret;
+
+ kernel->l_type = user.l_type;
+ kernel->l_whence = user.l_whence;
+ kernel->l_start = user.l_start;
+ kernel->l_len = user.l_len;
+ kernel->l_pid = user.l_pid;
+
+ return 0;
+}
+
+static int put_oabi_flock(struct flock64 *kernel, struct oabi_flock64 __user *arg)
+{
+ struct oabi_flock64 user;
+
+ user.l_type = kernel->l_type;
+ user.l_whence = kernel->l_whence;
+ user.l_start = kernel->l_start;
+ user.l_len = kernel->l_len;
+ user.l_pid = kernel->l_pid;
+
+ if (copy_to_user((struct oabi_flock64 __user *)arg,
+ &user, sizeof(user)))
+ return -EFAULT;
+
+ return 0;
}
asmlinkage long sys_oabi_fcntl64(unsigned int fd, unsigned int cmd,
unsigned long arg)
{
+ void __user *argp = (void __user *)arg;
+ struct fd f = fdget_raw(fd);
+ struct flock64 flock;
+ long err = -EBADF;
+
+ if (!f.file)
+ goto out;
+
switch (cmd) {
- case F_OFD_GETLK:
- case F_OFD_SETLK:
- case F_OFD_SETLKW:
case F_GETLK64:
+ case F_OFD_GETLK:
+ err = security_file_fcntl(f.file, cmd, arg);
+ if (err)
+ break;
+ err = get_oabi_flock(&flock, argp);
+ if (err)
+ break;
+ err = fcntl_getlk64(f.file, cmd, &flock);
+ if (!err)
+ err = put_oabi_flock(&flock, argp);
+ break;
case F_SETLK64:
case F_SETLKW64:
- return do_locks(fd, cmd, arg);
-
+ case F_OFD_SETLK:
+ case F_OFD_SETLKW:
+ err = security_file_fcntl(f.file, cmd, arg);
+ if (err)
+ break;
+ err = get_oabi_flock(&flock, argp);
+ if (err)
+ break;
+ err = fcntl_setlk64(fd, f.file, cmd, &flock);
+ break;
default:
- return sys_fcntl64(fd, cmd, arg);
+ err = sys_fcntl64(fd, cmd, arg);
+ break;
}
+ fdput(f);
+out:
+ return err;
}
struct oabi_epoll_event {
--
2.27.0
^ permalink raw reply related [flat|nested] 18+ messages in thread* [PATCH 8/9] ARM: uaccess: add __{get,put}_kernel_nofault
2020-10-30 15:49 ` [PATCH 1/9] mm/maccess: fix unaligned copy_{from,to}_kernel_nofault Arnd Bergmann
` (5 preceding siblings ...)
2020-10-30 15:49 ` [PATCH 7/9] ARM: oabi-compat: rework fcntl64() emulation Arnd Bergmann
@ 2020-10-30 15:49 ` Arnd Bergmann
2020-10-30 15:49 ` [PATCH 9/9] ARM: uaccess: remove set_fs() implementation Arnd Bergmann
2020-11-06 8:57 ` [PATCH 1/9] mm/maccess: fix unaligned copy_{from,to}_kernel_nofault Linus Walleij
8 siblings, 0 replies; 18+ messages in thread
From: Arnd Bergmann @ 2020-10-30 15:49 UTC (permalink / raw)
To: Russell King, Christoph Hellwig
Cc: linux-kernel, linux-arm-kernel, linux-arch, linux-mm, viro,
linus.walleij, arnd
From: Arnd Bergmann <arnd@arndb.de>
These mimic the behavior of get_user and put_user, except
for domain switching, address limit checking and handling
of mismatched sizes, none of which are relevant here.
To work with pre-Armv6 kernels, this has to avoid TUSER()
inside of the new macros, the new approach passes the "t"
string along with the opcode, which is a bit uglier but
avoids duplicating more code.
