* [PATCH 6/6] powerpc/64s/kuap: conditionally restore AMR in kuap_restore_amr asm
From: Nicholas Piggin @ 2020-04-29 6:56 UTC (permalink / raw)
To: linuxppc-dev; +Cc: Nicholas Piggin
In-Reply-To: <20200429065654.1677541-1-npiggin@gmail.com>
Similar to the C code change, make the AMR restore conditional on
whether the register has changed.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
arch/powerpc/include/asm/book3s/64/kup-radix.h | 10 +++++++---
arch/powerpc/kernel/entry_64.S | 8 ++++----
arch/powerpc/kernel/exceptions-64s.S | 4 ++--
3 files changed, 13 insertions(+), 9 deletions(-)
diff --git a/arch/powerpc/include/asm/book3s/64/kup-radix.h b/arch/powerpc/include/asm/book3s/64/kup-radix.h
index ec8970958a26..e82df54f5681 100644
--- a/arch/powerpc/include/asm/book3s/64/kup-radix.h
+++ b/arch/powerpc/include/asm/book3s/64/kup-radix.h
@@ -12,13 +12,17 @@
#ifdef __ASSEMBLY__
-.macro kuap_restore_amr gpr
+.macro kuap_restore_amr gpr1, gpr2
#ifdef CONFIG_PPC_KUAP
BEGIN_MMU_FTR_SECTION_NESTED(67)
- ld \gpr, STACK_REGS_KUAP(r1)
+ mfspr \gpr1, SPRN_AMR
+ ld \gpr2, STACK_REGS_KUAP(r1)
+ cmpd \gpr1, \gpr2
+ beq 998f
isync
- mtspr SPRN_AMR, \gpr
+ mtspr SPRN_AMR, \gpr2
/* No isync required, see kuap_restore_amr() */
+998:
END_MMU_FTR_SECTION_NESTED_IFSET(MMU_FTR_RADIX_KUAP, 67)
#endif
.endm
diff --git a/arch/powerpc/kernel/entry_64.S b/arch/powerpc/kernel/entry_64.S
index b3c9f15089b6..9d49338e0c85 100644
--- a/arch/powerpc/kernel/entry_64.S
+++ b/arch/powerpc/kernel/entry_64.S
@@ -479,11 +479,11 @@ END_FTR_SECTION_IFCLR(CPU_FTR_ARCH_207S)
fast_interrupt_return:
_ASM_NOKPROBE_SYMBOL(fast_interrupt_return)
kuap_check_amr r3, r4
- ld r4,_MSR(r1)
- andi. r0,r4,MSR_PR
+ ld r5,_MSR(r1)
+ andi. r0,r5,MSR_PR
bne .Lfast_user_interrupt_return
- kuap_restore_amr r3
- andi. r0,r4,MSR_RI
+ kuap_restore_amr r3, r4
+ andi. r0,r5,MSR_RI
li r3,0 /* 0 return value, no EMULATE_STACK_STORE */
bne+ .Lfast_kernel_interrupt_return
addi r3,r1,STACK_FRAME_OVERHEAD
diff --git a/arch/powerpc/kernel/exceptions-64s.S b/arch/powerpc/kernel/exceptions-64s.S
index b0ad930cbae5..ef4a90212664 100644
--- a/arch/powerpc/kernel/exceptions-64s.S
+++ b/arch/powerpc/kernel/exceptions-64s.S
@@ -971,7 +971,7 @@ EXC_COMMON_BEGIN(system_reset_common)
ld r10,SOFTE(r1)
stb r10,PACAIRQSOFTMASK(r13)
- kuap_restore_amr r10
+ kuap_restore_amr r9, r10
EXCEPTION_RESTORE_REGS
RFI_TO_USER_OR_KERNEL
@@ -2757,7 +2757,7 @@ EXC_COMMON_BEGIN(soft_nmi_common)
ld r10,SOFTE(r1)
stb r10,PACAIRQSOFTMASK(r13)
- kuap_restore_amr r10
+ kuap_restore_amr r9, r10
EXCEPTION_RESTORE_REGS hsrr=0
RFI_TO_KERNEL
--
2.23.0
^ permalink raw reply related
* [PATCH 5/6] powerpc/64s/kuap: restore AMR in fast_interrupt_return
From: Nicholas Piggin @ 2020-04-29 6:56 UTC (permalink / raw)
To: linuxppc-dev; +Cc: Nicholas Piggin
In-Reply-To: <20200429065654.1677541-1-npiggin@gmail.com>
Interrupts that use fast_interrupt_return actually do lock AMR, but they
have been ones which tend to come from userspace (or kernel bugs) in
radix mode. With kuap on hash, segment interrupts are taken in kernel
often, which quickly breaks due to the missing restore.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
arch/powerpc/kernel/entry_64.S | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/arch/powerpc/kernel/entry_64.S b/arch/powerpc/kernel/entry_64.S
index 9a1e5d636dea..b3c9f15089b6 100644
--- a/arch/powerpc/kernel/entry_64.S
+++ b/arch/powerpc/kernel/entry_64.S
@@ -472,15 +472,17 @@ END_FTR_SECTION_IFCLR(CPU_FTR_ARCH_207S)
#ifdef CONFIG_PPC_BOOK3S
/*
* If MSR EE/RI was never enabled, IRQs not reconciled, NVGPRs not
- * touched, AMR not set, no exit work created, then this can be used.
+ * touched, no exit work created, then this can be used.
*/
.balign IFETCH_ALIGN_BYTES
.globl fast_interrupt_return
fast_interrupt_return:
_ASM_NOKPROBE_SYMBOL(fast_interrupt_return)
+ kuap_check_amr r3, r4
ld r4,_MSR(r1)
andi. r0,r4,MSR_PR
bne .Lfast_user_interrupt_return
+ kuap_restore_amr r3
andi. r0,r4,MSR_RI
li r3,0 /* 0 return value, no EMULATE_STACK_STORE */
bne+ .Lfast_kernel_interrupt_return
--
2.23.0
^ permalink raw reply related
* [PATCH 4/6] powerpc/64s/kuap: restore AMR in system reset exception
From: Nicholas Piggin @ 2020-04-29 6:56 UTC (permalink / raw)
To: linuxppc-dev; +Cc: Nicholas Piggin
In-Reply-To: <20200429065654.1677541-1-npiggin@gmail.com>
system reset interrupt handler locks AMR and exits with
PTION_RESTORE_REGS without restoring AMR. Similarly to the soft-NMI
ler, it needs to restore.
ed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
arch/powerpc/kernel/exceptions-64s.S | 1 +
1 file changed, 1 insertion(+)
diff --git a/arch/powerpc/kernel/exceptions-64s.S b/arch/powerpc/kernel/exceptions-64s.S
index 728ccb0f560c..b0ad930cbae5 100644
--- a/arch/powerpc/kernel/exceptions-64s.S
+++ b/arch/powerpc/kernel/exceptions-64s.S
@@ -971,6 +971,7 @@ EXC_COMMON_BEGIN(system_reset_common)
ld r10,SOFTE(r1)
stb r10,PACAIRQSOFTMASK(r13)
+ kuap_restore_amr r10
EXCEPTION_RESTORE_REGS
RFI_TO_USER_OR_KERNEL
--
2.23.0
^ permalink raw reply related
* [PATCH 3/6] powerpc/64/kuap: interrupt exit conditionally restore AMR
From: Nicholas Piggin @ 2020-04-29 6:56 UTC (permalink / raw)
To: linuxppc-dev; +Cc: Nicholas Piggin
In-Reply-To: <20200429065654.1677541-1-npiggin@gmail.com>
The AMR update is made conditional on AMR actually changing, which
should be the less common case on most workloads (though kernel page
faults on uaccess could be frequent, this doesn't significantly slow
down that case).
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
.../powerpc/include/asm/book3s/64/kup-radix.h | 36 ++++++++++++++-----
arch/powerpc/kernel/syscall_64.c | 14 +++++---
2 files changed, 37 insertions(+), 13 deletions(-)
diff --git a/arch/powerpc/include/asm/book3s/64/kup-radix.h b/arch/powerpc/include/asm/book3s/64/kup-radix.h
index 8dc5f292b806..ec8970958a26 100644
--- a/arch/powerpc/include/asm/book3s/64/kup-radix.h
+++ b/arch/powerpc/include/asm/book3s/64/kup-radix.h
@@ -62,19 +62,32 @@
#include <asm/mmu.h>
#include <asm/ptrace.h>
-static inline void kuap_restore_amr(struct pt_regs *regs)
+static inline void kuap_restore_amr(struct pt_regs *regs, unsigned long amr)
{
if (mmu_has_feature(MMU_FTR_RADIX_KUAP)) {
- isync();
- mtspr(SPRN_AMR, regs->kuap);
- /*
- * No isync required here because we are about to rfi
- * back to previous context before any user accesses
- * would be made, which is a CSI.
- */
+ if (unlikely(regs->kuap != amr)) {
+ isync();
+ mtspr(SPRN_AMR, regs->kuap);
+ /*
+ * No isync required here because we are about to rfi
+ * back to previous context before any user accesses
+ * would be made, which is a CSI.
+ */
+ }
}
}
+static inline unsigned long kuap_get_and_check_amr(void)
+{
+ if (mmu_has_feature(MMU_FTR_RADIX_KUAP)) {
+ unsigned long amr = mfspr(SPRN_AMR);
+ if (IS_ENABLED(CONFIG_PPC_KUAP_DEBUG)) /* kuap_check_amr() */
+ WARN_ON_ONCE(amr != AMR_KUAP_BLOCKED);
+ return amr;
+ }
+ return 0;
+}
+
static inline void kuap_check_amr(void)
{
if (IS_ENABLED(CONFIG_PPC_KUAP_DEBUG) && mmu_has_feature(MMU_FTR_RADIX_KUAP))
@@ -151,13 +164,18 @@ bad_kuap_fault(struct pt_regs *regs, unsigned long address, bool is_write)
"Bug: %s fault blocked by AMR!", is_write ? "Write" : "Read");
}
#else /* CONFIG_PPC_KUAP */
-static inline void kuap_restore_amr(struct pt_regs *regs)
+static inline void kuap_restore_amr(struct pt_regs *regs, unsigned long amr)
{
}
static inline void kuap_check_amr(void)
{
}
+
+static inline unsigned long kuap_get_and_check_amr(void)
+{
+ return 0;
+}
#endif /* CONFIG_PPC_KUAP */
#endif /* __ASSEMBLY__ */
diff --git a/arch/powerpc/kernel/syscall_64.c b/arch/powerpc/kernel/syscall_64.c
index a37c7717424f..edeab10c6888 100644
--- a/arch/powerpc/kernel/syscall_64.c
+++ b/arch/powerpc/kernel/syscall_64.c
@@ -242,6 +242,10 @@ notrace unsigned long interrupt_exit_user_prepare(struct pt_regs *regs, unsigned
BUG_ON(!FULL_REGS(regs));
BUG_ON(regs->softe != IRQS_ENABLED);
+ /*
+ * We don't need to restore AMR on the way back to userspace for KUAP.