As there is no __get_user_asm_dword(), I work around it
by copying 32 bit at a time, which is possible because
the output size is known.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
---
arch/arm/include/asm/uaccess.h | 123 ++++++++++++++++++++++-----------
1 file changed, 83 insertions(+), 40 deletions(-)
diff --git a/arch/arm/include/asm/uaccess.h b/arch/arm/include/asm/uaccess.h
index a13d90206472..4f60638755c4 100644
--- a/arch/arm/include/asm/uaccess.h
+++ b/arch/arm/include/asm/uaccess.h
@@ -308,11 +308,11 @@ static inline void set_fs(mm_segment_t fs)
#define __get_user(x, ptr) \
({ \
long __gu_err = 0; \
- __get_user_err((x), (ptr), __gu_err); \
+ __get_user_err((x), (ptr), __gu_err, TUSER()); \
__gu_err; \
})
-#define __get_user_err(x, ptr, err) \
+#define __get_user_err(x, ptr, err, __t) \
do { \
unsigned long __gu_addr = (unsigned long)(ptr); \
unsigned long __gu_val; \
@@ -321,18 +321,19 @@ do { \
might_fault(); \
__ua_flags = uaccess_save_and_enable(); \
switch (sizeof(*(ptr))) { \
- case 1: __get_user_asm_byte(__gu_val, __gu_addr, err); break; \
- case 2: __get_user_asm_half(__gu_val, __gu_addr, err); break; \
- case 4: __get_user_asm_word(__gu_val, __gu_addr, err); break; \
+ case 1: __get_user_asm_byte(__gu_val, __gu_addr, err, __t); break; \
+ case 2: __get_user_asm_half(__gu_val, __gu_addr, err, __t); break; \
+ case 4: __get_user_asm_word(__gu_val, __gu_addr, err, __t); break; \
default: (__gu_val) = __get_user_bad(); \
} \
uaccess_restore(__ua_flags); \
(x) = (__typeof__(*(ptr)))__gu_val; \
} while (0)
+#endif
#define __get_user_asm(x, addr, err, instr) \
__asm__ __volatile__( \
- "1: " TUSER(instr) " %1, [%2], #0\n" \
+ "1: " instr " %1, [%2], #0\n" \
"2:\n" \
" .pushsection .text.fixup,\"ax\"\n" \
" .align 2\n" \
@@ -348,40 +349,38 @@ do { \
: "r" (addr), "i" (-EFAULT) \
: "cc")
-#define __get_user_asm_byte(x, addr, err) \
- __get_user_asm(x, addr, err, ldrb)
+#define __get_user_asm_byte(x, addr, err, __t) \
+ __get_user_asm(x, addr, err, "ldrb" __t)
#if __LINUX_ARM_ARCH__ >= 6
-#define __get_user_asm_half(x, addr, err) \
- __get_user_asm(x, addr, err, ldrh)
+#define __get_user_asm_half(x, addr, err, __t) \
+ __get_user_asm(x, addr, err, "ldrh" __t)
#else
#ifndef __ARMEB__
-#define __get_user_asm_half(x, __gu_addr, err) \
+#define __get_user_asm_half(x, __gu_addr, err, __t) \
({ \
unsigned long __b1, __b2; \
- __get_user_asm_byte(__b1, __gu_addr, err); \
- __get_user_asm_byte(__b2, __gu_addr + 1, err); \
+ __get_user_asm_byte(__b1, __gu_addr, err, __t); \
+ __get_user_asm_byte(__b2, __gu_addr + 1, err, __t); \
(x) = __b1 | (__b2 << 8); \
})
#else
-#define __get_user_asm_half(x, __gu_addr, err) \
+#define __get_user_asm_half(x, __gu_addr, err, __t) \
({ \
unsigned long __b1, __b2; \
- __get_user_asm_byte(__b1, __gu_addr, err); \
- __get_user_asm_byte(__b2, __gu_addr + 1, err); \
+ __get_user_asm_byte(__b1, __gu_addr, err, __t); \
+ __get_user_asm_byte(__b2, __gu_addr + 1, err, __t); \
(x) = (__b1 << 8) | __b2; \
})
#endif
#endif /* __LINUX_ARM_ARCH__ >= 6 */
-#define __get_user_asm_word(x, addr, err) \
- __get_user_asm(x, addr, err, ldr)
-#endif
-
+#define __get_user_asm_word(x, addr, err, __t) \
+ __get_user_asm(x, addr, err, "ldr" __t)
#define __put_user_switch(x, ptr, __err, __fn) \
do { \
@@ -425,7 +424,7 @@ do { \
#define __put_user_nocheck(x, __pu_ptr, __err, __size) \
do { \
unsigned long __pu_addr = (unsigned long)__pu_ptr; \
- __put_user_nocheck_##__size(x, __pu_addr, __err); \
+ __put_user_nocheck_##__size(x, __pu_addr, __err, TUSER());\
} while (0)
#define __put_user_nocheck_1 __put_user_asm_byte
@@ -433,9 +432,11 @@ do { \
#define __put_user_nocheck_4 __put_user_asm_word
#define __put_user_nocheck_8 __put_user_asm_dword
+#endif /* !CONFIG_CPU_SPECTRE */
+
#define __put_user_asm(x, __pu_addr, err, instr) \
__asm__ __volatile__( \
- "1: " TUSER(instr) " %1, [%2], #0\n" \
+ "1: " instr " %1, [%2], #0\n" \
"2:\n" \
" .pushsection .text.fixup,\"ax\"\n" \
" .align 2\n" \
@@ -450,36 +451,36 @@ do { \
: "r" (x), "r" (__pu_addr), "i" (-EFAULT) \
: "cc")
-#define __put_user_asm_byte(x, __pu_addr, err) \
- __put_user_asm(x, __pu_addr, err, strb)
+#define __put_user_asm_byte(x, __pu_addr, err, __t) \
+ __put_user_asm(x, __pu_addr, err, "strb" __t)
#if __LINUX_ARM_ARCH__ >= 6
-#define __put_user_asm_half(x, __pu_addr, err) \
- __put_user_asm(x, __pu_addr, err, strh)
+#define __put_user_asm_half(x, __pu_addr, err, __t) \
+ __put_user_asm(x, __pu_addr, err, "strh" __t)
#else
#ifndef __ARMEB__
-#define __put_user_asm_half(x, __pu_addr, err) \
+#define __put_user_asm_half(x, __pu_addr, err, __t) \
({ \
unsigned long __temp = (__force unsigned long)(x); \
- __put_user_asm_byte(__temp, __pu_addr, err); \
- __put_user_asm_byte(__temp >> 8, __pu_addr + 1, err); \
+ __put_user_asm_byte(__temp, __pu_addr, err, __t); \
+ __put_user_asm_byte(__temp >> 8, __pu_addr + 1, err, __t);\
})
#else
-#define __put_user_asm_half(x, __pu_addr, err) \
+#define __put_user_asm_half(x, __pu_addr, err, __t) \
({ \
unsigned long __temp = (__force unsigned long)(x); \
- __put_user_asm_byte(__temp >> 8, __pu_addr, err); \
- __put_user_asm_byte(__temp, __pu_addr + 1, err); \
+ __put_user_asm_byte(__temp >> 8, __pu_addr, err, __t); \
+ __put_user_asm_byte(__temp, __pu_addr + 1, err, __t); \
})
#endif
#endif /* __LINUX_ARM_ARCH__ >= 6 */
-#define __put_user_asm_word(x, __pu_addr, err) \
- __put_user_asm(x, __pu_addr, err, str)
+#define __put_user_asm_word(x, __pu_addr, err, __t) \
+ __put_user_asm(x, __pu_addr, err, "str" __t)
#ifndef __ARMEB__
#define __reg_oper0 "%R2"
@@ -489,12 +490,12 @@ do { \
#define __reg_oper1 "%R2"
#endif
-#define __put_user_asm_dword(x, __pu_addr, err) \
+#define __put_user_asm_dword(x, __pu_addr, err, __t) \
__asm__ __volatile__( \
- ARM( "1: " TUSER(str) " " __reg_oper1 ", [%1], #4\n" ) \
- ARM( "2: " TUSER(str) " " __reg_oper0 ", [%1]\n" ) \
- THUMB( "1: " TUSER(str) " " __reg_oper1 ", [%1]\n" ) \
- THUMB( "2: " TUSER(str) " " __reg_oper0 ", [%1, #4]\n" ) \