+ * AMR can only have been unlocked if we interrupted the kernel.
+ */
kuap_check_amr();
local_irq_save(flags);
@@ -313,13 +317,14 @@ notrace unsigned long interrupt_exit_kernel_prepare(struct pt_regs *regs, unsign
unsigned long *ti_flagsp = ¤t_thread_info()->flags;
unsigned long flags;
unsigned long ret = 0;
+ unsigned long amr;
if (IS_ENABLED(CONFIG_PPC_BOOK3S) && unlikely(!(regs->msr & MSR_RI)))
unrecoverable_exception(regs);
BUG_ON(regs->msr & MSR_PR);
BUG_ON(!FULL_REGS(regs));
- kuap_check_amr();
+ amr = kuap_get_and_check_amr();
if (unlikely(*ti_flagsp & _TIF_EMULATE_STACK_STORE)) {
clear_bits(_TIF_EMULATE_STACK_STORE, ti_flagsp);
@@ -367,10 +372,11 @@ notrace unsigned long interrupt_exit_kernel_prepare(struct pt_regs *regs, unsign
#endif
/*
- * We don't need to restore AMR on the way back to userspace for KUAP.
- * The value of AMR only matters while we're in the kernel.
+ * Don't want to mfspr(SPRN_AMR) here, because this comes after
+ * mtmsr, which would cause RAW stalls. Hence, we take the AMR value
+ * from the check above.
*/
- kuap_restore_amr(regs);
+ kuap_restore_amr(regs, amr);
return ret;
}
--
2.23.0
^ permalink raw reply related
* [PATCH 2/6] powerpc/64s/kuap: kuap_restore missing isync
From: Nicholas Piggin @ 2020-04-29 6:56 UTC (permalink / raw)
To: linuxppc-dev; +Cc: Nicholas Piggin
In-Reply-To: <20200429065654.1677541-1-npiggin@gmail.com>
Writing the AMR register is documented to require context synchronizing
operations before and after, for it to take effect as expected. The kuap
restore at interrupt exit time deliberately avoids the isync after the
AMR update because it only needs to take effect after the context
synchronizing rfid that soon follows. Add a comment for this.
The missing isync before the update doesn't have an obvious
justification, and seems it could theorietically allow a rogue user
access to leak past the AMR update. Add isyncs for these.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
arch/powerpc/include/asm/book3s/64/kup-radix.h | 11 ++++++++++-
1 file changed, 10 insertions(+), 1 deletion(-)
diff --git a/arch/powerpc/include/asm/book3s/64/kup-radix.h b/arch/powerpc/include/asm/book3s/64/kup-radix.h
index 3bcef989a35d..8dc5f292b806 100644
--- a/arch/powerpc/include/asm/book3s/64/kup-radix.h
+++ b/arch/powerpc/include/asm/book3s/64/kup-radix.h
@@ -16,7 +16,9 @@
#ifdef CONFIG_PPC_KUAP
BEGIN_MMU_FTR_SECTION_NESTED(67)
ld \gpr, STACK_REGS_KUAP(r1)
+ isync
mtspr SPRN_AMR, \gpr
+ /* No isync required, see kuap_restore_amr() */
END_MMU_FTR_SECTION_NESTED_IFSET(MMU_FTR_RADIX_KUAP, 67)
#endif
.endm
@@ -62,8 +64,15 @@
static inline void kuap_restore_amr(struct pt_regs *regs)
{
- if (mmu_has_feature(MMU_FTR_RADIX_KUAP))
+ if (mmu_has_feature(MMU_FTR_RADIX_KUAP)) {
+ isync();
mtspr(SPRN_AMR, regs->kuap);
+ /*
+ * No isync required here because we are about to rfi
+ * back to previous context before any user accesses
+ * would be made, which is a CSI.
+ */
+ }
}
static inline void kuap_check_amr(void)
--
2.23.0
^ permalink raw reply related
* [PATCH 1/6] powerpc/64/kuap: move kuap checks out of MSR[RI]=0 regions of exit code
From: Nicholas Piggin @ 2020-04-29 6:56 UTC (permalink / raw)
To: linuxppc-dev; +Cc: Nicholas Piggin
In-Reply-To: <20200429065654.1677541-1-npiggin@gmail.com>
Any kind of WARN causes a program check that will crash with
unrecoverable exception if it occurs when RI is clear.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
arch/powerpc/kernel/syscall_64.c | 14 ++++++++------
1 file changed, 8 insertions(+), 6 deletions(-)
diff --git a/arch/powerpc/kernel/syscall_64.c b/arch/powerpc/kernel/syscall_64.c
index 8f7e268f3294..a37c7717424f 100644
--- a/arch/powerpc/kernel/syscall_64.c
+++ b/arch/powerpc/kernel/syscall_64.c
@@ -35,6 +35,8 @@ notrace long system_call_exception(long r3, long r4, long r5,
BUG_ON(!FULL_REGS(regs));
BUG_ON(regs->softe != IRQS_ENABLED);
+ kuap_check_amr();
+
account_cpu_user_entry();
#ifdef CONFIG_PPC_SPLPAR
@@ -47,8 +49,6 @@ notrace long system_call_exception(long r3, long r4, long r5,
}
#endif
- kuap_check_amr();
-
/*
* This is not required for the syscall exit path, but makes the
* stack frame look nicer. If this was initialised in the first stack
@@ -142,6 +142,8 @@ notrace unsigned long syscall_exit_prepare(unsigned long r3,
unsigned long ti_flags;
unsigned long ret = 0;
+ kuap_check_amr();
+
regs->result = r3;
/* Check whether the syscall is issued inside a restartable sequence */
@@ -218,8 +220,6 @@ notrace unsigned long syscall_exit_prepare(unsigned long r3,
local_paca->tm_scratch = regs->msr;
#endif
- kuap_check_amr();
-
account_cpu_user_exit();
return ret;
@@ -242,6 +242,8 @@ notrace unsigned long interrupt_exit_user_prepare(struct pt_regs *regs, unsigned
BUG_ON(!FULL_REGS(regs));
BUG_ON(regs->softe != IRQS_ENABLED);
+ kuap_check_amr();
+
local_irq_save(flags);
again:
@@ -298,8 +300,6 @@ notrace unsigned long interrupt_exit_user_prepare(struct pt_regs *regs, unsigned
local_paca->tm_scratch = regs->msr;
#endif
- kuap_check_amr();
-
account_cpu_user_exit();
return ret;
@@ -319,6 +319,8 @@ notrace unsigned long interrupt_exit_kernel_prepare(struct pt_regs *regs, unsign
BUG_ON(regs->msr & MSR_PR);
BUG_ON(!FULL_REGS(regs));
+ kuap_check_amr();
+
if (unlikely(*ti_flagsp & _TIF_EMULATE_STACK_STORE)) {
clear_bits(_TIF_EMULATE_STACK_STORE, ti_flagsp);
ret = 1;
--
2.23.0
^ permalink raw reply related
* [PATCH 0/6] assorted kuap fixes (try again)
From: Nicholas Piggin @ 2020-04-29 6:56 UTC (permalink / raw)
To: linuxppc-dev; +Cc: Nicholas Piggin
Well the last series was a disaster, I'll try again sending the
patches with proper subject and changelogs written.
Nicholas Piggin (6):
powerpc/64/kuap: move kuap checks out of MSR[RI]=0 regions of exit
code
powerpc/64s/kuap: kuap_restore missing isync
powerpc/64/kuap: interrupt exit conditionally restore AMR
powerpc/64s/kuap: restore AMR in system reset exception
powerpc/64s/kuap: restore AMR in fast_interrupt_return
powerpc/64s/kuap: conditionally restore AMR in kuap_restore_amr asm
.../powerpc/include/asm/book3s/64/kup-radix.h | 45 ++++++++++++++++---
arch/powerpc/kernel/entry_64.S | 10 +++--
arch/powerpc/kernel/exceptions-64s.S | 3 +-
arch/powerpc/kernel/syscall_64.c | 26 +++++++----
4 files changed, 63 insertions(+), 21 deletions(-)
--
2.23.0
^ permalink raw reply
* Re: [PATCH v2 2/3] powerpc/numa: Prefer node id queried from vphn
From: Gautham R Shenoy @ 2020-04-29 6:52 UTC (permalink / raw)
To: Srikar Dronamraju
Cc: Michal Hocko, Linus Torvalds, linux-kernel, linux-mm, Mel Gorman,
Kirill A. Shutemov, Andrew Morton, linuxppc-dev,
Christopher Lameter, Vlastimil Babka
In-Reply-To: <20200428093836.27190-3-srikar@linux.vnet.ibm.com>
Hello Srikar,
On Tue, Apr 28, 2020 at 03:08:35PM +0530, Srikar Dronamraju wrote:
> Node id queried from the static device tree may not
> be correct. For example: it may always show 0 on a shared processor.
> Hence prefer the node id queried from vphn and fallback on the device tree
> based node id if vphn query fails.
>
> Cc: linuxppc-dev@lists.ozlabs.org
> Cc: linux-mm@kvack.org
> Cc: linux-kernel@vger.kernel.org
> Cc: Michal Hocko <mhocko@suse.com>
> Cc: Mel Gorman <mgorman@suse.de>
> Cc: Vlastimil Babka <vbabka@suse.cz>
> Cc: "Kirill A. Shutemov" <kirill@shutemov.name>
> Cc: Christopher Lameter <cl@linux.com>
> Cc: Michael Ellerman <mpe@ellerman.id.au>
> Cc: Andrew Morton <akpm@linux-foundation.org>
> Cc: Linus Torvalds <torvalds@linux-foundation.org>
> Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
> ---
> Changelog v1:->v2:
> - Rebased to v5.7-rc3
>
> arch/powerpc/mm/numa.c | 16 ++++++++--------
> 1 file changed, 8 insertions(+), 8 deletions(-)
>
> diff --git a/arch/powerpc/mm/numa.c b/arch/powerpc/mm/numa.c
> index b3615b7fdbdf..281531340230 100644
> --- a/arch/powerpc/mm/numa.c
> +++ b/arch/powerpc/mm/numa.c
> @@ -719,20 +719,20 @@ static int __init parse_numa_properties(void)
> */
> for_each_present_cpu(i) {
> struct device_node *cpu;
> - int nid;
> -
> - cpu = of_get_cpu_node(i, NULL);
> - BUG_ON(!cpu);
> - nid = of_node_to_nid_single(cpu);
> - of_node_put(cpu);
> + int nid = vphn_get_nid(i);
>
> /*
> * Don't fall back to default_nid yet -- we will plug
> * cpus into nodes once the memory scan has discovered
> * the topology.
> */
> - if (nid < 0)
> - continue;
> + if (nid == NUMA_NO_NODE) {
> + cpu = of_get_cpu_node(i, NULL);
> + if (cpu) {
Why are we not retaining the BUG_ON(!cpu) assert here ?
> + nid = of_node_to_nid_single(cpu);
> + of_node_put(cpu);
> + }
> + }
Is it possible at this point that both vphn_get_nid(i) and
of_node_to_nid_single(cpu) returns NUMA_NO_NODE ? If so,
should we still call node_set_online() below ?
> node_set_online(nid);
> }
>
> --
> 2.20.1
>
--
Thanks and Regards
gautham.
^ permalink raw reply
* Re: [PATCH] fixup! signal: factor copy_siginfo_to_external32 from copy_siginfo_to_user32
From: Christoph Hellwig @ 2020-04-29 6:44 UTC (permalink / raw)
To: Arnd Bergmann
Cc: linux-kernel, Alexander Viro, Jeremy Kerr, linux-fsdevel,
Andrew Morton, linuxppc-dev, Christoph Hellwig,
Eric W . Biederman
In-Reply-To: <20200428195645.1365019-1-arnd@arndb.de>
On Tue, Apr 28, 2020 at 09:56:26PM +0200, Arnd Bergmann wrote:
> I think I found a way to improve the x32 handling:
>
> This is a simplification over Christoph's "[PATCH 2/7] signal: factor
> copy_siginfo_to_external32 from copy_siginfo_to_user32", reducing the
> x32 specifics in the common code to a single #ifdef/#endif check, in
> order to keep it more readable for everyone else.
>
> Christoph, if you like it, please fold into your patch.
What do you think of this version? This one always overrides
copy_siginfo_to_user32 for the x86 compat case to keep the churn down,
and improves the copy_siginfo_to_external32 documentation a bit.
---
From 5ca642c4c744ce6460ebb91fe30ec7a064d28e96 Mon Sep 17 00:00:00 2001
From: Christoph Hellwig <hch@lst.de>
Date: Wed, 29 Apr 2020 08:38:07 +0200
Subject: signal: factor copy_siginfo_to_external32 from copy_siginfo_to_user32
To remove the use of set_fs in the coredump code there needs to be a
way to convert a kernel siginfo to a userspace compat siginfo.
Factour out a copy_siginfo_to_external32 helper from
copy_siginfo_to_user32 that fills out the compat_siginfo, but does so
on a kernel space data structure. Handling of the x32 SIGCHLD magic
is kept out of copy_siginfo_to_external32, as coredumps are never
caused by SIGCHLD. To simplify the mess that the current
copy_siginfo_to_user32 is, we allow architectures to override it, and
keep thus keep the x32 SIGCHLD magic confined to x86-64.
Contains improvements from Eric W. Biederman <ebiederm@xmission.com>
and Arnd Bergmann <arnd@arndb.de>.
Signed-off-by: Christoph Hellwig <hch@lst.de>
---
arch/Kconfig | 3 ++
arch/x86/Kconfig | 1 +
arch/x86/kernel/signal.c | 22 ++++++++
include/linux/compat.h | 2 +
kernel/signal.c | 108 ++++++++++++++++++++-------------------
5 files changed, 83 insertions(+), 53 deletions(-)
diff --git a/arch/Kconfig b/arch/Kconfig
index 786a85d4ad40d..d5bc9f1ee1747 100644
--- a/arch/Kconfig
+++ b/arch/Kconfig
@@ -815,6 +815,9 @@ config OLD_SIGACTION
config COMPAT_OLD_SIGACTION
bool
+config ARCH_COPY_SIGINFO_TO_USER32
+ bool
+
config COMPAT_32BIT_TIME
bool "Provide system calls for 32-bit time_t"
default !64BIT || COMPAT
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index 1197b5596d5ad..ad13facbe9ffe 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -2906,6 +2906,7 @@ config X86_X32
config COMPAT_32
def_bool y
depends on IA32_EMULATION || X86_32
+ select ARCH_COPY_SIGINFO_TO_USER32
select HAVE_UID16
select OLD_SIGSUSPEND3
diff --git a/arch/x86/kernel/signal.c b/arch/x86/kernel/signal.c
index 83b74fb38c8fc..d2b09866105a2 100644
--- a/arch/x86/kernel/signal.c
+++ b/arch/x86/kernel/signal.c
@@ -511,6 +511,28 @@ static int __setup_rt_frame(int sig, struct ksignal *ksig,
}
#endif /* CONFIG_X86_32 */
+#ifdef CONFIG_ARCH_COPY_SIGINFO_TO_USER32
+int __copy_siginfo_to_user32(struct compat_siginfo __user *to,
+ const struct kernel_siginfo *from, bool x32_ABI)
+{
+ struct compat_siginfo new;
+
+ copy_siginfo_to_external32(&new, from);
+ if (x32_ABI && from->si_signo == SIGCHLD) {
+ new._sifields._sigchld_x32._utime = from->si_utime;
+ new._sifields._sigchld_x32._stime = from->si_stime;
+ }
+ if (copy_to_user(to, &new, sizeof(struct compat_siginfo)))
+ return -EFAULT;
+ return 0;
+}
+int copy_siginfo_to_user32(struct compat_siginfo __user *to,
+ const struct kernel_siginfo *from)
+{
+ return __copy_siginfo_to_user32(to, from, in_x32_syscall());
+}
+#endif /* CONFIG_ARCH_COPY_SIGINFO_TO_USER32 */
+
static int x32_setup_rt_frame(struct ksignal *ksig,
compat_sigset_t *set,
struct pt_regs *regs)
diff --git a/include/linux/compat.h b/include/linux/compat.h
index 0480ba4db5929..adbfe8f688d92 100644
--- a/include/linux/compat.h
+++ b/include/linux/compat.h
@@ -402,6 +402,8 @@ long compat_get_bitmap(unsigned long *mask, const compat_ulong_t __user *umask,
unsigned long bitmap_size);
long compat_put_bitmap(compat_ulong_t __user *umask, unsigned long *mask,
unsigned long bitmap_size);
+void copy_siginfo_to_external32(struct compat_siginfo *to,
+ const struct kernel_siginfo *from);
int copy_siginfo_from_user32(kernel_siginfo_t *to, const struct compat_siginfo __user *from);
int copy_siginfo_to_user32(struct compat_siginfo __user *to, const kernel_siginfo_t *from);
int get_compat_sigevent(struct sigevent *event,
diff --git a/kernel/signal.c b/kernel/signal.c
index 284fc1600063b..710655df1269d 100644
--- a/kernel/signal.c
+++ b/kernel/signal.c
@@ -3235,96 +3235,98 @@ int copy_siginfo_from_user(kernel_siginfo_t *to, const siginfo_t __user *from)
}
#ifdef CONFIG_COMPAT
-int copy_siginfo_to_user32(struct compat_siginfo __user *to,
- const struct kernel_siginfo *from)
-#if defined(CONFIG_X86_X32_ABI) || defined(CONFIG_IA32_EMULATION)
-{
- return __copy_siginfo_to_user32(to, from, in_x32_syscall());
-}
-int __copy_siginfo_to_user32(struct compat_siginfo __user *to,
- const struct kernel_siginfo *from, bool x32_ABI)
-#endif
+/**
+ * copy_siginfo_to_external32: copy a kernel signinfo into a 32-bit user one
+ * @to: compat siginfo destination
+ * @from: kernel siginfo source
+ *
+ * This function does not work properly for SIGCHLD on x32, but it does not need
+ * to as SIGCHLD never causes a coredump as this function is only intended to
+ * be used either by the coredump code or to implement copy_siginfo_to_user32,
+ * which can have its own arch version to deal with things like x32.
+ */
+void copy_siginfo_to_external32(struct compat_siginfo *to,
+ const struct kernel_siginfo *from)
{
- struct compat_siginfo new;
- memset(&new, 0, sizeof(new));
+ memset(to, 0, sizeof(*to));
- new.si_signo = from->si_signo;
- new.si_errno = from->si_errno;
- new.si_code = from->si_code;
+ to->si_signo = from->si_signo;
+ to->si_errno = from->si_errno;
+ to->si_code = from->si_code;
switch(siginfo_layout(from->si_signo, from->si_code)) {
case SIL_KILL:
- new.si_pid = from->si_pid;
- new.si_uid = from->si_uid;
+ to->si_pid = from->si_pid;
+ to->si_uid = from->si_uid;
break;
case SIL_TIMER:
- new.si_tid = from->si_tid;
- new.si_overrun = from->si_overrun;
- new.si_int = from->si_int;
+ to->si_tid = from->si_tid;
+ to->si_overrun = from->si_overrun;
+ to->si_int = from->si_int;
break;
case SIL_POLL:
- new.si_band = from->si_band;
- new.si_fd = from->si_fd;
+ to->si_band = from->si_band;
+ to->si_fd = from->si_fd;
break;
case SIL_FAULT:
- new.si_addr = ptr_to_compat(from->si_addr);
+ to->si_addr = ptr_to_compat(from->si_addr);
#ifdef __ARCH_SI_TRAPNO
- new.si_trapno = from->si_trapno;
+ to->si_trapno = from->si_trapno;
#endif
break;
case SIL_FAULT_MCEERR:
- new.si_addr = ptr_to_compat(from->si_addr);
+ to->si_addr = ptr_to_compat(from->si_addr);
#ifdef __ARCH_SI_TRAPNO
- new.si_trapno = from->si_trapno;
+ to->si_trapno = from->si_trapno;
#endif
- new.si_addr_lsb = from->si_addr_lsb;
+ to->si_addr_lsb = from->si_addr_lsb;
break;
case SIL_FAULT_BNDERR:
- new.si_addr = ptr_to_compat(from->si_addr);
+ to->si_addr = ptr_to_compat(from->si_addr);
#ifdef __ARCH_SI_TRAPNO
- new.si_trapno = from->si_trapno;
+ to->si_trapno = from->si_trapno;
#endif
- new.si_lower = ptr_to_compat(from->si_lower);
- new.si_upper = ptr_to_compat(from->si_upper);
+ to->si_lower = ptr_to_compat(from->si_lower);
+ to->si_upper = ptr_to_compat(from->si_upper);
break;
case SIL_FAULT_PKUERR:
- new.si_addr = ptr_to_compat(from->si_addr);
+ to->si_addr = ptr_to_compat(from->si_addr);
#ifdef __ARCH_SI_TRAPNO
- new.si_trapno = from->si_trapno;
+ to->si_trapno = from->si_trapno;
#endif
- new.si_pkey = from->si_pkey;
+ to->si_pkey = from->si_pkey;
break;
case SIL_CHLD:
- new.si_pid = from->si_pid;
- new.si_uid = from->si_uid;
- new.si_status = from->si_status;
-#ifdef CONFIG_X86_X32_ABI
- if (x32_ABI) {
- new._sifields._sigchld_x32._utime = from->si_utime;
- new._sifields._sigchld_x32._stime = from->si_stime;
- } else
-#endif
- {
- new.si_utime = from->si_utime;
- new.si_stime = from->si_stime;
- }
+ to->si_pid = from->si_pid;
+ to->si_uid = from->si_uid;
+ to->si_status = from->si_status;
+ to->si_utime = from->si_utime;
+ to->si_stime = from->si_stime;
break;
case SIL_RT:
- new.si_pid = from->si_pid;
- new.si_uid = from->si_uid;
- new.si_int = from->si_int;
+ to->si_pid = from->si_pid;
+ to->si_uid = from->si_uid;
+ to->si_int = from->si_int;
break;
case SIL_SYS:
- new.si_call_addr = ptr_to_compat(from->si_call_addr);
- new.si_syscall = from->si_syscall;
- new.si_arch = from->si_arch;
+ to->si_call_addr = ptr_to_compat(from->si_call_addr);
+ to->si_syscall = from->si_syscall;
+ to->si_arch = from->si_arch;
break;
}
+}
+#ifndef CONFIG_ARCH_COPY_SIGINFO_TO_USER32
+int copy_siginfo_to_user32(struct compat_siginfo __user *to,
+ const struct kernel_siginfo *from)
+{
+ struct compat_siginfo new;
+
+ copy_siginfo_to_external32(&new, from);
if (copy_to_user(to, &new, sizeof(struct compat_siginfo)))
return -EFAULT;
-
return 0;
}
+#endif /* ARCH_COPY_SIGINFO_TO_USER32 */
static int post_copy_siginfo_from_user32(kernel_siginfo_t *to,
const struct compat_siginfo *from)
--
2.26.2
^ permalink raw reply related
* Re: Section mismatch in reference from the function .early_init_mmu() to the function .init.text:.radix__early_init_mmu() after PowerPC updates 5.7-1
From: Christian Zigotzky @ 2020-04-29 6:41 UTC (permalink / raw)
To: Nicholas Piggin, linuxppc-dev
Cc: Darren Stevens, contact@a-eon.com, R.T.Dickinson,
Christian Zigotzky
In-Reply-To: <1588136973.mf0j03tcz0.astroid@bobo.none>
On 29 April 2020 at 07:13 am, Nicholas Piggin wrote:
> Excerpts from Christian Zigotzky's message of April 29, 2020 2:53 pm:
>> Hi All,
>>
>> The issue still exists in the RC3. (kernel config attached)
>>
>> Please help me to fix this issue.