+ ARM( "1: str" __t " " __reg_oper1 ", [%1], #4\n" ) \
+ ARM( "2: str" __t " " __reg_oper0 ", [%1]\n" ) \
+ THUMB( "1: str" __t " " __reg_oper1 ", [%1]\n" ) \
+ THUMB( "2: str" __t " " __reg_oper0 ", [%1, #4]\n" ) \
"3:\n" \
" .pushsection .text.fixup,\"ax\"\n" \
" .align 2\n" \
@@ -510,7 +511,49 @@ do { \
: "r" (x), "i" (-EFAULT) \
: "cc")
-#endif /* !CONFIG_CPU_SPECTRE */
+#define HAVE_GET_KERNEL_NOFAULT
+
+#define __get_kernel_nofault(dst, src, type, err_label) \
+do { \
+ const type *__pk_ptr = (src); \
+ unsigned long __src = (unsigned long)(__pk_ptr); \
+ type __val; \
+ int __err = 0; \
+ switch (sizeof(type)) { \
+ case 1: __get_user_asm_byte(__val, __src, __err, ""); break; \
+ case 2: __get_user_asm_half(__val, __src, __err, ""); break; \
+ case 4: __get_user_asm_word(__val, __src, __err, ""); break; \
+ case 8: { \
+ u32 *__v32 = (u32*)&__val; \
+ __get_user_asm_word(__v32[0], __src, __err, ""); \
+ if (__err) \
+ break; \
+ __get_user_asm_word(__v32[1], __src+4, __err, ""); \
+ break; \
+ } \
+ default: __err = __get_user_bad(); break; \
+ } \
+ *(type *)(dst) = __val; \
+ if (__err) \
+ goto err_label; \
+} while (0)
+
+#define __put_kernel_nofault(dst, src, type, err_label) \
+do { \
+ const type *__pk_ptr = (dst); \
+ unsigned long __dst = (unsigned long)__pk_ptr; \
+ int __err = 0; \
+ type __val = *(type *)src; \
+ switch (sizeof(type)) { \
+ case 1: __put_user_asm_byte(__val, __dst, __err, ""); break; \
+ case 2: __put_user_asm_half(__val, __dst, __err, ""); break; \
+ case 4: __put_user_asm_word(__val, __dst, __err, ""); break; \
+ case 8: __put_user_asm_dword(__val, __dst, __err, ""); break; \
+ default: __err = __put_user_bad(); break; \
+ } \
+ if (__err) \
+ goto err_label; \
+} while (0)
#ifdef CONFIG_MMU
extern unsigned long __must_check
--
2.27.0
^ permalink raw reply related [flat|nested] 18+ messages in thread* [PATCH 9/9] ARM: uaccess: remove set_fs() implementation
2020-10-30 15:49 ` [PATCH 1/9] mm/maccess: fix unaligned copy_{from,to}_kernel_nofault Arnd Bergmann
` (6 preceding siblings ...)
2020-10-30 15:49 ` [PATCH 8/9] ARM: uaccess: add __{get,put}_kernel_nofault Arnd Bergmann
@ 2020-10-30 15:49 ` Arnd Bergmann
2020-11-06 8:57 ` [PATCH 1/9] mm/maccess: fix unaligned copy_{from,to}_kernel_nofault Linus Walleij
8 siblings, 0 replies; 18+ messages in thread
From: Arnd Bergmann @ 2020-10-30 15:49 UTC (permalink / raw)
To: Russell King, Christoph Hellwig
Cc: linux-kernel, linux-arm-kernel, linux-arch, linux-mm, viro,
linus.walleij, arnd
From: Arnd Bergmann <arnd@arndb.de>
There are no remaining callers of set_fs(), so just remove it
along with all associated code that operates on
thread_info->addr_limit.
There are still further optimizations that can be done:
- In get_user(), the address check could be moved entirely
into the out of line code, rather than passing a constant
as an argument,
- I assume the DACR handling can be simplified as we now
only change it during user access when CONFIG_CPU_SW_DOMAIN_PAN
is set, but not during set_fs().