> Huh, looks like maybe early_init_mmu() got uninlined because the
> compiler decided it was unlikely.
>
> Does this fix it?
>
> Thanks,
> Nick
> --
Hi Nick,
Many thanks for your patch! It solved the issue.
Have a nice day!
Cheers,
Christian
>
> diff --git a/arch/powerpc/include/asm/book3s/64/mmu.h b/arch/powerpc/include/asm/book3s/64/mmu.h
> index bb3deb76c951..3ffe5f967483 100644
> --- a/arch/powerpc/include/asm/book3s/64/mmu.h
> +++ b/arch/powerpc/include/asm/book3s/64/mmu.h
> @@ -208,7 +208,7 @@ void hash__early_init_devtree(void);
> void radix__early_init_devtree(void);
> extern void hash__early_init_mmu(void);
> extern void radix__early_init_mmu(void);
> -static inline void early_init_mmu(void)
> +static inline void __init early_init_mmu(void)
> {
> if (radix_enabled())
> return radix__early_init_mmu();
>
^ permalink raw reply
* Re: [RFC PATCH] powerpc/spufs: fix copy_to_user while atomic
From: Jeremy Kerr @ 2020-04-29 6:33 UTC (permalink / raw)
To: Christoph Hellwig
Cc: Arnd Bergmann, linuxppc-dev, linux-kernel, Eric W . Biederman,
linux-fsdevel, Andrew Morton, Linus Torvalds, Alexander Viro
In-Reply-To: <20200429061514.GD30946@lst.de>
Hi Christoph,
> And another one that should go on top of this one to address Al's other
> compaint:
Yeah, I was pondering that one. The access_ok() is kinda redundant, but
it does avoid forcing a SPU context save on those errors.
However, it's not like we really need to optimise for the case of
invalid addresses from userspace. So, I'll include this change in the
submission to Michael's tree. Arnd - let me know if you have any
objections.
Cheers,
Jeremy
^ permalink raw reply
* Re: [PATCH] fixup! signal: factor copy_siginfo_to_external32 from copy_siginfo_to_user32
From: Christoph Hellwig @ 2020-04-29 6:29 UTC (permalink / raw)
To: Christophe Leroy
Cc: Arnd Bergmann, linux-kernel, Alexander Viro, Eric W . Biederman,
linux-fsdevel, Andrew Morton, linuxppc-dev, Christoph Hellwig,
Jeremy Kerr
In-Reply-To: <d00b4436-9a97-5e90-2a6f-e79f90be9736@c-s.fr>
On Wed, Apr 29, 2020 at 08:17:22AM +0200, Christophe Leroy wrote:
>> +#ifndef CONFIG_X86_X32_ABI
>
> Can it be declared __weak instead of enclosing it in an #ifndef ?
I really hate the __weak ifdefs. But my plan was to move to a
CONFIG_ARCH_COPY_SIGINFO_TO_USER32 and have x86 select it.
^ permalink raw reply
* [PATCH 6/6] powerpc/64/kuap: conditionally restore AMR in kuap_restore_amr asm
From: Nicholas Piggin @ 2020-04-29 6:26 UTC (permalink / raw)
To: linuxppc-dev; +Cc: Nicholas Piggin
In-Reply-To: <20200429062607.1675792-1-npiggin@gmail.com>
Similar to the C code change, make the AMR restore conditional on
whether the register has changed.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
arch/powerpc/include/asm/book3s/64/kup-radix.h | 10 +++++++---
arch/powerpc/kernel/entry_64.S | 8 ++++----
arch/powerpc/kernel/exceptions-64s.S | 4 ++--
3 files changed, 13 insertions(+), 9 deletions(-)
diff --git a/arch/powerpc/include/asm/book3s/64/kup-radix.h b/arch/powerpc/include/asm/book3s/64/kup-radix.h
index ec8970958a26..e82df54f5681 100644
--- a/arch/powerpc/include/asm/book3s/64/kup-radix.h
+++ b/arch/powerpc/include/asm/book3s/64/kup-radix.h
@@ -12,13 +12,17 @@
#ifdef __ASSEMBLY__
-.macro kuap_restore_amr gpr
+.macro kuap_restore_amr gpr1, gpr2
#ifdef CONFIG_PPC_KUAP
BEGIN_MMU_FTR_SECTION_NESTED(67)
- ld \gpr, STACK_REGS_KUAP(r1)
+ mfspr \gpr1, SPRN_AMR
+ ld \gpr2, STACK_REGS_KUAP(r1)
+ cmpd \gpr1, \gpr2
+ beq 998f
isync
- mtspr SPRN_AMR, \gpr
+ mtspr SPRN_AMR, \gpr2
/* No isync required, see kuap_restore_amr() */
+998:
END_MMU_FTR_SECTION_NESTED_IFSET(MMU_FTR_RADIX_KUAP, 67)
#endif
.endm
diff --git a/arch/powerpc/kernel/entry_64.S b/arch/powerpc/kernel/entry_64.S
index b3c9f15089b6..9d49338e0c85 100644
--- a/arch/powerpc/kernel/entry_64.S
+++ b/arch/powerpc/kernel/entry_64.S
@@ -479,11 +479,11 @@ END_FTR_SECTION_IFCLR(CPU_FTR_ARCH_207S)
fast_interrupt_return:
_ASM_NOKPROBE_SYMBOL(fast_interrupt_return)
kuap_check_amr r3, r4
- ld r4,_MSR(r1)
- andi. r0,r4,MSR_PR
+ ld r5,_MSR(r1)
+ andi. r0,r5,MSR_PR
bne .Lfast_user_interrupt_return
- kuap_restore_amr r3
- andi. r0,r4,MSR_RI
+ kuap_restore_amr r3, r4
+ andi. r0,r5,MSR_RI
li r3,0 /* 0 return value, no EMULATE_STACK_STORE */
bne+ .Lfast_kernel_interrupt_return
addi r3,r1,STACK_FRAME_OVERHEAD
diff --git a/arch/powerpc/kernel/exceptions-64s.S b/arch/powerpc/kernel/exceptions-64s.S
index b0ad930cbae5..ef4a90212664 100644
--- a/arch/powerpc/kernel/exceptions-64s.S
+++ b/arch/powerpc/kernel/exceptions-64s.S
@@ -971,7 +971,7 @@ EXC_COMMON_BEGIN(system_reset_common)
ld r10,SOFTE(r1)
stb r10,PACAIRQSOFTMASK(r13)
- kuap_restore_amr r10
+ kuap_restore_amr r9, r10
EXCEPTION_RESTORE_REGS
RFI_TO_USER_OR_KERNEL
@@ -2757,7 +2757,7 @@ EXC_COMMON_BEGIN(soft_nmi_common)
ld r10,SOFTE(r1)
stb r10,PACAIRQSOFTMASK(r13)
- kuap_restore_amr r10
+ kuap_restore_amr r9, r10
EXCEPTION_RESTORE_REGS hsrr=0
RFI_TO_KERNEL
--
2.23.0
^ permalink raw reply related
* [PATCH 5/6] powerpc/64/kuap: restore AMR in fast_interrupt_return
From: Nicholas Piggin @ 2020-04-29 6:26 UTC (permalink / raw)
To: linuxppc-dev; +Cc: Nicholas Piggin
In-Reply-To: <20200429062607.1675792-1-npiggin@gmail.com>
Interrupts that use fast_interrupt_return actually do lock AMR, but they
have been ones which tend to come from userspace (or kernel bugs) in
radix mode. With kuap on hash, segment interrupts are taken in kernel
often, which quickly breaks due to the missing restore.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
arch/powerpc/kernel/entry_64.S | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/arch/powerpc/kernel/entry_64.S b/arch/powerpc/kernel/entry_64.S
index 9a1e5d636dea..b3c9f15089b6 100644
--- a/arch/powerpc/kernel/entry_64.S
+++ b/arch/powerpc/kernel/entry_64.S
@@ -472,15 +472,17 @@ END_FTR_SECTION_IFCLR(CPU_FTR_ARCH_207S)
#ifdef CONFIG_PPC_BOOK3S
/*
* If MSR EE/RI was never enabled, IRQs not reconciled, NVGPRs not
- * touched, AMR not set, no exit work created, then this can be used.
+ * touched, no exit work created, then this can be used.
*/
.balign IFETCH_ALIGN_BYTES
.globl fast_interrupt_return
fast_interrupt_return:
_ASM_NOKPROBE_SYMBOL(fast_interrupt_return)
+ kuap_check_amr r3, r4
ld r4,_MSR(r1)
andi. r0,r4,MSR_PR
bne .Lfast_user_interrupt_return
+ kuap_restore_amr r3
andi. r0,r4,MSR_RI
li r3,0 /* 0 return value, no EMULATE_STACK_STORE */
bne+ .Lfast_kernel_interrupt_return
--
2.23.0
^ permalink raw reply related
* [PATCH 4/6] rpc/64/kuap: restore AMR in system reset exception
From: Nicholas Piggin @ 2020-04-29 6:26 UTC (permalink / raw)
To: linuxppc-dev; +Cc: Nicholas Piggin
In-Reply-To: <20200429062607.1675792-1-npiggin@gmail.com>
system reset interrupt handler locks AMR and exits with
PTION_RESTORE_REGS without restoring AMR. Similarly to the soft-NMI
ler, it needs to restore.
ed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
arch/powerpc/kernel/exceptions-64s.S | 1 +
1 file changed, 1 insertion(+)
diff --git a/arch/powerpc/kernel/exceptions-64s.S b/arch/powerpc/kernel/exceptions-64s.S
index 728ccb0f560c..b0ad930cbae5 100644
--- a/arch/powerpc/kernel/exceptions-64s.S
+++ b/arch/powerpc/kernel/exceptions-64s.S
@@ -971,6 +971,7 @@ EXC_COMMON_BEGIN(system_reset_common)
ld r10,SOFTE(r1)
stb r10,PACAIRQSOFTMASK(r13)
+ kuap_restore_amr r10
EXCEPTION_RESTORE_REGS
RFI_TO_USER_OR_KERNEL
--
2.23.0
^ permalink raw reply related
* [PATCH 3/6] powerpc/64/kuap: interrupt exit kuap restore add missing isync, conditionally restore AMR
From: Nicholas Piggin @ 2020-04-29 6:26 UTC (permalink / raw)
To: linuxppc-dev; +Cc: Nicholas Piggin
In-Reply-To: <20200429062607.1675792-1-npiggin@gmail.com>
This fixes a missing isync before the mtspr(AMR), which ensures previous
memory accesses execute before the mtspr, so they can't slip past the
AMR check and access user memory if we are returning to a context where
kuap is allowed.
The AMR is made conditional, and only done if the AMR must change, which
should be the less common case on most workloads (though kernel page
faults on uaccess could be frequent, this doesn't significantly slow
down that case).
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
.../powerpc/include/asm/book3s/64/kup-radix.h | 36 ++++++++++++++-----
arch/powerpc/kernel/syscall_64.c | 14 +++++---
2 files changed, 37 insertions(+), 13 deletions(-)
diff --git a/arch/powerpc/include/asm/book3s/64/kup-radix.h b/arch/powerpc/include/asm/book3s/64/kup-radix.h
index 8dc5f292b806..ec8970958a26 100644
--- a/arch/powerpc/include/asm/book3s/64/kup-radix.h
+++ b/arch/powerpc/include/asm/book3s/64/kup-radix.h
@@ -62,19 +62,32 @@
#include <asm/mmu.h>
#include <asm/ptrace.h>
-static inline void kuap_restore_amr(struct pt_regs *regs)
+static inline void kuap_restore_amr(struct pt_regs *regs, unsigned long amr)
{
if (mmu_has_feature(MMU_FTR_RADIX_KUAP)) {
- isync();
- mtspr(SPRN_AMR, regs->kuap);
- /*
- * No isync required here because we are about to rfi
- * back to previous context before any user accesses
- * would be made, which is a CSI.
- */
+ if (unlikely(regs->kuap != amr)) {
+ isync();
+ mtspr(SPRN_AMR, regs->kuap);
+ /*
+ * No isync required here because we are about to rfi
+ * back to previous context before any user accesses
+ * would be made, which is a CSI.
+ */
+ }
}
}
+static inline unsigned long kuap_get_and_check_amr(void)
+{
+ if (mmu_has_feature(MMU_FTR_RADIX_KUAP)) {
+ unsigned long amr = mfspr(SPRN_AMR);
+ if (IS_ENABLED(CONFIG_PPC_KUAP_DEBUG)) /* kuap_check_amr() */
+ WARN_ON_ONCE(amr != AMR_KUAP_BLOCKED);
+ return amr;
+ }
+ return 0;
+}
+
static inline void kuap_check_amr(void)
{
if (IS_ENABLED(CONFIG_PPC_KUAP_DEBUG) && mmu_has_feature(MMU_FTR_RADIX_KUAP))
@@ -151,13 +164,18 @@ bad_kuap_fault(struct pt_regs *regs, unsigned long address, bool is_write)
"Bug: %s fault blocked by AMR!", is_write ? "Write" : "Read");
}
#else /* CONFIG_PPC_KUAP */
-static inline void kuap_restore_amr(struct pt_regs *regs)
+static inline void kuap_restore_amr(struct pt_regs *regs, unsigned long amr)
{
}
static inline void kuap_check_amr(void)
{
}
+
+static inline unsigned long kuap_get_and_check_amr(void)
+{
+ return 0;
+}
#endif /* CONFIG_PPC_KUAP */
#endif /* __ASSEMBLY__ */
diff --git a/arch/powerpc/kernel/syscall_64.c b/arch/powerpc/kernel/syscall_64.c
index a37c7717424f..edeab10c6888 100644
--- a/arch/powerpc/kernel/syscall_64.c
+++ b/arch/powerpc/kernel/syscall_64.c
@@ -242,6 +242,10 @@ notrace unsigned long interrupt_exit_user_prepare(struct pt_regs *regs, unsigned
BUG_ON(!FULL_REGS(regs));
BUG_ON(regs->softe != IRQS_ENABLED);
+ /*
+ * We don't need to restore AMR on the way back to userspace for KUAP.
+ * AMR can only have been unlocked if we interrupted the kernel.
+ */
kuap_check_amr();
local_irq_save(flags);
@@ -313,13 +317,14 @@ notrace unsigned long interrupt_exit_kernel_prepare(struct pt_regs *regs, unsign
unsigned long *ti_flagsp = ¤t_thread_info()->flags;
unsigned long flags;
unsigned long ret = 0;
+ unsigned long amr;
if (IS_ENABLED(CONFIG_PPC_BOOK3S) && unlikely(!(regs->msr & MSR_RI)))
unrecoverable_exception(regs);
BUG_ON(regs->msr & MSR_PR);
BUG_ON(!FULL_REGS(regs));
- kuap_check_amr();
+ amr = kuap_get_and_check_amr();
if (unlikely(*ti_flagsp & _TIF_EMULATE_STACK_STORE)) {
clear_bits(_TIF_EMULATE_STACK_STORE, ti_flagsp);
@@ -367,10 +372,11 @@ notrace unsigned long interrupt_exit_kernel_prepare(struct pt_regs *regs, unsign
#endif
/*
- * We don't need to restore AMR on the way back to userspace for KUAP.
- * The value of AMR only matters while we're in the kernel.
+ * Don't want to mfspr(SPRN_AMR) here, because this comes after
+ * mtmsr, which would cause RAW stalls. Hence, we take the AMR value
+ * from the check above.
*/
- kuap_restore_amr(regs);
+ kuap_restore_amr(regs, amr);
return ret;
}
--
2.23.0
^ permalink raw reply related
* [PATCH 2/6] missing isync
From: Nicholas Piggin @ 2020-04-29 6:26 UTC (permalink / raw)
To: linuxppc-dev; +Cc: Nicholas Piggin
In-Reply-To: <20200429062607.1675792-1-npiggin@gmail.com>
---
arch/powerpc/include/asm/book3s/64/kup-radix.h | 11 ++++++++++-
1 file changed, 10 insertions(+), 1 deletion(-)
diff --git a/arch/powerpc/include/asm/book3s/64/kup-radix.h b/arch/powerpc/include/asm/book3s/64/kup-radix.h
index 3bcef989a35d..8dc5f292b806 100644
--- a/arch/powerpc/include/asm/book3s/64/kup-radix.h
+++ b/arch/powerpc/include/asm/book3s/64/kup-radix.h
@@ -16,7 +16,9 @@
#ifdef CONFIG_PPC_KUAP
BEGIN_MMU_FTR_SECTION_NESTED(67)
ld \gpr, STACK_REGS_KUAP(r1)
+ isync
mtspr SPRN_AMR, \gpr
+ /* No isync required, see kuap_restore_amr() */
END_MMU_FTR_SECTION_NESTED_IFSET(MMU_FTR_RADIX_KUAP, 67)
#endif
.endm
@@ -62,8 +64,15 @@
static inline void kuap_restore_amr(struct pt_regs *regs)
{
- if (mmu_has_feature(MMU_FTR_RADIX_KUAP))
+ if (mmu_has_feature(MMU_FTR_RADIX_KUAP)) {
+ isync();
mtspr(SPRN_AMR, regs->kuap);
+ /*
+ * No isync required here because we are about to rfi
+ * back to previous context before any user accesses
+ * would be made, which is a CSI.
+ */
+ }
}
static inline void kuap_check_amr(void)
--
2.23.0
^ permalink raw reply related
* [PATCH 1/6] powerpc/64/kuap: move kuap checks out of MSR[RI]=0 regions of exit code
From: Nicholas Piggin @ 2020-04-29 6:26 UTC (permalink / raw)
To: linuxppc-dev; +Cc: Nicholas Piggin
In-Reply-To: <20200429062607.1675792-1-npiggin@gmail.com>
Any kind of WARN causes a program check that will crash with
unrecoverable exception if it occurs when RI is clear.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
arch/powerpc/kernel/syscall_64.c | 14 ++++++++------
1 file changed, 8 insertions(+), 6 deletions(-)
diff --git a/arch/powerpc/kernel/syscall_64.c b/arch/powerpc/kernel/syscall_64.c
index 8f7e268f3294..a37c7717424f 100644
--- a/arch/powerpc/kernel/syscall_64.c
+++ b/arch/powerpc/kernel/syscall_64.c
@@ -35,6 +35,8 @@ notrace long system_call_exception(long r3, long r4, long r5,
BUG_ON(!FULL_REGS(regs));
BUG_ON(regs->softe != IRQS_ENABLED);
+ kuap_check_amr();
+
account_cpu_user_entry();
#ifdef CONFIG_PPC_SPLPAR
@@ -47,8 +49,6 @@ notrace long system_call_exception(long r3, long r4, long r5,
}
#endif
- kuap_check_amr();
-
/*
* This is not required for the syscall exit path, but makes the
* stack frame look nicer. If this was initialised in the first stack
@@ -142,6 +142,8 @@ notrace unsigned long syscall_exit_prepare(unsigned long r3,
unsigned long ti_flags;
unsigned long ret = 0;
+ kuap_check_amr();
+
regs->result = r3;
/* Check whether the syscall is issued inside a restartable sequence */
@@ -218,8 +220,6 @@ notrace unsigned long syscall_exit_prepare(unsigned long r3,
local_paca->tm_scratch = regs->msr;
#endif
- kuap_check_amr();
-
account_cpu_user_exit();
return ret;
@@ -242,6 +242,8 @@ notrace unsigned long interrupt_exit_user_prepare(struct pt_regs *regs, unsigned
BUG_ON(!FULL_REGS(regs));
BUG_ON(regs->softe != IRQS_ENABLED);
+ kuap_check_amr();
+
local_irq_save(flags);
again:
@@ -298,8 +300,6 @@ notrace unsigned long interrupt_exit_user_prepare(struct pt_regs *regs, unsigned
local_paca->tm_scratch = regs->msr;
#endif
- kuap_check_amr();
-
account_cpu_user_exit();
return ret;
@@ -319,6 +319,8 @@ notrace unsigned long interrupt_exit_kernel_prepare(struct pt_regs *regs, unsign
BUG_ON(regs->msr & MSR_PR);
BUG_ON(!FULL_REGS(regs));
+ kuap_check_amr();
+
if (unlikely(*ti_flagsp & _TIF_EMULATE_STACK_STORE)) {
clear_bits(_TIF_EMULATE_STACK_STORE, ti_flagsp);
ret = 1;
--
2.23.0
^ permalink raw reply related
* [PATCH 0/6] assorted kuap fixes
From: Nicholas Piggin @ 2020-04-29 6:26 UTC (permalink / raw)
To: linuxppc-dev; +Cc: Nicholas Piggin
Here's a bunch of fixes I collected, and some that Aneesh needs for
his kuap on hash series.