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
---
arch/arm/Kconfig | 1 -
arch/arm/include/asm/ptrace.h | 1 -
arch/arm/include/asm/thread_info.h | 4 ---
arch/arm/include/asm/uaccess-asm.h | 6 ----
arch/arm/include/asm/uaccess.h | 46 +++---------------------------
arch/arm/kernel/asm-offsets.c | 2 --
arch/arm/kernel/entry-common.S | 9 ------
arch/arm/kernel/process.c | 7 +----
arch/arm/kernel/signal.c | 8 ------
arch/arm/lib/copy_from_user.S | 3 +-
arch/arm/lib/copy_to_user.S | 3 +-
11 files changed, 7 insertions(+), 83 deletions(-)
diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig
index fe2f17eb2b50..55a8892dd5d8 100644
--- a/arch/arm/Kconfig
+++ b/arch/arm/Kconfig
@@ -120,7 +120,6 @@ config ARM
select PCI_SYSCALL if PCI
select PERF_USE_VMALLOC
select RTC_LIB
- select SET_FS
select SYS_SUPPORTS_APM_EMULATION
# Above selects are sorted alphabetically; please add new ones
# according to that. Thanks.
diff --git a/arch/arm/include/asm/ptrace.h b/arch/arm/include/asm/ptrace.h
index 91d6b7856be4..93051e2f402c 100644
--- a/arch/arm/include/asm/ptrace.h
+++ b/arch/arm/include/asm/ptrace.h
@@ -19,7 +19,6 @@ struct pt_regs {
struct svc_pt_regs {
struct pt_regs regs;
u32 dacr;
- u32 addr_limit;
};
#define to_svc_pt_regs(r) container_of(r, struct svc_pt_regs, regs)
diff --git a/arch/arm/include/asm/thread_info.h b/arch/arm/include/asm/thread_info.h
index 536b6b979f63..8b705f611216 100644
--- a/arch/arm/include/asm/thread_info.h
+++ b/arch/arm/include/asm/thread_info.h
@@ -23,8 +23,6 @@ struct task_struct;
#include <asm/types.h>
-typedef unsigned long mm_segment_t;
-
struct cpu_context_save {
__u32 r4;
__u32 r5;
@@ -46,7 +44,6 @@ struct cpu_context_save {
struct thread_info {
unsigned long flags; /* low level flags */
int preempt_count; /* 0 => preemptable, <0 => bug */
- mm_segment_t addr_limit; /* address limit */
struct task_struct *task; /* main task structure */
__u32 cpu; /* cpu */
__u32 cpu_domain; /* cpu domain */
@@ -72,7 +69,6 @@ struct thread_info {
.task = &tsk, \
.flags = 0, \
.preempt_count = INIT_PREEMPT_COUNT, \
- .addr_limit = KERNEL_DS, \
}
/*
diff --git a/arch/arm/include/asm/uaccess-asm.h b/arch/arm/include/asm/uaccess-asm.h
index 907571fd05c6..6451a433912c 100644
--- a/arch/arm/include/asm/uaccess-asm.h
+++ b/arch/arm/include/asm/uaccess-asm.h
@@ -84,12 +84,8 @@
* if \disable is set.
*/
.macro uaccess_entry, tsk, tmp0, tmp1, tmp2, disable
- ldr \tmp1, [\tsk, #TI_ADDR_LIMIT]
- mov \tmp2, #TASK_SIZE
- str \tmp2, [\tsk, #TI_ADDR_LIMIT]
DACR( mrc p15, 0, \tmp0, c3, c0, 0)
DACR( str \tmp0, [sp, #SVC_DACR])
- str \tmp1, [sp, #SVC_ADDR_LIMIT]
.if \disable && IS_ENABLED(CONFIG_CPU_SW_DOMAIN_PAN)
/* kernel=client, user=no access */
mov \tmp2, #DACR_UACCESS_DISABLE
@@ -106,9 +102,7 @@
/* Restore the user access state previously saved by uaccess_entry */
.macro uaccess_exit, tsk, tmp0, tmp1
- ldr \tmp1, [sp, #SVC_ADDR_LIMIT]
DACR( ldr \tmp0, [sp, #SVC_DACR])
- str \tmp1, [\tsk, #TI_ADDR_LIMIT]
DACR( mcr p15, 0, \tmp0, c3, c0, 0)
.endm
diff --git a/arch/arm/include/asm/uaccess.h b/arch/arm/include/asm/uaccess.h
index 4f60638755c4..084d1c07c2d0 100644
--- a/arch/arm/include/asm/uaccess.h
+++ b/arch/arm/include/asm/uaccess.h
@@ -52,32 +52,8 @@ static __always_inline void uaccess_restore(unsigned int flags)
extern int __get_user_bad(void);
extern int __put_user_bad(void);
-/*
- * Note that this is actually 0x1,0000,0000
- */
-#define KERNEL_DS 0x00000000
-
#ifdef CONFIG_MMU
-#define USER_DS TASK_SIZE
-#define get_fs() (current_thread_info()->addr_limit)
-
-static inline void set_fs(mm_segment_t fs)
-{
- current_thread_info()->addr_limit = fs;
-
- /*
- * Prevent a mispredicted conditional call to set_fs from forwarding
- * the wrong address limit to access_ok under speculation.