Nicholas Piggin (6):
powerpc/64/kuap: move kuap checks out of MSR[RI]=0 regions of exit
code
missing isync
powerpc/64/kuap: interrupt exit kuap restore add missing isync,
conditionally restore AMR
rpc/64/kuap: restore AMR in system reset exception
powerpc/64/kuap: restore AMR in fast_interrupt_return
powerpc/64/kuap: conditionally restore AMR in kuap_restore_amr asm
.../powerpc/include/asm/book3s/64/kup-radix.h | 45 ++++++++++++++++---
arch/powerpc/kernel/entry_64.S | 10 +++--
arch/powerpc/kernel/exceptions-64s.S | 3 +-
arch/powerpc/kernel/syscall_64.c | 26 +++++++----
4 files changed, 63 insertions(+), 21 deletions(-)
--
2.23.0
^ permalink raw reply
* [PATCH] powerpc/64: refactor interrupt exit irq disabling sequence
From: Nicholas Piggin @ 2020-04-29 6:24 UTC (permalink / raw)
To: linuxppc-dev; +Cc: Nicholas Piggin
The same complicated sequence for juggling EE, RI, soft mask, and
irq tracing is repeated 3 times, tidy these up into one function.
This differs qiute a bit between sub architectures, so this makes
the ppc32 port cleaner as well.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
arch/powerpc/kernel/syscall_64.c | 58 +++++++++++++++-----------------
1 file changed, 28 insertions(+), 30 deletions(-)
diff --git a/arch/powerpc/kernel/syscall_64.c b/arch/powerpc/kernel/syscall_64.c
index c74295a7765b..8f7e268f3294 100644
--- a/arch/powerpc/kernel/syscall_64.c
+++ b/arch/powerpc/kernel/syscall_64.c
@@ -101,6 +101,31 @@ notrace long system_call_exception(long r3, long r4, long r5,
return f(r3, r4, r5, r6, r7, r8);
}
+/*
+ * local irqs must be disabled. Returns false if the caller must re-enable
+ * them, check for new work, and try again.
+ */
+static notrace inline bool prep_irq_for_enabled_exit(void)
+{
+ /* This must be done with RI=1 because tracing may touch vmaps */
+ trace_hardirqs_on();
+
+ /* This pattern matches prep_irq_for_idle */
+ __hard_EE_RI_disable();
+ if (unlikely(lazy_irq_pending())) {
+ /* Took an interrupt, may have more exit work to do. */
+ __hard_RI_enable();
+ trace_hardirqs_off();
+ local_paca->irq_happened |= PACA_IRQ_HARD_DIS;
+
+ return false;
+ }
+ local_paca->irq_happened = 0;
+ irq_soft_mask_set(IRQS_ENABLED);
+
+ return true;
+}
+
/*
* This should be called after a syscall returns, with r3 the return value
* from the syscall. If this function returns non-zero, the system call
@@ -184,21 +209,10 @@ notrace unsigned long syscall_exit_prepare(unsigned long r3,
}
}
- /* This must be done with RI=1 because tracing may touch vmaps */
- trace_hardirqs_on();
-
- /* This pattern matches prep_irq_for_idle */
- __hard_EE_RI_disable();
- if (unlikely(lazy_irq_pending())) {
- __hard_RI_enable();
- trace_hardirqs_off();
- local_paca->irq_happened |= PACA_IRQ_HARD_DIS;
+ if (unlikely(!prep_irq_for_enabled_exit())) {
local_irq_enable();
- /* Took an interrupt, may have more exit work to do. */
goto again;
}
- local_paca->irq_happened = 0;
- irq_soft_mask_set(IRQS_ENABLED);
#ifdef CONFIG_PPC_TRANSACTIONAL_MEM
local_paca->tm_scratch = regs->msr;
@@ -262,19 +276,11 @@ notrace unsigned long interrupt_exit_user_prepare(struct pt_regs *regs, unsigned
}
}
- trace_hardirqs_on();
- __hard_EE_RI_disable();
- if (unlikely(lazy_irq_pending())) {
- __hard_RI_enable();
- trace_hardirqs_off();
- local_paca->irq_happened |= PACA_IRQ_HARD_DIS;
+ if (unlikely(!prep_irq_for_enabled_exit())) {
local_irq_enable();
local_irq_disable();
- /* Took an interrupt, may have more exit work to do. */
goto again;
}
- local_paca->irq_happened = 0;
- irq_soft_mask_set(IRQS_ENABLED);
#ifdef CONFIG_PPC_BOOK3E
if (unlikely(ts->debug.dbcr0 & DBCR0_IDM)) {
@@ -332,13 +338,7 @@ notrace unsigned long interrupt_exit_kernel_prepare(struct pt_regs *regs, unsign
}
}
- trace_hardirqs_on();
- __hard_EE_RI_disable();
- if (unlikely(lazy_irq_pending())) {
- __hard_RI_enable();
- irq_soft_mask_set(IRQS_ALL_DISABLED);
- trace_hardirqs_off();
- local_paca->irq_happened |= PACA_IRQ_HARD_DIS;
+ if (unlikely(!prep_irq_for_enabled_exit())) {
/*
* Can't local_irq_restore to replay if we were in
* interrupt context. Must replay directly.
@@ -352,8 +352,6 @@ notrace unsigned long interrupt_exit_kernel_prepare(struct pt_regs *regs, unsign
/* Took an interrupt, may have more exit work to do. */
goto again;
}
- local_paca->irq_happened = 0;
- irq_soft_mask_set(IRQS_ENABLED);
} else {
/* Returning to a kernel context with local irqs disabled. */
__hard_EE_RI_disable();
--
2.23.0
^ permalink raw reply related
* Re: [PATCH] fixup! signal: factor copy_siginfo_to_external32 from copy_siginfo_to_user32
From: Christophe Leroy @ 2020-04-29 6:17 UTC (permalink / raw)
To: Arnd Bergmann, Christoph Hellwig
Cc: linux-kernel, Alexander Viro, Eric W . Biederman, linux-fsdevel,
Andrew Morton, linuxppc-dev, Jeremy Kerr
In-Reply-To: <20200428195645.1365019-1-arnd@arndb.de>
Le 28/04/2020 à 21:56, Arnd Bergmann a écrit :
> I think I found a way to improve the x32 handling:
>
> This is a simplification over Christoph's "[PATCH 2/7] signal: factor
> copy_siginfo_to_external32 from copy_siginfo_to_user32", reducing the
> x32 specifics in the common code to a single #ifdef/#endif check, in
> order to keep it more readable for everyone else.
>
> Christoph, if you like it, please fold into your patch.
>
> Cc: Andrew Morton <akpm@linux-foundation.org>
> Cc: Alexander Viro <viro@zeniv.linux.org.uk>
> Cc: Jeremy Kerr <jk@ozlabs.org>
> Cc: Arnd Bergmann <arnd@arndb.de>
> Cc: "Eric W . Biederman" <ebiederm@xmission.com>
> Cc: linuxppc-dev@lists.ozlabs.org
> Cc: linux-fsdevel@vger.kernel.org
> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
> ---
> arch/x86/include/asm/compat.h | 10 ++++++++++
> arch/x86/kernel/signal.c | 23 +++++++++++++++++++++++
> kernel/signal.c | 15 ++-------------
> 3 files changed, 35 insertions(+), 13 deletions(-)
>
[...]
> diff --git a/kernel/signal.c b/kernel/signal.c
> index 1a81602050b4..935facca4860 100644
> --- a/kernel/signal.c
> +++ b/kernel/signal.c
> @@ -3318,29 +3318,18 @@ void copy_siginfo_to_external32(struct compat_siginfo *to,
> }
> }
>
> +#ifndef CONFIG_X86_X32_ABI
Can it be declared __weak instead of enclosing it in an #ifndef ?
> int copy_siginfo_to_user32(struct compat_siginfo __user *to,
> const struct kernel_siginfo *from)
> -#if defined(CONFIG_X86_X32_ABI) || defined(CONFIG_IA32_EMULATION)
> -{
> - return __copy_siginfo_to_user32(to, from, in_x32_syscall());
> -}
> -int __copy_siginfo_to_user32(struct compat_siginfo __user *to,
> - const struct kernel_siginfo *from, bool x32_ABI)
> -#endif
> {
> struct compat_siginfo new;
>
> copy_siginfo_to_external32(&new, from);
> -#ifdef CONFIG_X86_X32_ABI
> - if (x32_ABI && from->si_signo == SIGCHLD) {
> - new._sifields._sigchld_x32._utime = from->si_utime;
> - new._sifields._sigchld_x32._stime = from->si_stime;
> - }
> -#endif
> if (copy_to_user(to, &new, sizeof(struct compat_siginfo)))
> return -EFAULT;
> return 0;
> }
> +#endif
>
> static int post_copy_siginfo_from_user32(kernel_siginfo_t *to,
> const struct compat_siginfo *from)
>
Christophe
^ permalink raw reply
* Re: [RFC PATCH] powerpc/spufs: fix copy_to_user while atomic
From: Christoph Hellwig @ 2020-04-29 6:15 UTC (permalink / raw)
To: Jeremy Kerr
Cc: Arnd Bergmann, linuxppc-dev, linux-kernel, Alexander Viro,
linux-fsdevel, Andrew Morton, Linus Torvalds, Christoph Hellwig,
Eric W . Biederman
In-Reply-To: <e1ebea36b162e8a3b4b24ecbc1051f8081ff5e53.camel@ozlabs.org>
And another one that should go on top of this one to address Al's other
compaint:
---
From 1b7ced3de0b3a4addec61f61ac5278c3ff141657 Mon Sep 17 00:00:00 2001
From: Christoph Hellwig <hch@lst.de>
Date: Wed, 22 Apr 2020 09:05:30 +0200
Subject: powerpc/spufs: stop using access_ok
Just use the proper non __-prefixed get/put_user variants where that is
not done yet.