- */
- dsb(nsh);
- isb();
-
- modify_domain(DOMAIN_KERNEL, fs ? DOMAIN_CLIENT : DOMAIN_MANAGER);
-}
-
-#define uaccess_kernel() (get_fs() == KERNEL_DS)
-
/*
* We use 33-bit arithmetic here. Success returns zero, failure returns
* addr_limit. We take advantage that addr_limit will be zero for KERNEL_DS,
@@ -89,7 +65,7 @@ static inline void set_fs(mm_segment_t fs)
__asm__(".syntax unified\n" \
"adds %1, %2, %3; sbcscc %1, %1, %0; movcc %0, #0" \
: "=&r" (flag), "=&r" (roksum) \
- : "r" (addr), "Ir" (size), "0" (current_thread_info()->addr_limit) \
+ : "r" (addr), "Ir" (size), "0" (TASK_SIZE) \
: "cc"); \
flag; })
@@ -120,7 +96,7 @@ static inline void __user *__uaccess_mask_range_ptr(const void __user *ptr,
" subshs %1, %1, %2\n"
" movlo %0, #0\n"
: "+r" (safe_ptr), "=&r" (tmp)
- : "r" (size), "r" (current_thread_info()->addr_limit)
+ : "r" (size), "r" (TASK_SIZE)
: "cc");
csdb();
@@ -194,7 +170,7 @@ extern int __get_user_64t_4(void *);
#define __get_user_check(x, p) \
({ \
- unsigned long __limit = current_thread_info()->addr_limit - 1; \
+ unsigned long __limit = TASK_SIZE - 1; \
register typeof(*(p)) __user *__p asm("r0") = (p); \
register __inttype(x) __r2 asm("r2"); \
register unsigned long __l asm("r1") = __limit; \
@@ -245,7 +221,7 @@ extern int __put_user_8(void *, unsigned long long);
#define __put_user_check(__pu_val, __ptr, __err, __s) \
({ \
- unsigned long __limit = current_thread_info()->addr_limit - 1; \
+ unsigned long __limit = TASK_SIZE - 1; \
register typeof(__pu_val) __r2 asm("r2") = __pu_val; \
register const void __user *__p asm("r0") = __ptr; \
register unsigned long __l asm("r1") = __limit; \
@@ -262,19 +238,8 @@ extern int __put_user_8(void *, unsigned long long);
#else /* CONFIG_MMU */
-/*
- * uClinux has only one addr space, so has simplified address limits.
- */
-#define USER_DS KERNEL_DS
-
-#define uaccess_kernel() (true)
#define __addr_ok(addr) ((void)(addr), 1)
#define __range_ok(addr, size) ((void)(addr), 0)
-#define get_fs() (KERNEL_DS)
-
-static inline void set_fs(mm_segment_t fs)
-{
-}
#define get_user(x, p) __get_user(x, p)
#define __put_user_check __put_user_nocheck
@@ -283,9 +248,6 @@ static inline void set_fs(mm_segment_t fs)
#define access_ok(addr, size) (__range_ok(addr, size) == 0)
-#define user_addr_max() \
- (uaccess_kernel() ? ~0UL : get_fs())
-
#ifdef CONFIG_CPU_SPECTRE
/*
* When mitigating Spectre variant 1, it is not worth fixing the non-
diff --git a/arch/arm/kernel/asm-offsets.c b/arch/arm/kernel/asm-offsets.c
index 97af6735172b..78f0a25baf2d 100644
--- a/arch/arm/kernel/asm-offsets.c
+++ b/arch/arm/kernel/asm-offsets.c
@@ -41,7 +41,6 @@ int main(void)
BLANK();
DEFINE(TI_FLAGS, offsetof(struct thread_info, flags));
DEFINE(TI_PREEMPT, offsetof(struct thread_info, preempt_count));
- DEFINE(TI_ADDR_LIMIT, offsetof(struct thread_info, addr_limit));
DEFINE(TI_TASK, offsetof(struct thread_info, task));
DEFINE(TI_CPU, offsetof(struct thread_info, cpu));
DEFINE(TI_CPU_DOMAIN, offsetof(struct thread_info, cpu_domain));
@@ -90,7 +89,6 @@ int main(void)
DEFINE(S_OLD_R0, offsetof(struct pt_regs, ARM_ORIG_r0));
DEFINE(PT_REGS_SIZE, sizeof(struct pt_regs));
DEFINE(SVC_DACR, offsetof(struct svc_pt_regs, dacr));
- DEFINE(SVC_ADDR_LIMIT, offsetof(struct svc_pt_regs, addr_limit));
DEFINE(SVC_REGS_SIZE, sizeof(struct svc_pt_regs));
BLANK();
DEFINE(SIGFRAME_RC3_OFFSET, offsetof(struct sigframe, retcode[3]));
diff --git a/arch/arm/kernel/entry-common.S b/arch/arm/kernel/entry-common.S
index 9a76467bbb47..2c0bde14fba6 100644
--- a/arch/arm/kernel/entry-common.S
+++ b/arch/arm/kernel/entry-common.S
@@ -49,9 +49,6 @@ __ret_fast_syscall:
UNWIND(.fnstart )
UNWIND(.cantunwind )
disable_irq_notrace @ disable interrupts
- ldr r2, [tsk, #TI_ADDR_LIMIT]
- cmp r2, #TASK_SIZE
- blne addr_limit_check_failed
ldr r1, [tsk, #TI_FLAGS] @ re-check for syscall tracing
tst r1, #_TIF_SYSCALL_WORK | _TIF_WORK_MASK
bne fast_work_pending
@@ -86,9 +83,6 @@ __ret_fast_syscall:
bl do_rseq_syscall
#endif
disable_irq_notrace @ disable interrupts
- ldr r2, [tsk, #TI_ADDR_LIMIT]
- cmp r2, #TASK_SIZE
- blne addr_limit_check_failed
ldr r1, [tsk, #TI_FLAGS] @ re-check for syscall tracing
tst r1, #_TIF_SYSCALL_WORK | _TIF_WORK_MASK
beq no_work_pending
@@ -127,9 +121,6 @@ ret_slow_syscall:
#endif
disable_irq_notrace @ disable interrupts
ENTRY(ret_to_user_from_irq)
- ldr r2, [tsk, #TI_ADDR_LIMIT]
- cmp r2, #TASK_SIZE
- blne addr_limit_check_failed
ldr r1, [tsk, #TI_FLAGS]
tst r1, #_TIF_WORK_MASK
bne slow_work_pending
diff --git a/arch/arm/kernel/process.c b/arch/arm/kernel/process.c
index 8e6ace03e960..28a1a4a9dd77 100644
--- a/arch/arm/kernel/process.c
+++ b/arch/arm/kernel/process.c
@@ -97,7 +97,7 @@ void __show_regs(struct pt_regs *regs)
unsigned long flags;
char buf[64];
#ifndef CONFIG_CPU_V7M
- unsigned int domain, fs;
+ unsigned int domain;
#ifdef CONFIG_CPU_SW_DOMAIN_PAN
/*
* Get the domain register for the parent context. In user
@@ -106,14 +106,11 @@ void __show_regs(struct pt_regs *regs)
*/
if (user_mode(regs)) {
domain = DACR_UACCESS_ENABLE;
- fs = get_fs();
} else {
domain = to_svc_pt_regs(regs)->dacr;
- fs = to_svc_pt_regs(regs)->addr_limit;
}
#else
domain = get_domain();
- fs = get_fs();
#endif
#endif
@@ -149,8 +146,6 @@ void __show_regs(struct pt_regs *regs)
if ((domain & domain_mask(DOMAIN_USER)) ==
domain_val(DOMAIN_USER, DOMAIN_NOACCESS))
segment = "none";
- else if (fs == KERNEL_DS)
- segment = "kernel";
else
segment = "user";
diff --git a/arch/arm/kernel/signal.c b/arch/arm/kernel/signal.c
index 585edbfccf6d..e2e4db8fee42 100644
--- a/arch/arm/kernel/signal.c
+++ b/arch/arm/kernel/signal.c
@@ -709,14 +709,6 @@ struct page *get_signal_page(void)
return page;
}
-/* Defer to generic check */
-asmlinkage void addr_limit_check_failed(void)
-{
-#ifdef CONFIG_MMU
- addr_limit_user_check();
-#endif
-}
-
#ifdef CONFIG_DEBUG_RSEQ
asmlinkage void do_rseq_syscall(struct pt_regs *regs)
{
diff --git a/arch/arm/lib/copy_from_user.S b/arch/arm/lib/copy_from_user.S
index f8016e3db65d..f481ef789a93 100644
--- a/arch/arm/lib/copy_from_user.S
+++ b/arch/arm/lib/copy_from_user.S
@@ -109,8 +109,7 @@
ENTRY(arm_copy_from_user)
#ifdef CONFIG_CPU_SPECTRE
- get_thread_info r3
- ldr r3, [r3, #TI_ADDR_LIMIT]
+ mov r3, #TASK_SIZE
uaccess_mask_range_ptr r1, r2, r3, ip
#endif
diff --git a/arch/arm/lib/copy_to_user.S b/arch/arm/lib/copy_to_user.S
index ebfe4cb3d912..215da16c7d6e 100644
--- a/arch/arm/lib/copy_to_user.S
+++ b/arch/arm/lib/copy_to_user.S
@@ -109,8 +109,7 @@
ENTRY(__copy_to_user_std)
WEAK(arm_copy_to_user)
#ifdef CONFIG_CPU_SPECTRE
- get_thread_info r3
- ldr r3, [r3, #TI_ADDR_LIMIT]
+ mov r3, #TASK_SIZE
uaccess_mask_range_ptr r0, r2, r3, ip
#endif
--
2.27.0
^ permalink raw reply related [flat|nested] 18+ messages in thread* Re: [PATCH 1/9] mm/maccess: fix unaligned copy_{from,to}_kernel_nofault
2020-10-30 15:49 ` [PATCH 1/9] mm/maccess: fix unaligned copy_{from,to}_kernel_nofault Arnd Bergmann
` (7 preceding siblings ...)
2020-10-30 15:49 ` [PATCH 9/9] ARM: uaccess: remove set_fs() implementation Arnd Bergmann
@ 2020-11-06 8:57 ` Linus Walleij
8 siblings, 0 replies; 18+ messages in thread
From: Linus Walleij @ 2020-11-06 8:57 UTC (permalink / raw)
To: Arnd Bergmann
Cc: Russell King, Christoph Hellwig, linux-kernel, Linux ARM,
Linux-Arch, Linux Memory Management List, Al Viro, Arnd Bergmann
On Fri, Oct 30, 2020 at 4:49 PM Arnd Bergmann <arnd@kernel.org> wrote:
> From: Arnd Bergmann <arnd@arndb.de>
>
> On machines such as ARMv5 that trap unaligned accesses, these
> two functions can be slow when each access needs to be emulated,
> or they might not work at all.
>
> Change them so that each loop is only used when both the src
> and dst pointers are naturally aligned.
>
> Reviewed-by: Christoph Hellwig <hch@lst.de>
> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Yours,
Linus Walleij
^ permalink raw reply [flat|nested] 18+ messages in thread