Signed-off-by: Christoph Hellwig <hch@lst.de>
---
arch/powerpc/platforms/cell/spufs/file.c | 42 +++++-------------------
1 file changed, 8 insertions(+), 34 deletions(-)
diff --git a/arch/powerpc/platforms/cell/spufs/file.c b/arch/powerpc/platforms/cell/spufs/file.c
index b4e1ef650b406..cd7d10f27fad1 100644
--- a/arch/powerpc/platforms/cell/spufs/file.c
+++ b/arch/powerpc/platforms/cell/spufs/file.c
@@ -590,17 +590,12 @@ static ssize_t spufs_mbox_read(struct file *file, char __user *buf,
size_t len, loff_t *pos)
{
struct spu_context *ctx = file->private_data;
- u32 mbox_data, __user *udata;
+ u32 mbox_data, __user *udata = (void __user *)buf;
ssize_t count;
if (len < 4)
return -EINVAL;
- if (!access_ok(buf, len))
- return -EFAULT;
-
- udata = (void __user *)buf;
-
count = spu_acquire(ctx);
if (count)
return count;
@@ -616,7 +611,7 @@ static ssize_t spufs_mbox_read(struct file *file, char __user *buf,
* but still need to return the data we have
* read successfully so far.
*/
- ret = __put_user(mbox_data, udata);
+ ret = put_user(mbox_data, udata);
if (ret) {
if (!count)
count = -EFAULT;
@@ -698,17 +693,12 @@ static ssize_t spufs_ibox_read(struct file *file, char __user *buf,
size_t len, loff_t *pos)
{
struct spu_context *ctx = file->private_data;
- u32 ibox_data, __user *udata;
+ u32 ibox_data, __user *udata = (void __user *)buf;
ssize_t count;
if (len < 4)
return -EINVAL;
- if (!access_ok(buf, len))
- return -EFAULT;
-
- udata = (void __user *)buf;
-
count = spu_acquire(ctx);
if (count)
goto out;
@@ -727,7 +717,7 @@ static ssize_t spufs_ibox_read(struct file *file, char __user *buf,
}
/* if we can't write at all, return -EFAULT */
- count = __put_user(ibox_data, udata);
+ count = put_user(ibox_data, udata);
if (count)
goto out_unlock;
@@ -741,7 +731,7 @@ static ssize_t spufs_ibox_read(struct file *file, char __user *buf,
* but still need to return the data we have
* read successfully so far.
*/
- ret = __put_user(ibox_data, udata);
+ ret = put_user(ibox_data, udata);
if (ret)
break;
}
@@ -836,17 +826,13 @@ static ssize_t spufs_wbox_write(struct file *file, const char __user *buf,
size_t len, loff_t *pos)
{
struct spu_context *ctx = file->private_data;
- u32 wbox_data, __user *udata;
+ u32 wbox_data, __user *udata = (void __user *)buf;
ssize_t count;
if (len < 4)
return -EINVAL;
- udata = (void __user *)buf;
- if (!access_ok(buf, len))
- return -EFAULT;
-
- if (__get_user(wbox_data, udata))
+ if (get_user(wbox_data, udata))
return -EFAULT;
count = spu_acquire(ctx);
@@ -873,7 +859,7 @@ static ssize_t spufs_wbox_write(struct file *file, const char __user *buf,
/* write as much as possible */
for (count = 4, udata++; (count + 4) <= len; count += 4, udata++) {
int ret;
- ret = __get_user(wbox_data, udata);
+ ret = get_user(wbox_data, udata);
if (ret)
break;
@@ -1982,9 +1968,6 @@ static ssize_t spufs_mbox_info_read(struct file *file, char __user *buf,
u32 stat, data;
int ret;
- if (!access_ok(buf, len))
- return -EFAULT;
-
ret = spu_acquire_saved(ctx);
if (ret)
return ret;
@@ -2028,9 +2011,6 @@ static ssize_t spufs_ibox_info_read(struct file *file, char __user *buf,
u32 stat, data;
int ret;
- if (!access_ok(buf, len))
- return -EFAULT;
-
ret = spu_acquire_saved(ctx);
if (ret)
return ret;
@@ -2082,9 +2062,6 @@ static ssize_t spufs_wbox_info_read(struct file *file, char __user *buf,
u32 data[ARRAY_SIZE(ctx->csa.spu_mailbox_data)];
int ret, count;
- if (!access_ok(buf, len))
- return -EFAULT;
-
ret = spu_acquire_saved(ctx);
if (ret)
return ret;
@@ -2143,9 +2120,6 @@ static ssize_t spufs_dma_info_read(struct file *file, char __user *buf,
struct spu_dma_info info;
int ret;
- if (!access_ok(buf, len))
- return -EFAULT;
-
ret = spu_acquire_saved(ctx);
if (ret)
return ret;
--
2.26.2
^ permalink raw reply related
* Re: [RFC PATCH] powerpc/spufs: fix copy_to_user while atomic
From: Christoph Hellwig @ 2020-04-29 6:13 UTC (permalink / raw)
To: Jeremy Kerr
Cc: Arnd Bergmann, linuxppc-dev, linux-kernel, Alexander Viro,
linux-fsdevel, Andrew Morton, Linus Torvalds, Christoph Hellwig,
Eric W . Biederman
In-Reply-To: <20200429060553.GA30946@lst.de>
On Wed, Apr 29, 2020 at 08:05:53AM +0200, Christoph Hellwig wrote:
> On Wed, Apr 29, 2020 at 09:36:30AM +0800, Jeremy Kerr wrote:
> > Hi Christoph,
> >
> > > FYI, these little hunks reduce the difference to my version, maybe
> > > you can fold them in?
> >
> > Sure, no problem.
> >
> > How do you want to coordinate these? I can submit mine through mpe, but
> > that may make it tricky to synchronise with your changes. Or, you can
> > include this change in your series if you prefer.
>
> Maybe you can feed your patch directly to Linus through Michael
> ASAP, and I'll wait for that before resubmitting this series?
Btw, turns out my fold patch didn't actually compile without the
rebased patch on top, sorry. Here is the proper one:
diff --git a/arch/powerpc/platforms/cell/spufs/file.c b/arch/powerpc/platforms/cell/spufs/file.c
index c62d77ddaf7d3..b4e1ef650b406 100644
--- a/arch/powerpc/platforms/cell/spufs/file.c
+++ b/arch/powerpc/platforms/cell/spufs/file.c
@@ -2107,7 +2107,6 @@ static const struct file_operations spufs_wbox_info_fops = {
static void ___spufs_dma_info_read(struct spu_context *ctx,
struct spu_dma_info *info)
{
- struct mfc_cq_sr *qp, *spuqp;
int i;
info->dma_info_type = ctx->csa.priv2.spu_tag_status_query_RW;
@@ -2116,8 +2115,8 @@ static void ___spufs_dma_info_read(struct spu_context *ctx,
info->dma_info_stall_and_notify = ctx->csa.spu_chnldata_RW[25];
info->dma_info_atomic_command_status = ctx->csa.spu_chnldata_RW[27];
for (i = 0; i < 16; i++) {
- qp = &info->dma_info_command_data[i];
- spuqp = &ctx->csa.priv2.spuq[i];
+ struct mfc_cq_sr *qp = &info->dma_info_command_data[i];
+ struct mfc_cq_sr *spuqp = &ctx->csa.priv2.spuq[i];
qp->mfc_cq_data0_RW = spuqp->mfc_cq_data0_RW;
qp->mfc_cq_data1_RW = spuqp->mfc_cq_data1_RW;
@@ -2166,7 +2165,7 @@ static const struct file_operations spufs_dma_info_fops = {
};
static void ___spufs_proxydma_info_read(struct spu_context *ctx,
- struct spu_proxydma_info *info)
+ struct spu_proxydma_info *info)
{
int i;
^ permalink raw reply related
* Re: [RFC PATCH] powerpc/spufs: fix copy_to_user while atomic
From: Christoph Hellwig @ 2020-04-29 6:05 UTC (permalink / raw)
To: Jeremy Kerr
Cc: Arnd Bergmann, linuxppc-dev, linux-kernel, Alexander Viro,
linux-fsdevel, Andrew Morton, Linus Torvalds, Christoph Hellwig,
Eric W . Biederman
In-Reply-To: <e1ebea36b162e8a3b4b24ecbc1051f8081ff5e53.camel@ozlabs.org>
On Wed, Apr 29, 2020 at 09:36:30AM +0800, Jeremy Kerr wrote:
> Hi Christoph,
>
> > FYI, these little hunks reduce the difference to my version, maybe
> > you can fold them in?
>
> Sure, no problem.
>
> How do you want to coordinate these? I can submit mine through mpe, but
> that may make it tricky to synchronise with your changes. Or, you can
> include this change in your series if you prefer.
Maybe you can feed your patch directly to Linus through Michael
ASAP, and I'll wait for that before resubmitting this series?
^ permalink raw reply
* [PATCH 2/2] powerpc/perf: Add support for outputting extended regs in perf intr_regs
From: Anju T Sudhakar @ 2020-04-29 6:04 UTC (permalink / raw)
To: linuxppc-dev; +Cc: ravi.bangoria, maddy, linux-kernel, acme, anju, jolsa
In-Reply-To: <20200429060415.25930-1-anju@linux.vnet.ibm.com>
The capability flag PERF_PMU_CAP_EXTENDED_REGS, is used to indicate the
PMU which support extended registers. The generic code define the mask
of extended registers as 0 for non supported architectures.
Add support for extended registers in POWER9 architecture. For POWER9,
the extended registers are mmcr0, mmc1 and mmcr2.
REG_RESERVED mask is redefined to accommodate the extended registers.
With patch:
----------------
# perf record -I?
available registers: r0 r1 r2 r3 r4 r5 r6 r7 r8 r9 r10 r11 r12 r13 r14
r15 r16 r17 r18 r19 r20 r21 r22 r23 r24 r25 r26 r27 r28 r29 r30 r31 nip
msr orig_r3 ctr link xer ccr softe trap dar dsisr sier mmcra mmcr0
mmcr1 mmcr2
# perf record -I ls
# perf script -D
PERF_RECORD_SAMPLE(IP, 0x1): 9019/9019: 0 period: 1 addr: 0
... intr regs: mask 0xffffffffffff ABI 64-bit
.... r0 0xc00000000011b12c
.... r1 0xc000003f9a98b930
.... r2 0xc000000001a32100
.... r3 0xc000003f8fe9a800
.... r4 0xc000003fd1810000
.... r5 0x3e32557150
.... r6 0xc000003f9a98b908
.... r7 0xffffffc1cdae06ac
.... r8 0x818
[.....]
.... r31 0xc000003ffd047230
.... nip 0xc00000000011b2c0
.... msr 0x9000000000009033
.... orig_r3 0xc00000000011b21c
.... ctr 0xc000000000119380
.... link 0xc00000000011b12c
.... xer 0x0
.... ccr 0x28002222
.... softe 0x1
.... trap 0xf00
.... dar 0x0
.... dsisr 0x80000000000
.... sier 0x0
.... mmcra 0x80000000000
.... mmcr0 0x82008090
.... mmcr1 0x1e000000
.... mmcr2 0x0
... thread: perf:9019
Signed-off-by: Anju T Sudhakar <anju@linux.vnet.ibm.com>
---
arch/powerpc/include/asm/perf_event_server.h | 5 +++
arch/powerpc/include/uapi/asm/perf_regs.h | 13 +++++++-
arch/powerpc/perf/core-book3s.c | 1 +
arch/powerpc/perf/perf_regs.c | 29 ++++++++++++++--
arch/powerpc/perf/power9-pmu.c | 1 +
.../arch/powerpc/include/uapi/asm/perf_regs.h | 13 +++++++-
tools/perf/arch/powerpc/include/perf_regs.h | 6 +++-
tools/perf/arch/powerpc/util/perf_regs.c | 33 +++++++++++++++++++
8 files changed, 95 insertions(+), 6 deletions(-)
diff --git a/arch/powerpc/include/asm/perf_event_server.h b/arch/powerpc/include/asm/perf_event_server.h
index 3e9703f44c7c..1d15953bd99e 100644
--- a/arch/powerpc/include/asm/perf_event_server.h
+++ b/arch/powerpc/include/asm/perf_event_server.h
@@ -55,6 +55,11 @@ struct power_pmu {
int *blacklist_ev;
/* BHRB entries in the PMU */
int bhrb_nr;
+ /*
+ * set this flag with `PERF_PMU_CAP_EXTENDED_REGS` if
+ * the pmu supports extended perf regs capability
+ */
+ int capabilities;
};
/*
diff --git a/arch/powerpc/include/uapi/asm/perf_regs.h b/arch/powerpc/include/uapi/asm/perf_regs.h
index f599064dd8dc..604b831378fe 100644
--- a/arch/powerpc/include/uapi/asm/perf_regs.h
+++ b/arch/powerpc/include/uapi/asm/perf_regs.h
@@ -48,6 +48,17 @@ enum perf_event_powerpc_regs {
PERF_REG_POWERPC_DSISR,
PERF_REG_POWERPC_SIER,
PERF_REG_POWERPC_MMCRA,
- PERF_REG_POWERPC_MAX,
+ /* Extended registers */
+ PERF_REG_POWERPC_MMCR0,
+ PERF_REG_POWERPC_MMCR1,
+ PERF_REG_POWERPC_MMCR2,
+ PERF_REG_EXTENDED_MAX,
+ /* Max regs without the extended regs */
+ PERF_REG_POWERPC_MAX = PERF_REG_POWERPC_MMCRA + 1,
};
+
+#define PERF_REG_PMU_MASK ((1ULL << PERF_REG_POWERPC_MAX) - 1)
+#define PERF_REG_EXTENDED_MASK (((1ULL << (PERF_REG_EXTENDED_MAX)) \
+ - 1) - PERF_REG_PMU_MASK)
+
#endif /* _UAPI_ASM_POWERPC_PERF_REGS_H */
diff --git a/arch/powerpc/perf/core-book3s.c b/arch/powerpc/perf/core-book3s.c
index 3dcfecf858f3..f56b77800a7b 100644
--- a/arch/powerpc/perf/core-book3s.c
+++ b/arch/powerpc/perf/core-book3s.c
@@ -2276,6 +2276,7 @@ int register_power_pmu(struct power_pmu *pmu)
power_pmu.attr_groups = ppmu->attr_groups;
+ power_pmu.capabilities |= (ppmu->capabilities & PERF_PMU_CAP_EXTENDED_REGS);
#ifdef MSR_HV
/*
* Use FCHV to ignore kernel events if MSR.HV is set.
diff --git a/arch/powerpc/perf/perf_regs.c b/arch/powerpc/perf/perf_regs.c
index a213a0aa5d25..57aa02568caf 100644
--- a/arch/powerpc/perf/perf_regs.c
+++ b/arch/powerpc/perf/perf_regs.c
@@ -15,7 +15,8 @@
#define PT_REGS_OFFSET(id, r) [id] = offsetof(struct pt_regs, r)
-#define REG_RESERVED (~((1ULL << PERF_REG_POWERPC_MAX) - 1))
+#define REG_RESERVED (~(PERF_REG_EXTENDED_MASK) & \
+ (~((1ULL << PERF_REG_POWERPC_MAX) - 1)))
static unsigned int pt_regs_offset[PERF_REG_POWERPC_MAX] = {
PT_REGS_OFFSET(PERF_REG_POWERPC_R0, gpr[0]),
@@ -69,10 +70,22 @@ static unsigned int pt_regs_offset[PERF_REG_POWERPC_MAX] = {
PT_REGS_OFFSET(PERF_REG_POWERPC_MMCRA, dsisr),
};
+/* Function to return the extended register values */
+static u64 get_ext_regs_value(int idx)
+{
+ switch (idx) {
+ case PERF_REG_POWERPC_MMCR0:
+ return mfspr(SPRN_MMCR0);
+ case PERF_REG_POWERPC_MMCR1:
+ return mfspr(SPRN_MMCR1);
+ case PERF_REG_POWERPC_MMCR2:
+ return mfspr(SPRN_MMCR2);
+ default: return 0;
+ }
+}
+
u64 perf_reg_value(struct pt_regs *regs, int idx)
{
- if (WARN_ON_ONCE(idx >= PERF_REG_POWERPC_MAX))
- return 0;
if (idx == PERF_REG_POWERPC_SIER &&
(IS_ENABLED(CONFIG_FSL_EMB_PERF_EVENT) ||
@@ -85,6 +98,16 @@ u64 perf_reg_value(struct pt_regs *regs, int idx)
IS_ENABLED(CONFIG_PPC32)))
return 0;
+ if (idx >= PERF_REG_POWERPC_MAX && idx < PERF_REG_EXTENDED_MAX)
+ return get_ext_regs_value(idx);
+
+ /*
+ * If the idx is referring to value beyond the
+ * supported registers, return 0 with a warning
+ */
+ if (WARN_ON_ONCE(idx >= PERF_REG_EXTENDED_MAX))
+ return 0;
+
return regs_get_register(regs, pt_regs_offset[idx]);
}
diff --git a/arch/powerpc/perf/power9-pmu.c b/arch/powerpc/perf/power9-pmu.c
index 08c3ef796198..c37193b3e73f 100644
--- a/arch/powerpc/perf/power9-pmu.c
+++ b/arch/powerpc/perf/power9-pmu.c
@@ -434,6 +434,7 @@ static struct power_pmu power9_pmu = {
.cache_events = &power9_cache_events,
.attr_groups = power9_pmu_attr_groups,
.bhrb_nr = 32,
+ .capabilities = PERF_PMU_CAP_EXTENDED_REGS,
};
int init_power9_pmu(void)
diff --git a/tools/arch/powerpc/include/uapi/asm/perf_regs.h b/tools/arch/powerpc/include/uapi/asm/perf_regs.h
index f599064dd8dc..d66953294c73 100644
--- a/tools/arch/powerpc/include/uapi/asm/perf_regs.h
+++ b/tools/arch/powerpc/include/uapi/asm/perf_regs.h
@@ -48,6 +48,17 @@ enum perf_event_powerpc_regs {
PERF_REG_POWERPC_DSISR,
PERF_REG_POWERPC_SIER,
PERF_REG_POWERPC_MMCRA,
- PERF_REG_POWERPC_MAX,
+ /* Extended arch registers */
+ PERF_REG_POWERPC_MMCR0,
+ PERF_REG_POWERPC_MMCR1,
+ PERF_REG_POWERPC_MMCR2,
+ PERF_REG_EXTENDED_MAX,
+ /* Max regs without extended arch regs */
+ PERF_REG_POWERPC_MAX = PERF_REG_POWERPC_MMCRA + 1,
+
};
+#define PERF_REG_PMU_MASK ((1ULL << PERF_REG_POWERPC_MAX) - 1)
+#define PERF_REG_EXTENDED_MASK (((1ULL << (PERF_REG_EXTENDED_MAX))\
+ - 1) - PERF_REG_PMU_MASK)
+
#endif /* _UAPI_ASM_POWERPC_PERF_REGS_H */
diff --git a/tools/perf/arch/powerpc/include/perf_regs.h b/tools/perf/arch/powerpc/include/perf_regs.h
index e18a3556f5e3..f7bbdb816f88 100644
--- a/tools/perf/arch/powerpc/include/perf_regs.h
+++ b/tools/perf/arch/powerpc/include/perf_regs.h
@@ -64,7 +64,11 @@ static const char *reg_names[] = {
[PERF_REG_POWERPC_DAR] = "dar",
[PERF_REG_POWERPC_DSISR] = "dsisr",
[PERF_REG_POWERPC_SIER] = "sier",
- [PERF_REG_POWERPC_MMCRA] = "mmcra"
+ [PERF_REG_POWERPC_MMCRA] = "mmcra",
+ [PERF_REG_POWERPC_MMCR0] = "mmcr0",
+ [PERF_REG_POWERPC_MMCR1] = "mmcr1",
+ [PERF_REG_POWERPC_MMCR2] = "mmcr2",
+
};
static inline const char *perf_reg_name(int id)
diff --git a/tools/perf/arch/powerpc/util/perf_regs.c b/tools/perf/arch/powerpc/util/perf_regs.c
index 0a5242900248..37b150f9d1a1 100644
--- a/tools/perf/arch/powerpc/util/perf_regs.c
+++ b/tools/perf/arch/powerpc/util/perf_regs.c
@@ -6,6 +6,8 @@
#include "../../../util/perf_regs.h"
#include "../../../util/debug.h"
+#include "../../../util/event.h"
+#include "../../../perf-sys.h"
#include <linux/kernel.h>
@@ -55,6 +57,9 @@ const struct sample_reg sample_reg_masks[] = {
SMPL_REG(dsisr, PERF_REG_POWERPC_DSISR),
SMPL_REG(sier, PERF_REG_POWERPC_SIER),
SMPL_REG(mmcra, PERF_REG_POWERPC_MMCRA),
+ SMPL_REG(mmcr0, PERF_REG_POWERPC_MMCR0),
+ SMPL_REG(mmcr1, PERF_REG_POWERPC_MMCR1),
+ SMPL_REG(mmcr2, PERF_REG_POWERPC_MMCR2),
SMPL_REG_END
};
@@ -163,3 +168,31 @@ int arch_sdt_arg_parse_op(char *old_op, char **new_op)
return SDT_ARG_VALID;
}
+
+uint64_t arch__intr_reg_mask(void)
+{
+ struct perf_event_attr attr = {
+ .type = PERF_TYPE_HARDWARE,
+ .config = PERF_COUNT_HW_CPU_CYCLES,
+ .sample_type = PERF_SAMPLE_REGS_INTR,
+ .sample_regs_intr = PERF_REG_EXTENDED_MASK,
+ .precise_ip = 1,
+ .disabled = 1,
+ .exclude_kernel = 1,
+ };
+ int fd;
+
+ attr.sample_period = 1;
+ event_attr_init(&attr);
+
+ /*
+ * check if the pmu supports perf extended regs, before
+ * returning the register mask to sample.
+ */
+ fd = sys_perf_event_open(&attr, 0, -1, -1, 0);
+ if (fd != -1) {
+ close(fd);
+ return (PERF_REG_EXTENDED_MASK | PERF_REGS_MASK);
+ }
+ return PERF_REGS_MASK;
+}
--
2.20.1
^ permalink raw reply related
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox