* [PATCH v2 0/2] kdump: simplify code
@ 2021-12-11 3:33 ` Tiezhu Yang
0 siblings, 0 replies; 61+ messages in thread
From: Tiezhu Yang @ 2021-12-11 3:33 UTC (permalink / raw)
To: Dave Young, Baoquan He, Vivek Goyal, Andrew Morton
Cc: linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li
v2:
-- add copy_to_user_or_kernel() in lib/usercopy.c
-- define userbuf as bool type
Tiezhu Yang (2):
kdump: vmcore: remove copy_to() and add copy_to_user_or_kernel()
kdump: crashdump: use copy_to_user_or_kernel() to simplify code
arch/arm/kernel/crash_dump.c | 12 +++---------
arch/arm64/kernel/crash_dump.c | 12 +++---------
arch/ia64/kernel/crash_dump.c | 12 +++++-------
arch/mips/kernel/crash_dump.c | 11 +++--------
arch/powerpc/kernel/crash_dump.c | 11 ++++-------
arch/riscv/kernel/crash_dump.c | 11 +++--------
arch/sh/kernel/crash_dump.c | 11 +++--------
arch/x86/kernel/crash_dump_32.c | 11 +++--------
arch/x86/kernel/crash_dump_64.c | 15 +++++----------
fs/proc/vmcore.c | 32 +++++++++-----------------------
include/linux/crash_dump.h | 8 ++++----
include/linux/uaccess.h | 1 +
lib/usercopy.c | 15 +++++++++++++++
13 files changed, 61 insertions(+), 101 deletions(-)
--
2.1.0
^ permalink raw reply [flat|nested] 61+ messages in thread
* [PATCH v2 0/2] kdump: simplify code
@ 2021-12-11 3:33 ` Tiezhu Yang
0 siblings, 0 replies; 61+ messages in thread
From: Tiezhu Yang @ 2021-12-11 3:33 UTC (permalink / raw)
To: Dave Young, Baoquan He, Vivek Goyal, Andrew Morton
Cc: linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li
v2:
-- add copy_to_user_or_kernel() in lib/usercopy.c
-- define userbuf as bool type
Tiezhu Yang (2):
kdump: vmcore: remove copy_to() and add copy_to_user_or_kernel()
kdump: crashdump: use copy_to_user_or_kernel() to simplify code
arch/arm/kernel/crash_dump.c | 12 +++---------
arch/arm64/kernel/crash_dump.c | 12 +++---------
arch/ia64/kernel/crash_dump.c | 12 +++++-------
arch/mips/kernel/crash_dump.c | 11 +++--------
arch/powerpc/kernel/crash_dump.c | 11 ++++-------
arch/riscv/kernel/crash_dump.c | 11 +++--------
arch/sh/kernel/crash_dump.c | 11 +++--------
arch/x86/kernel/crash_dump_32.c | 11 +++--------
arch/x86/kernel/crash_dump_64.c | 15 +++++----------
fs/proc/vmcore.c | 32 +++++++++-----------------------
include/linux/crash_dump.h | 8 ++++----
include/linux/uaccess.h | 1 +
lib/usercopy.c | 15 +++++++++++++++
13 files changed, 61 insertions(+), 101 deletions(-)
--
2.1.0
_______________________________________________
linux-riscv mailing list
linux-riscv@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-riscv
^ permalink raw reply [flat|nested] 61+ messages in thread
* [PATCH v2 0/2] kdump: simplify code
@ 2021-12-11 3:33 ` Tiezhu Yang
0 siblings, 0 replies; 61+ messages in thread
From: Tiezhu Yang @ 2021-12-11 3:33 UTC (permalink / raw)
To: Dave Young, Baoquan He, Vivek Goyal, Andrew Morton
Cc: linux-ia64, linux-sh, Xuefeng Li, x86, kexec, linux-mips,
linux-kernel, linux-fsdevel, linux-riscv, linuxppc-dev,
linux-arm-kernel
v2:
-- add copy_to_user_or_kernel() in lib/usercopy.c
-- define userbuf as bool type
Tiezhu Yang (2):
kdump: vmcore: remove copy_to() and add copy_to_user_or_kernel()
kdump: crashdump: use copy_to_user_or_kernel() to simplify code
arch/arm/kernel/crash_dump.c | 12 +++---------
arch/arm64/kernel/crash_dump.c | 12 +++---------
arch/ia64/kernel/crash_dump.c | 12 +++++-------
arch/mips/kernel/crash_dump.c | 11 +++--------
arch/powerpc/kernel/crash_dump.c | 11 ++++-------
arch/riscv/kernel/crash_dump.c | 11 +++--------
arch/sh/kernel/crash_dump.c | 11 +++--------
arch/x86/kernel/crash_dump_32.c | 11 +++--------
arch/x86/kernel/crash_dump_64.c | 15 +++++----------
fs/proc/vmcore.c | 32 +++++++++-----------------------
include/linux/crash_dump.h | 8 ++++----
include/linux/uaccess.h | 1 +
lib/usercopy.c | 15 +++++++++++++++
13 files changed, 61 insertions(+), 101 deletions(-)
--
2.1.0
^ permalink raw reply [flat|nested] 61+ messages in thread
* [PATCH v2 0/2] kdump: simplify code
@ 2021-12-11 3:33 ` Tiezhu Yang
0 siblings, 0 replies; 61+ messages in thread
From: Tiezhu Yang @ 2021-12-11 3:33 UTC (permalink / raw)
To: Dave Young, Baoquan He, Vivek Goyal, Andrew Morton
Cc: linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li
v2:
-- add copy_to_user_or_kernel() in lib/usercopy.c
-- define userbuf as bool type
Tiezhu Yang (2):
kdump: vmcore: remove copy_to() and add copy_to_user_or_kernel()
kdump: crashdump: use copy_to_user_or_kernel() to simplify code
arch/arm/kernel/crash_dump.c | 12 +++---------
arch/arm64/kernel/crash_dump.c | 12 +++---------
arch/ia64/kernel/crash_dump.c | 12 +++++-------
arch/mips/kernel/crash_dump.c | 11 +++--------
arch/powerpc/kernel/crash_dump.c | 11 ++++-------
arch/riscv/kernel/crash_dump.c | 11 +++--------
arch/sh/kernel/crash_dump.c | 11 +++--------
arch/x86/kernel/crash_dump_32.c | 11 +++--------
arch/x86/kernel/crash_dump_64.c | 15 +++++----------
fs/proc/vmcore.c | 32 +++++++++-----------------------
include/linux/crash_dump.h | 8 ++++----
include/linux/uaccess.h | 1 +
lib/usercopy.c | 15 +++++++++++++++
13 files changed, 61 insertions(+), 101 deletions(-)
--
2.1.0
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 61+ messages in thread
* [PATCH v2 0/2] kdump: simplify code
@ 2021-12-11 3:33 ` Tiezhu Yang
0 siblings, 0 replies; 61+ messages in thread
From: Tiezhu Yang @ 2021-12-11 3:33 UTC (permalink / raw)
To: Dave Young, Baoquan He, Vivek Goyal, Andrew Morton
Cc: linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li
v2:
-- add copy_to_user_or_kernel() in lib/usercopy.c
-- define userbuf as bool type
Tiezhu Yang (2):
kdump: vmcore: remove copy_to() and add copy_to_user_or_kernel()
kdump: crashdump: use copy_to_user_or_kernel() to simplify code
arch/arm/kernel/crash_dump.c | 12 +++---------
arch/arm64/kernel/crash_dump.c | 12 +++---------
arch/ia64/kernel/crash_dump.c | 12 +++++-------
arch/mips/kernel/crash_dump.c | 11 +++--------
arch/powerpc/kernel/crash_dump.c | 11 ++++-------
arch/riscv/kernel/crash_dump.c | 11 +++--------
arch/sh/kernel/crash_dump.c | 11 +++--------
arch/x86/kernel/crash_dump_32.c | 11 +++--------
arch/x86/kernel/crash_dump_64.c | 15 +++++----------
fs/proc/vmcore.c | 32 +++++++++-----------------------
include/linux/crash_dump.h | 8 ++++----
include/linux/uaccess.h | 1 +
lib/usercopy.c | 15 +++++++++++++++
13 files changed, 61 insertions(+), 101 deletions(-)
--
2.1.0
_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec
^ permalink raw reply [flat|nested] 61+ messages in thread
* [PATCH v2 0/2] kdump: simplify code
@ 2021-12-11 3:33 ` Tiezhu Yang
0 siblings, 0 replies; 61+ messages in thread
From: Tiezhu Yang @ 2021-12-11 3:33 UTC (permalink / raw)
To: Dave Young, Baoquan He, Vivek Goyal, Andrew Morton
Cc: linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li
v2:
-- add copy_to_user_or_kernel() in lib/usercopy.c
-- define userbuf as bool type
Tiezhu Yang (2):
kdump: vmcore: remove copy_to() and add copy_to_user_or_kernel()
kdump: crashdump: use copy_to_user_or_kernel() to simplify code
arch/arm/kernel/crash_dump.c | 12 +++---------
arch/arm64/kernel/crash_dump.c | 12 +++---------
arch/ia64/kernel/crash_dump.c | 12 +++++-------
arch/mips/kernel/crash_dump.c | 11 +++--------
arch/powerpc/kernel/crash_dump.c | 11 ++++-------
arch/riscv/kernel/crash_dump.c | 11 +++--------
arch/sh/kernel/crash_dump.c | 11 +++--------
arch/x86/kernel/crash_dump_32.c | 11 +++--------
arch/x86/kernel/crash_dump_64.c | 15 +++++----------
fs/proc/vmcore.c | 32 +++++++++-----------------------
include/linux/crash_dump.h | 8 ++++----
include/linux/uaccess.h | 1 +
lib/usercopy.c | 15 +++++++++++++++
13 files changed, 61 insertions(+), 101 deletions(-)
--
2.1.0
^ permalink raw reply [flat|nested] 61+ messages in thread
* [PATCH v2 1/2] kdump: vmcore: remove copy_to() and add copy_to_user_or_kernel()
2021-12-11 3:33 ` Tiezhu Yang
` (3 preceding siblings ...)
(?)
@ 2021-12-11 3:33 ` Tiezhu Yang
-1 siblings, 0 replies; 61+ messages in thread
From: Tiezhu Yang @ 2021-12-11 3:33 UTC (permalink / raw)
To: Dave Young, Baoquan He, Vivek Goyal, Andrew Morton
Cc: linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li
In arch/*/kernel/crash_dump*.c, there exist many similar code
about copy_oldmem_page(), remove copy_to() in fs/proc/vmcore.c
and add copy_to_user_or_kernel() in lib/usercopy.c, then we can
use copy_to_user_or_kernel() to simplify the related code.
Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
---
fs/proc/vmcore.c | 28 +++++++---------------------
include/linux/uaccess.h | 1 +
lib/usercopy.c | 15 +++++++++++++++
3 files changed, 23 insertions(+), 21 deletions(-)
diff --git a/fs/proc/vmcore.c b/fs/proc/vmcore.c
index 509f851..f67fd77 100644
--- a/fs/proc/vmcore.c
+++ b/fs/proc/vmcore.c
@@ -238,22 +238,8 @@ copy_oldmem_page_encrypted(unsigned long pfn, char *buf, size_t csize,
return copy_oldmem_page(pfn, buf, csize, offset, userbuf);
}
-/*
- * Copy to either kernel or user space
- */
-static int copy_to(void *target, void *src, size_t size, int userbuf)
-{
- if (userbuf) {
- if (copy_to_user((char __user *) target, src, size))
- return -EFAULT;
- } else {
- memcpy(target, src, size);
- }
- return 0;
-}
-
#ifdef CONFIG_PROC_VMCORE_DEVICE_DUMP
-static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, int userbuf)
+static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, bool userbuf)
{
struct vmcoredd_node *dump;
u64 offset = 0;
@@ -266,7 +252,7 @@ static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, int userbuf)
if (start < offset + dump->size) {
tsz = min(offset + (u64)dump->size - start, (u64)size);
buf = dump->buf + start - offset;
- if (copy_to(dst, buf, tsz, userbuf)) {
+ if (copy_to_user_or_kernel(dst, buf, tsz, userbuf)) {
ret = -EFAULT;
goto out_unlock;
}
@@ -330,7 +316,7 @@ static int vmcoredd_mmap_dumps(struct vm_area_struct *vma, unsigned long dst,
* returned otherwise number of bytes read are returned.
*/
static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
- int userbuf)
+ bool userbuf)
{
ssize_t acc = 0, tmp;
size_t tsz;
@@ -347,7 +333,7 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
/* Read ELF core header */
if (*fpos < elfcorebuf_sz) {
tsz = min(elfcorebuf_sz - (size_t)*fpos, buflen);
- if (copy_to(buffer, elfcorebuf + *fpos, tsz, userbuf))
+ if (copy_to_user_or_kernel(buffer, elfcorebuf + *fpos, tsz, userbuf))
return -EFAULT;
buflen -= tsz;
*fpos += tsz;
@@ -395,7 +381,7 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
/* Read remaining elf notes */
tsz = min(elfcorebuf_sz + elfnotes_sz - (size_t)*fpos, buflen);
kaddr = elfnotes_buf + *fpos - elfcorebuf_sz - vmcoredd_orig_sz;
- if (copy_to(buffer, kaddr, tsz, userbuf))
+ if (copy_to_user_or_kernel(buffer, kaddr, tsz, userbuf))
return -EFAULT;
buflen -= tsz;
@@ -435,7 +421,7 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
static ssize_t read_vmcore(struct file *file, char __user *buffer,
size_t buflen, loff_t *fpos)
{
- return __read_vmcore((__force char *) buffer, buflen, fpos, 1);
+ return __read_vmcore((__force char *) buffer, buflen, fpos, true);
}
/*
@@ -461,7 +447,7 @@ static vm_fault_t mmap_vmcore_fault(struct vm_fault *vmf)
if (!PageUptodate(page)) {
offset = (loff_t) index << PAGE_SHIFT;
buf = __va((page_to_pfn(page) << PAGE_SHIFT));
- rc = __read_vmcore(buf, PAGE_SIZE, &offset, 0);
+ rc = __read_vmcore(buf, PAGE_SIZE, &offset, false);
if (rc < 0) {
unlock_page(page);
put_page(page);
diff --git a/include/linux/uaccess.h b/include/linux/uaccess.h
index ac03940..a25e682e 100644
--- a/include/linux/uaccess.h
+++ b/include/linux/uaccess.h
@@ -283,6 +283,7 @@ __copy_from_user_inatomic_nocache(void *to, const void __user *from,
#endif /* ARCH_HAS_NOCACHE_UACCESS */
extern __must_check int check_zeroed_user(const void __user *from, size_t size);
+extern __must_check int copy_to_user_or_kernel(void *target, void *src, size_t size, bool userbuf);
/**
* copy_struct_from_user: copy a struct from userspace
diff --git a/lib/usercopy.c b/lib/usercopy.c
index 7413dd3..7431b1b 100644
--- a/lib/usercopy.c
+++ b/lib/usercopy.c
@@ -90,3 +90,18 @@ int check_zeroed_user(const void __user *from, size_t size)
return -EFAULT;
}
EXPORT_SYMBOL(check_zeroed_user);
+
+/*
+ * Copy to either user or kernel space
+ */
+int copy_to_user_or_kernel(void *target, void *src, size_t size, bool userbuf)
+{
+ if (userbuf) {
+ if (copy_to_user((char __user *) target, src, size))
+ return -EFAULT;
+ } else {
+ memcpy(target, src, size);
+ }
+ return 0;
+}
+EXPORT_SYMBOL(copy_to_user_or_kernel);
--
2.1.0
^ permalink raw reply related [flat|nested] 61+ messages in thread
* [PATCH v2 1/2] kdump: vmcore: remove copy_to() and add copy_to_user_or_kernel()
@ 2021-12-11 3:33 ` Tiezhu Yang
0 siblings, 0 replies; 61+ messages in thread
From: Tiezhu Yang @ 2021-12-11 3:33 UTC (permalink / raw)
To: Dave Young, Baoquan He, Vivek Goyal, Andrew Morton
Cc: linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li
In arch/*/kernel/crash_dump*.c, there exist many similar code
about copy_oldmem_page(), remove copy_to() in fs/proc/vmcore.c
and add copy_to_user_or_kernel() in lib/usercopy.c, then we can
use copy_to_user_or_kernel() to simplify the related code.
Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
---
fs/proc/vmcore.c | 28 +++++++---------------------
include/linux/uaccess.h | 1 +
lib/usercopy.c | 15 +++++++++++++++
3 files changed, 23 insertions(+), 21 deletions(-)
diff --git a/fs/proc/vmcore.c b/fs/proc/vmcore.c
index 509f851..f67fd77 100644
--- a/fs/proc/vmcore.c
+++ b/fs/proc/vmcore.c
@@ -238,22 +238,8 @@ copy_oldmem_page_encrypted(unsigned long pfn, char *buf, size_t csize,
return copy_oldmem_page(pfn, buf, csize, offset, userbuf);
}
-/*
- * Copy to either kernel or user space
- */
-static int copy_to(void *target, void *src, size_t size, int userbuf)
-{
- if (userbuf) {
- if (copy_to_user((char __user *) target, src, size))
- return -EFAULT;
- } else {
- memcpy(target, src, size);
- }
- return 0;
-}
-
#ifdef CONFIG_PROC_VMCORE_DEVICE_DUMP
-static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, int userbuf)
+static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, bool userbuf)
{
struct vmcoredd_node *dump;
u64 offset = 0;
@@ -266,7 +252,7 @@ static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, int userbuf)
if (start < offset + dump->size) {
tsz = min(offset + (u64)dump->size - start, (u64)size);
buf = dump->buf + start - offset;
- if (copy_to(dst, buf, tsz, userbuf)) {
+ if (copy_to_user_or_kernel(dst, buf, tsz, userbuf)) {
ret = -EFAULT;
goto out_unlock;
}
@@ -330,7 +316,7 @@ static int vmcoredd_mmap_dumps(struct vm_area_struct *vma, unsigned long dst,
* returned otherwise number of bytes read are returned.
*/
static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
- int userbuf)
+ bool userbuf)
{
ssize_t acc = 0, tmp;
size_t tsz;
@@ -347,7 +333,7 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
/* Read ELF core header */
if (*fpos < elfcorebuf_sz) {
tsz = min(elfcorebuf_sz - (size_t)*fpos, buflen);
- if (copy_to(buffer, elfcorebuf + *fpos, tsz, userbuf))
+ if (copy_to_user_or_kernel(buffer, elfcorebuf + *fpos, tsz, userbuf))
return -EFAULT;
buflen -= tsz;
*fpos += tsz;
@@ -395,7 +381,7 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
/* Read remaining elf notes */
tsz = min(elfcorebuf_sz + elfnotes_sz - (size_t)*fpos, buflen);
kaddr = elfnotes_buf + *fpos - elfcorebuf_sz - vmcoredd_orig_sz;
- if (copy_to(buffer, kaddr, tsz, userbuf))
+ if (copy_to_user_or_kernel(buffer, kaddr, tsz, userbuf))
return -EFAULT;
buflen -= tsz;
@@ -435,7 +421,7 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
static ssize_t read_vmcore(struct file *file, char __user *buffer,
size_t buflen, loff_t *fpos)
{
- return __read_vmcore((__force char *) buffer, buflen, fpos, 1);
+ return __read_vmcore((__force char *) buffer, buflen, fpos, true);
}
/*
@@ -461,7 +447,7 @@ static vm_fault_t mmap_vmcore_fault(struct vm_fault *vmf)
if (!PageUptodate(page)) {
offset = (loff_t) index << PAGE_SHIFT;
buf = __va((page_to_pfn(page) << PAGE_SHIFT));
- rc = __read_vmcore(buf, PAGE_SIZE, &offset, 0);
+ rc = __read_vmcore(buf, PAGE_SIZE, &offset, false);
if (rc < 0) {
unlock_page(page);
put_page(page);
diff --git a/include/linux/uaccess.h b/include/linux/uaccess.h
index ac03940..a25e682e 100644
--- a/include/linux/uaccess.h
+++ b/include/linux/uaccess.h
@@ -283,6 +283,7 @@ __copy_from_user_inatomic_nocache(void *to, const void __user *from,
#endif /* ARCH_HAS_NOCACHE_UACCESS */
extern __must_check int check_zeroed_user(const void __user *from, size_t size);
+extern __must_check int copy_to_user_or_kernel(void *target, void *src, size_t size, bool userbuf);
/**
* copy_struct_from_user: copy a struct from userspace
diff --git a/lib/usercopy.c b/lib/usercopy.c
index 7413dd3..7431b1b 100644
--- a/lib/usercopy.c
+++ b/lib/usercopy.c
@@ -90,3 +90,18 @@ int check_zeroed_user(const void __user *from, size_t size)
return -EFAULT;
}
EXPORT_SYMBOL(check_zeroed_user);
+
+/*
+ * Copy to either user or kernel space
+ */
+int copy_to_user_or_kernel(void *target, void *src, size_t size, bool userbuf)
+{
+ if (userbuf) {
+ if (copy_to_user((char __user *) target, src, size))
+ return -EFAULT;
+ } else {
+ memcpy(target, src, size);
+ }
+ return 0;
+}
+EXPORT_SYMBOL(copy_to_user_or_kernel);
--
2.1.0
_______________________________________________
linux-riscv mailing list
linux-riscv@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-riscv
^ permalink raw reply related [flat|nested] 61+ messages in thread
* [PATCH v2 1/2] kdump: vmcore: remove copy_to() and add copy_to_user_or_kernel()
@ 2021-12-11 3:33 ` Tiezhu Yang
0 siblings, 0 replies; 61+ messages in thread
From: Tiezhu Yang @ 2021-12-11 3:33 UTC (permalink / raw)
To: Dave Young, Baoquan He, Vivek Goyal, Andrew Morton
Cc: linux-ia64, linux-sh, Xuefeng Li, x86, kexec, linux-mips,
linux-kernel, linux-fsdevel, linux-riscv, linuxppc-dev,
linux-arm-kernel
In arch/*/kernel/crash_dump*.c, there exist many similar code
about copy_oldmem_page(), remove copy_to() in fs/proc/vmcore.c
and add copy_to_user_or_kernel() in lib/usercopy.c, then we can
use copy_to_user_or_kernel() to simplify the related code.
Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
---
fs/proc/vmcore.c | 28 +++++++---------------------
include/linux/uaccess.h | 1 +
lib/usercopy.c | 15 +++++++++++++++
3 files changed, 23 insertions(+), 21 deletions(-)
diff --git a/fs/proc/vmcore.c b/fs/proc/vmcore.c
index 509f851..f67fd77 100644
--- a/fs/proc/vmcore.c
+++ b/fs/proc/vmcore.c
@@ -238,22 +238,8 @@ copy_oldmem_page_encrypted(unsigned long pfn, char *buf, size_t csize,
return copy_oldmem_page(pfn, buf, csize, offset, userbuf);
}
-/*
- * Copy to either kernel or user space
- */
-static int copy_to(void *target, void *src, size_t size, int userbuf)
-{
- if (userbuf) {
- if (copy_to_user((char __user *) target, src, size))
- return -EFAULT;
- } else {
- memcpy(target, src, size);
- }
- return 0;
-}
-
#ifdef CONFIG_PROC_VMCORE_DEVICE_DUMP
-static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, int userbuf)
+static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, bool userbuf)
{
struct vmcoredd_node *dump;
u64 offset = 0;
@@ -266,7 +252,7 @@ static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, int userbuf)
if (start < offset + dump->size) {
tsz = min(offset + (u64)dump->size - start, (u64)size);
buf = dump->buf + start - offset;
- if (copy_to(dst, buf, tsz, userbuf)) {
+ if (copy_to_user_or_kernel(dst, buf, tsz, userbuf)) {
ret = -EFAULT;
goto out_unlock;
}
@@ -330,7 +316,7 @@ static int vmcoredd_mmap_dumps(struct vm_area_struct *vma, unsigned long dst,
* returned otherwise number of bytes read are returned.
*/
static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
- int userbuf)
+ bool userbuf)
{
ssize_t acc = 0, tmp;
size_t tsz;
@@ -347,7 +333,7 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
/* Read ELF core header */
if (*fpos < elfcorebuf_sz) {
tsz = min(elfcorebuf_sz - (size_t)*fpos, buflen);
- if (copy_to(buffer, elfcorebuf + *fpos, tsz, userbuf))
+ if (copy_to_user_or_kernel(buffer, elfcorebuf + *fpos, tsz, userbuf))
return -EFAULT;
buflen -= tsz;
*fpos += tsz;
@@ -395,7 +381,7 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
/* Read remaining elf notes */
tsz = min(elfcorebuf_sz + elfnotes_sz - (size_t)*fpos, buflen);
kaddr = elfnotes_buf + *fpos - elfcorebuf_sz - vmcoredd_orig_sz;
- if (copy_to(buffer, kaddr, tsz, userbuf))
+ if (copy_to_user_or_kernel(buffer, kaddr, tsz, userbuf))
return -EFAULT;
buflen -= tsz;
@@ -435,7 +421,7 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
static ssize_t read_vmcore(struct file *file, char __user *buffer,
size_t buflen, loff_t *fpos)
{
- return __read_vmcore((__force char *) buffer, buflen, fpos, 1);
+ return __read_vmcore((__force char *) buffer, buflen, fpos, true);
}
/*
@@ -461,7 +447,7 @@ static vm_fault_t mmap_vmcore_fault(struct vm_fault *vmf)
if (!PageUptodate(page)) {
offset = (loff_t) index << PAGE_SHIFT;
buf = __va((page_to_pfn(page) << PAGE_SHIFT));
- rc = __read_vmcore(buf, PAGE_SIZE, &offset, 0);
+ rc = __read_vmcore(buf, PAGE_SIZE, &offset, false);
if (rc < 0) {
unlock_page(page);
put_page(page);
diff --git a/include/linux/uaccess.h b/include/linux/uaccess.h
index ac03940..a25e682e 100644
--- a/include/linux/uaccess.h
+++ b/include/linux/uaccess.h
@@ -283,6 +283,7 @@ __copy_from_user_inatomic_nocache(void *to, const void __user *from,
#endif /* ARCH_HAS_NOCACHE_UACCESS */
extern __must_check int check_zeroed_user(const void __user *from, size_t size);
+extern __must_check int copy_to_user_or_kernel(void *target, void *src, size_t size, bool userbuf);
/**
* copy_struct_from_user: copy a struct from userspace
diff --git a/lib/usercopy.c b/lib/usercopy.c
index 7413dd3..7431b1b 100644
--- a/lib/usercopy.c
+++ b/lib/usercopy.c
@@ -90,3 +90,18 @@ int check_zeroed_user(const void __user *from, size_t size)
return -EFAULT;
}
EXPORT_SYMBOL(check_zeroed_user);
+
+/*
+ * Copy to either user or kernel space
+ */
+int copy_to_user_or_kernel(void *target, void *src, size_t size, bool userbuf)
+{
+ if (userbuf) {
+ if (copy_to_user((char __user *) target, src, size))
+ return -EFAULT;
+ } else {
+ memcpy(target, src, size);
+ }
+ return 0;
+}
+EXPORT_SYMBOL(copy_to_user_or_kernel);
--
2.1.0
^ permalink raw reply related [flat|nested] 61+ messages in thread
* [PATCH v2 1/2] kdump: vmcore: remove copy_to() and add copy_to_user_or_kernel()
@ 2021-12-11 3:33 ` Tiezhu Yang
0 siblings, 0 replies; 61+ messages in thread
From: Tiezhu Yang @ 2021-12-11 3:33 UTC (permalink / raw)
To: Dave Young, Baoquan He, Vivek Goyal, Andrew Morton
Cc: linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li
In arch/*/kernel/crash_dump*.c, there exist many similar code
about copy_oldmem_page(), remove copy_to() in fs/proc/vmcore.c
and add copy_to_user_or_kernel() in lib/usercopy.c, then we can
use copy_to_user_or_kernel() to simplify the related code.
Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
---
fs/proc/vmcore.c | 28 +++++++---------------------
include/linux/uaccess.h | 1 +
lib/usercopy.c | 15 +++++++++++++++
3 files changed, 23 insertions(+), 21 deletions(-)
diff --git a/fs/proc/vmcore.c b/fs/proc/vmcore.c
index 509f851..f67fd77 100644
--- a/fs/proc/vmcore.c
+++ b/fs/proc/vmcore.c
@@ -238,22 +238,8 @@ copy_oldmem_page_encrypted(unsigned long pfn, char *buf, size_t csize,
return copy_oldmem_page(pfn, buf, csize, offset, userbuf);
}
-/*
- * Copy to either kernel or user space
- */
-static int copy_to(void *target, void *src, size_t size, int userbuf)
-{
- if (userbuf) {
- if (copy_to_user((char __user *) target, src, size))
- return -EFAULT;
- } else {
- memcpy(target, src, size);
- }
- return 0;
-}
-
#ifdef CONFIG_PROC_VMCORE_DEVICE_DUMP
-static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, int userbuf)
+static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, bool userbuf)
{
struct vmcoredd_node *dump;
u64 offset = 0;
@@ -266,7 +252,7 @@ static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, int userbuf)
if (start < offset + dump->size) {
tsz = min(offset + (u64)dump->size - start, (u64)size);
buf = dump->buf + start - offset;
- if (copy_to(dst, buf, tsz, userbuf)) {
+ if (copy_to_user_or_kernel(dst, buf, tsz, userbuf)) {
ret = -EFAULT;
goto out_unlock;
}
@@ -330,7 +316,7 @@ static int vmcoredd_mmap_dumps(struct vm_area_struct *vma, unsigned long dst,
* returned otherwise number of bytes read are returned.
*/
static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
- int userbuf)
+ bool userbuf)
{
ssize_t acc = 0, tmp;
size_t tsz;
@@ -347,7 +333,7 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
/* Read ELF core header */
if (*fpos < elfcorebuf_sz) {
tsz = min(elfcorebuf_sz - (size_t)*fpos, buflen);
- if (copy_to(buffer, elfcorebuf + *fpos, tsz, userbuf))
+ if (copy_to_user_or_kernel(buffer, elfcorebuf + *fpos, tsz, userbuf))
return -EFAULT;
buflen -= tsz;
*fpos += tsz;
@@ -395,7 +381,7 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
/* Read remaining elf notes */
tsz = min(elfcorebuf_sz + elfnotes_sz - (size_t)*fpos, buflen);
kaddr = elfnotes_buf + *fpos - elfcorebuf_sz - vmcoredd_orig_sz;
- if (copy_to(buffer, kaddr, tsz, userbuf))
+ if (copy_to_user_or_kernel(buffer, kaddr, tsz, userbuf))
return -EFAULT;
buflen -= tsz;
@@ -435,7 +421,7 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
static ssize_t read_vmcore(struct file *file, char __user *buffer,
size_t buflen, loff_t *fpos)
{
- return __read_vmcore((__force char *) buffer, buflen, fpos, 1);
+ return __read_vmcore((__force char *) buffer, buflen, fpos, true);
}
/*
@@ -461,7 +447,7 @@ static vm_fault_t mmap_vmcore_fault(struct vm_fault *vmf)
if (!PageUptodate(page)) {
offset = (loff_t) index << PAGE_SHIFT;
buf = __va((page_to_pfn(page) << PAGE_SHIFT));
- rc = __read_vmcore(buf, PAGE_SIZE, &offset, 0);
+ rc = __read_vmcore(buf, PAGE_SIZE, &offset, false);
if (rc < 0) {
unlock_page(page);
put_page(page);
diff --git a/include/linux/uaccess.h b/include/linux/uaccess.h
index ac03940..a25e682e 100644
--- a/include/linux/uaccess.h
+++ b/include/linux/uaccess.h
@@ -283,6 +283,7 @@ __copy_from_user_inatomic_nocache(void *to, const void __user *from,
#endif /* ARCH_HAS_NOCACHE_UACCESS */
extern __must_check int check_zeroed_user(const void __user *from, size_t size);
+extern __must_check int copy_to_user_or_kernel(void *target, void *src, size_t size, bool userbuf);
/**
* copy_struct_from_user: copy a struct from userspace
diff --git a/lib/usercopy.c b/lib/usercopy.c
index 7413dd3..7431b1b 100644
--- a/lib/usercopy.c
+++ b/lib/usercopy.c
@@ -90,3 +90,18 @@ int check_zeroed_user(const void __user *from, size_t size)
return -EFAULT;
}
EXPORT_SYMBOL(check_zeroed_user);
+
+/*
+ * Copy to either user or kernel space
+ */
+int copy_to_user_or_kernel(void *target, void *src, size_t size, bool userbuf)
+{
+ if (userbuf) {
+ if (copy_to_user((char __user *) target, src, size))
+ return -EFAULT;
+ } else {
+ memcpy(target, src, size);
+ }
+ return 0;
+}
+EXPORT_SYMBOL(copy_to_user_or_kernel);
--
2.1.0
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply related [flat|nested] 61+ messages in thread
* [PATCH v2 1/2] kdump: vmcore: remove copy_to() and add copy_to_user_or_kernel()
@ 2021-12-11 3:33 ` Tiezhu Yang
0 siblings, 0 replies; 61+ messages in thread
From: Tiezhu Yang @ 2021-12-11 3:33 UTC (permalink / raw)
To: Dave Young, Baoquan He, Vivek Goyal, Andrew Morton
Cc: linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li
In arch/*/kernel/crash_dump*.c, there exist many similar code
about copy_oldmem_page(), remove copy_to() in fs/proc/vmcore.c
and add copy_to_user_or_kernel() in lib/usercopy.c, then we can
use copy_to_user_or_kernel() to simplify the related code.
Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
---
fs/proc/vmcore.c | 28 +++++++---------------------
include/linux/uaccess.h | 1 +
lib/usercopy.c | 15 +++++++++++++++
3 files changed, 23 insertions(+), 21 deletions(-)
diff --git a/fs/proc/vmcore.c b/fs/proc/vmcore.c
index 509f851..f67fd77 100644
--- a/fs/proc/vmcore.c
+++ b/fs/proc/vmcore.c
@@ -238,22 +238,8 @@ copy_oldmem_page_encrypted(unsigned long pfn, char *buf, size_t csize,
return copy_oldmem_page(pfn, buf, csize, offset, userbuf);
}
-/*
- * Copy to either kernel or user space
- */
-static int copy_to(void *target, void *src, size_t size, int userbuf)
-{
- if (userbuf) {
- if (copy_to_user((char __user *) target, src, size))
- return -EFAULT;
- } else {
- memcpy(target, src, size);
- }
- return 0;
-}
-
#ifdef CONFIG_PROC_VMCORE_DEVICE_DUMP
-static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, int userbuf)
+static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, bool userbuf)
{
struct vmcoredd_node *dump;
u64 offset = 0;
@@ -266,7 +252,7 @@ static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, int userbuf)
if (start < offset + dump->size) {
tsz = min(offset + (u64)dump->size - start, (u64)size);
buf = dump->buf + start - offset;
- if (copy_to(dst, buf, tsz, userbuf)) {
+ if (copy_to_user_or_kernel(dst, buf, tsz, userbuf)) {
ret = -EFAULT;
goto out_unlock;
}
@@ -330,7 +316,7 @@ static int vmcoredd_mmap_dumps(struct vm_area_struct *vma, unsigned long dst,
* returned otherwise number of bytes read are returned.
*/
static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
- int userbuf)
+ bool userbuf)
{
ssize_t acc = 0, tmp;
size_t tsz;
@@ -347,7 +333,7 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
/* Read ELF core header */
if (*fpos < elfcorebuf_sz) {
tsz = min(elfcorebuf_sz - (size_t)*fpos, buflen);
- if (copy_to(buffer, elfcorebuf + *fpos, tsz, userbuf))
+ if (copy_to_user_or_kernel(buffer, elfcorebuf + *fpos, tsz, userbuf))
return -EFAULT;
buflen -= tsz;
*fpos += tsz;
@@ -395,7 +381,7 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
/* Read remaining elf notes */
tsz = min(elfcorebuf_sz + elfnotes_sz - (size_t)*fpos, buflen);
kaddr = elfnotes_buf + *fpos - elfcorebuf_sz - vmcoredd_orig_sz;
- if (copy_to(buffer, kaddr, tsz, userbuf))
+ if (copy_to_user_or_kernel(buffer, kaddr, tsz, userbuf))
return -EFAULT;
buflen -= tsz;
@@ -435,7 +421,7 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
static ssize_t read_vmcore(struct file *file, char __user *buffer,
size_t buflen, loff_t *fpos)
{
- return __read_vmcore((__force char *) buffer, buflen, fpos, 1);
+ return __read_vmcore((__force char *) buffer, buflen, fpos, true);
}
/*
@@ -461,7 +447,7 @@ static vm_fault_t mmap_vmcore_fault(struct vm_fault *vmf)
if (!PageUptodate(page)) {
offset = (loff_t) index << PAGE_SHIFT;
buf = __va((page_to_pfn(page) << PAGE_SHIFT));
- rc = __read_vmcore(buf, PAGE_SIZE, &offset, 0);
+ rc = __read_vmcore(buf, PAGE_SIZE, &offset, false);
if (rc < 0) {
unlock_page(page);
put_page(page);
diff --git a/include/linux/uaccess.h b/include/linux/uaccess.h
index ac03940..a25e682e 100644
--- a/include/linux/uaccess.h
+++ b/include/linux/uaccess.h
@@ -283,6 +283,7 @@ __copy_from_user_inatomic_nocache(void *to, const void __user *from,
#endif /* ARCH_HAS_NOCACHE_UACCESS */
extern __must_check int check_zeroed_user(const void __user *from, size_t size);
+extern __must_check int copy_to_user_or_kernel(void *target, void *src, size_t size, bool userbuf);
/**
* copy_struct_from_user: copy a struct from userspace
diff --git a/lib/usercopy.c b/lib/usercopy.c
index 7413dd3..7431b1b 100644
--- a/lib/usercopy.c
+++ b/lib/usercopy.c
@@ -90,3 +90,18 @@ int check_zeroed_user(const void __user *from, size_t size)
return -EFAULT;
}
EXPORT_SYMBOL(check_zeroed_user);
+
+/*
+ * Copy to either user or kernel space
+ */
+int copy_to_user_or_kernel(void *target, void *src, size_t size, bool userbuf)
+{
+ if (userbuf) {
+ if (copy_to_user((char __user *) target, src, size))
+ return -EFAULT;
+ } else {
+ memcpy(target, src, size);
+ }
+ return 0;
+}
+EXPORT_SYMBOL(copy_to_user_or_kernel);
--
2.1.0
_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec
^ permalink raw reply related [flat|nested] 61+ messages in thread
* [PATCH v2 1/2] kdump: vmcore: remove copy_to() and add copy_to_user_or_kernel()
@ 2021-12-11 3:33 ` Tiezhu Yang
0 siblings, 0 replies; 61+ messages in thread
From: Tiezhu Yang @ 2021-12-11 3:33 UTC (permalink / raw)
To: Dave Young, Baoquan He, Vivek Goyal, Andrew Morton
Cc: linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li
In arch/*/kernel/crash_dump*.c, there exist many similar code
about copy_oldmem_page(), remove copy_to() in fs/proc/vmcore.c
and add copy_to_user_or_kernel() in lib/usercopy.c, then we can
use copy_to_user_or_kernel() to simplify the related code.
Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
---
fs/proc/vmcore.c | 28 +++++++---------------------
include/linux/uaccess.h | 1 +
lib/usercopy.c | 15 +++++++++++++++
3 files changed, 23 insertions(+), 21 deletions(-)
diff --git a/fs/proc/vmcore.c b/fs/proc/vmcore.c
index 509f851..f67fd77 100644
--- a/fs/proc/vmcore.c
+++ b/fs/proc/vmcore.c
@@ -238,22 +238,8 @@ copy_oldmem_page_encrypted(unsigned long pfn, char *buf, size_t csize,
return copy_oldmem_page(pfn, buf, csize, offset, userbuf);
}
-/*
- * Copy to either kernel or user space
- */
-static int copy_to(void *target, void *src, size_t size, int userbuf)
-{
- if (userbuf) {
- if (copy_to_user((char __user *) target, src, size))
- return -EFAULT;
- } else {
- memcpy(target, src, size);
- }
- return 0;
-}
-
#ifdef CONFIG_PROC_VMCORE_DEVICE_DUMP
-static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, int userbuf)
+static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, bool userbuf)
{
struct vmcoredd_node *dump;
u64 offset = 0;
@@ -266,7 +252,7 @@ static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, int userbuf)
if (start < offset + dump->size) {
tsz = min(offset + (u64)dump->size - start, (u64)size);
buf = dump->buf + start - offset;
- if (copy_to(dst, buf, tsz, userbuf)) {
+ if (copy_to_user_or_kernel(dst, buf, tsz, userbuf)) {
ret = -EFAULT;
goto out_unlock;
}
@@ -330,7 +316,7 @@ static int vmcoredd_mmap_dumps(struct vm_area_struct *vma, unsigned long dst,
* returned otherwise number of bytes read are returned.
*/
static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
- int userbuf)
+ bool userbuf)
{
ssize_t acc = 0, tmp;
size_t tsz;
@@ -347,7 +333,7 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
/* Read ELF core header */
if (*fpos < elfcorebuf_sz) {
tsz = min(elfcorebuf_sz - (size_t)*fpos, buflen);
- if (copy_to(buffer, elfcorebuf + *fpos, tsz, userbuf))
+ if (copy_to_user_or_kernel(buffer, elfcorebuf + *fpos, tsz, userbuf))
return -EFAULT;
buflen -= tsz;
*fpos += tsz;
@@ -395,7 +381,7 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
/* Read remaining elf notes */
tsz = min(elfcorebuf_sz + elfnotes_sz - (size_t)*fpos, buflen);
kaddr = elfnotes_buf + *fpos - elfcorebuf_sz - vmcoredd_orig_sz;
- if (copy_to(buffer, kaddr, tsz, userbuf))
+ if (copy_to_user_or_kernel(buffer, kaddr, tsz, userbuf))
return -EFAULT;
buflen -= tsz;
@@ -435,7 +421,7 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
static ssize_t read_vmcore(struct file *file, char __user *buffer,
size_t buflen, loff_t *fpos)
{
- return __read_vmcore((__force char *) buffer, buflen, fpos, 1);
+ return __read_vmcore((__force char *) buffer, buflen, fpos, true);
}
/*
@@ -461,7 +447,7 @@ static vm_fault_t mmap_vmcore_fault(struct vm_fault *vmf)
if (!PageUptodate(page)) {
offset = (loff_t) index << PAGE_SHIFT;
buf = __va((page_to_pfn(page) << PAGE_SHIFT));
- rc = __read_vmcore(buf, PAGE_SIZE, &offset, 0);
+ rc = __read_vmcore(buf, PAGE_SIZE, &offset, false);
if (rc < 0) {
unlock_page(page);
put_page(page);
diff --git a/include/linux/uaccess.h b/include/linux/uaccess.h
index ac03940..a25e682e 100644
--- a/include/linux/uaccess.h
+++ b/include/linux/uaccess.h
@@ -283,6 +283,7 @@ __copy_from_user_inatomic_nocache(void *to, const void __user *from,
#endif /* ARCH_HAS_NOCACHE_UACCESS */
extern __must_check int check_zeroed_user(const void __user *from, size_t size);
+extern __must_check int copy_to_user_or_kernel(void *target, void *src, size_t size, bool userbuf);
/**
* copy_struct_from_user: copy a struct from userspace
diff --git a/lib/usercopy.c b/lib/usercopy.c
index 7413dd3..7431b1b 100644
--- a/lib/usercopy.c
+++ b/lib/usercopy.c
@@ -90,3 +90,18 @@ int check_zeroed_user(const void __user *from, size_t size)
return -EFAULT;
}
EXPORT_SYMBOL(check_zeroed_user);
+
+/*
+ * Copy to either user or kernel space
+ */
+int copy_to_user_or_kernel(void *target, void *src, size_t size, bool userbuf)
+{
+ if (userbuf) {
+ if (copy_to_user((char __user *) target, src, size))
+ return -EFAULT;
+ } else {
+ memcpy(target, src, size);
+ }
+ return 0;
+}
+EXPORT_SYMBOL(copy_to_user_or_kernel);
--
2.1.0
^ permalink raw reply related [flat|nested] 61+ messages in thread
* [PATCH v2 2/2] kdump: crashdump: use copy_to_user_or_kernel() to simplify code
2021-12-11 3:33 ` Tiezhu Yang
` (3 preceding siblings ...)
(?)
@ 2021-12-11 3:33 ` Tiezhu Yang
-1 siblings, 0 replies; 61+ messages in thread
From: Tiezhu Yang @ 2021-12-11 3:33 UTC (permalink / raw)
To: Dave Young, Baoquan He, Vivek Goyal, Andrew Morton
Cc: linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li
Use copy_to_user_or_kernel() to simplify the related code about
copy_oldmem_page() in arch/*/kernel/crash_dump*.c files.
Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
---
arch/arm/kernel/crash_dump.c | 12 +++---------
arch/arm64/kernel/crash_dump.c | 12 +++---------
arch/ia64/kernel/crash_dump.c | 12 +++++-------
arch/mips/kernel/crash_dump.c | 11 +++--------
arch/powerpc/kernel/crash_dump.c | 11 ++++-------
arch/riscv/kernel/crash_dump.c | 11 +++--------
arch/sh/kernel/crash_dump.c | 11 +++--------
arch/x86/kernel/crash_dump_32.c | 11 +++--------
arch/x86/kernel/crash_dump_64.c | 15 +++++----------
fs/proc/vmcore.c | 4 ++--
include/linux/crash_dump.h | 8 ++++----
11 files changed, 38 insertions(+), 80 deletions(-)
diff --git a/arch/arm/kernel/crash_dump.c b/arch/arm/kernel/crash_dump.c
index 53cb924..a27c5df 100644
--- a/arch/arm/kernel/crash_dump.c
+++ b/arch/arm/kernel/crash_dump.c
@@ -29,7 +29,7 @@
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
size_t csize, unsigned long offset,
- int userbuf)
+ bool userbuf)
{
void *vaddr;
@@ -40,14 +40,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
if (!vaddr)
return -ENOMEM;
- if (userbuf) {
- if (copy_to_user(buf, vaddr + offset, csize)) {
- iounmap(vaddr);
- return -EFAULT;
- }
- } else {
- memcpy(buf, vaddr + offset, csize);
- }
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
iounmap(vaddr);
return csize;
diff --git a/arch/arm64/kernel/crash_dump.c b/arch/arm64/kernel/crash_dump.c
index 58303a9..d22988f 100644
--- a/arch/arm64/kernel/crash_dump.c
+++ b/arch/arm64/kernel/crash_dump.c
@@ -27,7 +27,7 @@
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
size_t csize, unsigned long offset,
- int userbuf)
+ bool userbuf)
{
void *vaddr;
@@ -38,14 +38,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
if (!vaddr)
return -ENOMEM;
- if (userbuf) {
- if (copy_to_user((char __user *)buf, vaddr + offset, csize)) {
- memunmap(vaddr);
- return -EFAULT;
- }
- } else {
- memcpy(buf, vaddr + offset, csize);
- }
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
memunmap(vaddr);
diff --git a/arch/ia64/kernel/crash_dump.c b/arch/ia64/kernel/crash_dump.c
index 0ed3c3d..12128f8 100644
--- a/arch/ia64/kernel/crash_dump.c
+++ b/arch/ia64/kernel/crash_dump.c
@@ -33,19 +33,17 @@
*/
ssize_t
copy_oldmem_page(unsigned long pfn, char *buf,
- size_t csize, unsigned long offset, int userbuf)
+ size_t csize, unsigned long offset, bool userbuf)
{
void *vaddr;
if (!csize)
return 0;
+
vaddr = __va(pfn<<PAGE_SHIFT);
- if (userbuf) {
- if (copy_to_user(buf, (vaddr + offset), csize)) {
- return -EFAULT;
- }
- } else
- memcpy(buf, (vaddr + offset), csize);
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ return -EFAULT;
+
return csize;
}
diff --git a/arch/mips/kernel/crash_dump.c b/arch/mips/kernel/crash_dump.c
index 2e50f551..7670915 100644
--- a/arch/mips/kernel/crash_dump.c
+++ b/arch/mips/kernel/crash_dump.c
@@ -16,7 +16,7 @@
* in the current kernel.
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
- size_t csize, unsigned long offset, int userbuf)
+ size_t csize, unsigned long offset, bool userbuf)
{
void *vaddr;
@@ -24,13 +24,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
return 0;
vaddr = kmap_local_pfn(pfn);
-
- if (!userbuf) {
- memcpy(buf, vaddr + offset, csize);
- } else {
- if (copy_to_user(buf, vaddr + offset, csize))
- csize = -EFAULT;
- }
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
kunmap_local(vaddr);
diff --git a/arch/powerpc/kernel/crash_dump.c b/arch/powerpc/kernel/crash_dump.c
index 5693e1c67..e2e9612 100644
--- a/arch/powerpc/kernel/crash_dump.c
+++ b/arch/powerpc/kernel/crash_dump.c
@@ -69,13 +69,10 @@ void __init setup_kdump_trampoline(void)
#endif /* CONFIG_NONSTATIC_KERNEL */
static size_t copy_oldmem_vaddr(void *vaddr, char *buf, size_t csize,
- unsigned long offset, int userbuf)
+ unsigned long offset, bool userbuf)
{
- if (userbuf) {
- if (copy_to_user((char __user *)buf, (vaddr + offset), csize))
- return -EFAULT;
- } else
- memcpy(buf, (vaddr + offset), csize);
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ return -EFAULT;
return csize;
}
@@ -94,7 +91,7 @@ static size_t copy_oldmem_vaddr(void *vaddr, char *buf, size_t csize,
* in the current kernel. We stitch up a pte, similar to kmap_atomic.
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
- size_t csize, unsigned long offset, int userbuf)
+ size_t csize, unsigned long offset, bool userbuf)
{
void *vaddr;
phys_addr_t paddr;
diff --git a/arch/riscv/kernel/crash_dump.c b/arch/riscv/kernel/crash_dump.c
index 86cc0ad..4167437 100644
--- a/arch/riscv/kernel/crash_dump.c
+++ b/arch/riscv/kernel/crash_dump.c
@@ -22,7 +22,7 @@
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
size_t csize, unsigned long offset,
- int userbuf)
+ bool userbuf)
{
void *vaddr;
@@ -33,13 +33,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
if (!vaddr)
return -ENOMEM;
- if (userbuf) {
- if (copy_to_user((char __user *)buf, vaddr + offset, csize)) {
- memunmap(vaddr);
- return -EFAULT;
- }
- } else
- memcpy(buf, vaddr + offset, csize);
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
memunmap(vaddr);
return csize;
diff --git a/arch/sh/kernel/crash_dump.c b/arch/sh/kernel/crash_dump.c
index 5b41b59..4bc071a 100644
--- a/arch/sh/kernel/crash_dump.c
+++ b/arch/sh/kernel/crash_dump.c
@@ -24,7 +24,7 @@
* in the current kernel. We stitch up a pte, similar to kmap_atomic.
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
- size_t csize, unsigned long offset, int userbuf)
+ size_t csize, unsigned long offset, bool userbuf)
{
void __iomem *vaddr;
@@ -33,13 +33,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
vaddr = ioremap(pfn << PAGE_SHIFT, PAGE_SIZE);
- if (userbuf) {
- if (copy_to_user((void __user *)buf, (vaddr + offset), csize)) {
- iounmap(vaddr);
- return -EFAULT;
- }
- } else
- memcpy(buf, (vaddr + offset), csize);
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
iounmap(vaddr);
return csize;
diff --git a/arch/x86/kernel/crash_dump_32.c b/arch/x86/kernel/crash_dump_32.c
index 5fcac46..3eff124 100644
--- a/arch/x86/kernel/crash_dump_32.c
+++ b/arch/x86/kernel/crash_dump_32.c
@@ -43,7 +43,7 @@ static inline bool is_crashed_pfn_valid(unsigned long pfn)
* in the current kernel.
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
- unsigned long offset, int userbuf)
+ unsigned long offset, bool userbuf)
{
void *vaddr;
@@ -54,13 +54,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
return -EFAULT;
vaddr = kmap_local_pfn(pfn);
-
- if (!userbuf) {
- memcpy(buf, vaddr + offset, csize);
- } else {
- if (copy_to_user(buf, vaddr + offset, csize))
- csize = -EFAULT;
- }
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
kunmap_local(vaddr);
diff --git a/arch/x86/kernel/crash_dump_64.c b/arch/x86/kernel/crash_dump_64.c
index a7f617a..e8fffdf 100644
--- a/arch/x86/kernel/crash_dump_64.c
+++ b/arch/x86/kernel/crash_dump_64.c
@@ -13,7 +13,7 @@
#include <linux/cc_platform.h>
static ssize_t __copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
- unsigned long offset, int userbuf,
+ unsigned long offset, bool userbuf,
bool encrypted)
{
void *vaddr;
@@ -29,13 +29,8 @@ static ssize_t __copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
if (!vaddr)
return -ENOMEM;
- if (userbuf) {
- if (copy_to_user((void __user *)buf, vaddr + offset, csize)) {
- iounmap((void __iomem *)vaddr);
- return -EFAULT;
- }
- } else
- memcpy(buf, vaddr + offset, csize);
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
set_iounmap_nonlazy();
iounmap((void __iomem *)vaddr);
@@ -56,7 +51,7 @@ static ssize_t __copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
* mapped in the current kernel. We stitch up a pte, similar to kmap_atomic.
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
- unsigned long offset, int userbuf)
+ unsigned long offset, bool userbuf)
{
return __copy_oldmem_page(pfn, buf, csize, offset, userbuf, false);
}
@@ -67,7 +62,7 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
* machines.
*/
ssize_t copy_oldmem_page_encrypted(unsigned long pfn, char *buf, size_t csize,
- unsigned long offset, int userbuf)
+ unsigned long offset, bool userbuf)
{
return __copy_oldmem_page(pfn, buf, csize, offset, userbuf, true);
}
diff --git a/fs/proc/vmcore.c b/fs/proc/vmcore.c
index f67fd77..bba52aa 100644
--- a/fs/proc/vmcore.c
+++ b/fs/proc/vmcore.c
@@ -133,7 +133,7 @@ static int open_vmcore(struct inode *inode, struct file *file)
/* Reads a page from the oldmem device from given offset. */
ssize_t read_from_oldmem(char *buf, size_t count,
- u64 *ppos, int userbuf,
+ u64 *ppos, bool userbuf,
bool encrypted)
{
unsigned long pfn, offset;
@@ -233,7 +233,7 @@ int __weak remap_oldmem_pfn_range(struct vm_area_struct *vma,
*/
ssize_t __weak
copy_oldmem_page_encrypted(unsigned long pfn, char *buf, size_t csize,
- unsigned long offset, int userbuf)
+ unsigned long offset, bool userbuf)
{
return copy_oldmem_page(pfn, buf, csize, offset, userbuf);
}
diff --git a/include/linux/crash_dump.h b/include/linux/crash_dump.h
index 6208215..033448b 100644
--- a/include/linux/crash_dump.h
+++ b/include/linux/crash_dump.h
@@ -25,10 +25,10 @@ extern int remap_oldmem_pfn_range(struct vm_area_struct *vma,
unsigned long size, pgprot_t prot);
extern ssize_t copy_oldmem_page(unsigned long, char *, size_t,
- unsigned long, int);
+ unsigned long, bool);
extern ssize_t copy_oldmem_page_encrypted(unsigned long pfn, char *buf,
size_t csize, unsigned long offset,
- int userbuf);
+ bool userbuf);
void vmcore_cleanup(void);
@@ -136,11 +136,11 @@ static inline int vmcore_add_device_dump(struct vmcoredd_data *data)
#ifdef CONFIG_PROC_VMCORE
ssize_t read_from_oldmem(char *buf, size_t count,
- u64 *ppos, int userbuf,
+ u64 *ppos, bool userbuf,
bool encrypted);
#else
static inline ssize_t read_from_oldmem(char *buf, size_t count,
- u64 *ppos, int userbuf,
+ u64 *ppos, bool userbuf,
bool encrypted)
{
return -EOPNOTSUPP;
--
2.1.0
^ permalink raw reply related [flat|nested] 61+ messages in thread
* [PATCH v2 2/2] kdump: crashdump: use copy_to_user_or_kernel() to simplify code
@ 2021-12-11 3:33 ` Tiezhu Yang
0 siblings, 0 replies; 61+ messages in thread
From: Tiezhu Yang @ 2021-12-11 3:33 UTC (permalink / raw)
To: Dave Young, Baoquan He, Vivek Goyal, Andrew Morton
Cc: linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li
Use copy_to_user_or_kernel() to simplify the related code about
copy_oldmem_page() in arch/*/kernel/crash_dump*.c files.
Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
---
arch/arm/kernel/crash_dump.c | 12 +++---------
arch/arm64/kernel/crash_dump.c | 12 +++---------
arch/ia64/kernel/crash_dump.c | 12 +++++-------
arch/mips/kernel/crash_dump.c | 11 +++--------
arch/powerpc/kernel/crash_dump.c | 11 ++++-------
arch/riscv/kernel/crash_dump.c | 11 +++--------
arch/sh/kernel/crash_dump.c | 11 +++--------
arch/x86/kernel/crash_dump_32.c | 11 +++--------
arch/x86/kernel/crash_dump_64.c | 15 +++++----------
fs/proc/vmcore.c | 4 ++--
include/linux/crash_dump.h | 8 ++++----
11 files changed, 38 insertions(+), 80 deletions(-)
diff --git a/arch/arm/kernel/crash_dump.c b/arch/arm/kernel/crash_dump.c
index 53cb924..a27c5df 100644
--- a/arch/arm/kernel/crash_dump.c
+++ b/arch/arm/kernel/crash_dump.c
@@ -29,7 +29,7 @@
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
size_t csize, unsigned long offset,
- int userbuf)
+ bool userbuf)
{
void *vaddr;
@@ -40,14 +40,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
if (!vaddr)
return -ENOMEM;
- if (userbuf) {
- if (copy_to_user(buf, vaddr + offset, csize)) {
- iounmap(vaddr);
- return -EFAULT;
- }
- } else {
- memcpy(buf, vaddr + offset, csize);
- }
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
iounmap(vaddr);
return csize;
diff --git a/arch/arm64/kernel/crash_dump.c b/arch/arm64/kernel/crash_dump.c
index 58303a9..d22988f 100644
--- a/arch/arm64/kernel/crash_dump.c
+++ b/arch/arm64/kernel/crash_dump.c
@@ -27,7 +27,7 @@
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
size_t csize, unsigned long offset,
- int userbuf)
+ bool userbuf)
{
void *vaddr;
@@ -38,14 +38,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
if (!vaddr)
return -ENOMEM;
- if (userbuf) {
- if (copy_to_user((char __user *)buf, vaddr + offset, csize)) {
- memunmap(vaddr);
- return -EFAULT;
- }
- } else {
- memcpy(buf, vaddr + offset, csize);
- }
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
memunmap(vaddr);
diff --git a/arch/ia64/kernel/crash_dump.c b/arch/ia64/kernel/crash_dump.c
index 0ed3c3d..12128f8 100644
--- a/arch/ia64/kernel/crash_dump.c
+++ b/arch/ia64/kernel/crash_dump.c
@@ -33,19 +33,17 @@
*/
ssize_t
copy_oldmem_page(unsigned long pfn, char *buf,
- size_t csize, unsigned long offset, int userbuf)
+ size_t csize, unsigned long offset, bool userbuf)
{
void *vaddr;
if (!csize)
return 0;
+
vaddr = __va(pfn<<PAGE_SHIFT);
- if (userbuf) {
- if (copy_to_user(buf, (vaddr + offset), csize)) {
- return -EFAULT;
- }
- } else
- memcpy(buf, (vaddr + offset), csize);
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ return -EFAULT;
+
return csize;
}
diff --git a/arch/mips/kernel/crash_dump.c b/arch/mips/kernel/crash_dump.c
index 2e50f551..7670915 100644
--- a/arch/mips/kernel/crash_dump.c
+++ b/arch/mips/kernel/crash_dump.c
@@ -16,7 +16,7 @@
* in the current kernel.
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
- size_t csize, unsigned long offset, int userbuf)
+ size_t csize, unsigned long offset, bool userbuf)
{
void *vaddr;
@@ -24,13 +24,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
return 0;
vaddr = kmap_local_pfn(pfn);
-
- if (!userbuf) {
- memcpy(buf, vaddr + offset, csize);
- } else {
- if (copy_to_user(buf, vaddr + offset, csize))
- csize = -EFAULT;
- }
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
kunmap_local(vaddr);
diff --git a/arch/powerpc/kernel/crash_dump.c b/arch/powerpc/kernel/crash_dump.c
index 5693e1c67..e2e9612 100644
--- a/arch/powerpc/kernel/crash_dump.c
+++ b/arch/powerpc/kernel/crash_dump.c
@@ -69,13 +69,10 @@ void __init setup_kdump_trampoline(void)
#endif /* CONFIG_NONSTATIC_KERNEL */
static size_t copy_oldmem_vaddr(void *vaddr, char *buf, size_t csize,
- unsigned long offset, int userbuf)
+ unsigned long offset, bool userbuf)
{
- if (userbuf) {
- if (copy_to_user((char __user *)buf, (vaddr + offset), csize))
- return -EFAULT;
- } else
- memcpy(buf, (vaddr + offset), csize);
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ return -EFAULT;
return csize;
}
@@ -94,7 +91,7 @@ static size_t copy_oldmem_vaddr(void *vaddr, char *buf, size_t csize,
* in the current kernel. We stitch up a pte, similar to kmap_atomic.
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
- size_t csize, unsigned long offset, int userbuf)
+ size_t csize, unsigned long offset, bool userbuf)
{
void *vaddr;
phys_addr_t paddr;
diff --git a/arch/riscv/kernel/crash_dump.c b/arch/riscv/kernel/crash_dump.c
index 86cc0ad..4167437 100644
--- a/arch/riscv/kernel/crash_dump.c
+++ b/arch/riscv/kernel/crash_dump.c
@@ -22,7 +22,7 @@
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
size_t csize, unsigned long offset,
- int userbuf)
+ bool userbuf)
{
void *vaddr;
@@ -33,13 +33,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
if (!vaddr)
return -ENOMEM;
- if (userbuf) {
- if (copy_to_user((char __user *)buf, vaddr + offset, csize)) {
- memunmap(vaddr);
- return -EFAULT;
- }
- } else
- memcpy(buf, vaddr + offset, csize);
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
memunmap(vaddr);
return csize;
diff --git a/arch/sh/kernel/crash_dump.c b/arch/sh/kernel/crash_dump.c
index 5b41b59..4bc071a 100644
--- a/arch/sh/kernel/crash_dump.c
+++ b/arch/sh/kernel/crash_dump.c
@@ -24,7 +24,7 @@
* in the current kernel. We stitch up a pte, similar to kmap_atomic.
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
- size_t csize, unsigned long offset, int userbuf)
+ size_t csize, unsigned long offset, bool userbuf)
{
void __iomem *vaddr;
@@ -33,13 +33,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
vaddr = ioremap(pfn << PAGE_SHIFT, PAGE_SIZE);
- if (userbuf) {
- if (copy_to_user((void __user *)buf, (vaddr + offset), csize)) {
- iounmap(vaddr);
- return -EFAULT;
- }
- } else
- memcpy(buf, (vaddr + offset), csize);
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
iounmap(vaddr);
return csize;
diff --git a/arch/x86/kernel/crash_dump_32.c b/arch/x86/kernel/crash_dump_32.c
index 5fcac46..3eff124 100644
--- a/arch/x86/kernel/crash_dump_32.c
+++ b/arch/x86/kernel/crash_dump_32.c
@@ -43,7 +43,7 @@ static inline bool is_crashed_pfn_valid(unsigned long pfn)
* in the current kernel.
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
- unsigned long offset, int userbuf)
+ unsigned long offset, bool userbuf)
{
void *vaddr;
@@ -54,13 +54,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
return -EFAULT;
vaddr = kmap_local_pfn(pfn);
-
- if (!userbuf) {
- memcpy(buf, vaddr + offset, csize);
- } else {
- if (copy_to_user(buf, vaddr + offset, csize))
- csize = -EFAULT;
- }
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
kunmap_local(vaddr);
diff --git a/arch/x86/kernel/crash_dump_64.c b/arch/x86/kernel/crash_dump_64.c
index a7f617a..e8fffdf 100644
--- a/arch/x86/kernel/crash_dump_64.c
+++ b/arch/x86/kernel/crash_dump_64.c
@@ -13,7 +13,7 @@
#include <linux/cc_platform.h>
static ssize_t __copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
- unsigned long offset, int userbuf,
+ unsigned long offset, bool userbuf,
bool encrypted)
{
void *vaddr;
@@ -29,13 +29,8 @@ static ssize_t __copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
if (!vaddr)
return -ENOMEM;
- if (userbuf) {
- if (copy_to_user((void __user *)buf, vaddr + offset, csize)) {
- iounmap((void __iomem *)vaddr);
- return -EFAULT;
- }
- } else
- memcpy(buf, vaddr + offset, csize);
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
set_iounmap_nonlazy();
iounmap((void __iomem *)vaddr);
@@ -56,7 +51,7 @@ static ssize_t __copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
* mapped in the current kernel. We stitch up a pte, similar to kmap_atomic.
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
- unsigned long offset, int userbuf)
+ unsigned long offset, bool userbuf)
{
return __copy_oldmem_page(pfn, buf, csize, offset, userbuf, false);
}
@@ -67,7 +62,7 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
* machines.
*/
ssize_t copy_oldmem_page_encrypted(unsigned long pfn, char *buf, size_t csize,
- unsigned long offset, int userbuf)
+ unsigned long offset, bool userbuf)
{
return __copy_oldmem_page(pfn, buf, csize, offset, userbuf, true);
}
diff --git a/fs/proc/vmcore.c b/fs/proc/vmcore.c
index f67fd77..bba52aa 100644
--- a/fs/proc/vmcore.c
+++ b/fs/proc/vmcore.c
@@ -133,7 +133,7 @@ static int open_vmcore(struct inode *inode, struct file *file)
/* Reads a page from the oldmem device from given offset. */
ssize_t read_from_oldmem(char *buf, size_t count,
- u64 *ppos, int userbuf,
+ u64 *ppos, bool userbuf,
bool encrypted)
{
unsigned long pfn, offset;
@@ -233,7 +233,7 @@ int __weak remap_oldmem_pfn_range(struct vm_area_struct *vma,
*/
ssize_t __weak
copy_oldmem_page_encrypted(unsigned long pfn, char *buf, size_t csize,
- unsigned long offset, int userbuf)
+ unsigned long offset, bool userbuf)
{
return copy_oldmem_page(pfn, buf, csize, offset, userbuf);
}
diff --git a/include/linux/crash_dump.h b/include/linux/crash_dump.h
index 6208215..033448b 100644
--- a/include/linux/crash_dump.h
+++ b/include/linux/crash_dump.h
@@ -25,10 +25,10 @@ extern int remap_oldmem_pfn_range(struct vm_area_struct *vma,
unsigned long size, pgprot_t prot);
extern ssize_t copy_oldmem_page(unsigned long, char *, size_t,
- unsigned long, int);
+ unsigned long, bool);
extern ssize_t copy_oldmem_page_encrypted(unsigned long pfn, char *buf,
size_t csize, unsigned long offset,
- int userbuf);
+ bool userbuf);
void vmcore_cleanup(void);
@@ -136,11 +136,11 @@ static inline int vmcore_add_device_dump(struct vmcoredd_data *data)
#ifdef CONFIG_PROC_VMCORE
ssize_t read_from_oldmem(char *buf, size_t count,
- u64 *ppos, int userbuf,
+ u64 *ppos, bool userbuf,
bool encrypted);
#else
static inline ssize_t read_from_oldmem(char *buf, size_t count,
- u64 *ppos, int userbuf,
+ u64 *ppos, bool userbuf,
bool encrypted)
{
return -EOPNOTSUPP;
--
2.1.0
_______________________________________________
linux-riscv mailing list
linux-riscv@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-riscv
^ permalink raw reply related [flat|nested] 61+ messages in thread
* [PATCH v2 2/2] kdump: crashdump: use copy_to_user_or_kernel() to simplify code
@ 2021-12-11 3:33 ` Tiezhu Yang
0 siblings, 0 replies; 61+ messages in thread
From: Tiezhu Yang @ 2021-12-11 3:33 UTC (permalink / raw)
To: Dave Young, Baoquan He, Vivek Goyal, Andrew Morton
Cc: linux-ia64, linux-sh, Xuefeng Li, x86, kexec, linux-mips,
linux-kernel, linux-fsdevel, linux-riscv, linuxppc-dev,
linux-arm-kernel
Use copy_to_user_or_kernel() to simplify the related code about
copy_oldmem_page() in arch/*/kernel/crash_dump*.c files.
Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
---
arch/arm/kernel/crash_dump.c | 12 +++---------
arch/arm64/kernel/crash_dump.c | 12 +++---------
arch/ia64/kernel/crash_dump.c | 12 +++++-------
arch/mips/kernel/crash_dump.c | 11 +++--------
arch/powerpc/kernel/crash_dump.c | 11 ++++-------
arch/riscv/kernel/crash_dump.c | 11 +++--------
arch/sh/kernel/crash_dump.c | 11 +++--------
arch/x86/kernel/crash_dump_32.c | 11 +++--------
arch/x86/kernel/crash_dump_64.c | 15 +++++----------
fs/proc/vmcore.c | 4 ++--
include/linux/crash_dump.h | 8 ++++----
11 files changed, 38 insertions(+), 80 deletions(-)
diff --git a/arch/arm/kernel/crash_dump.c b/arch/arm/kernel/crash_dump.c
index 53cb924..a27c5df 100644
--- a/arch/arm/kernel/crash_dump.c
+++ b/arch/arm/kernel/crash_dump.c
@@ -29,7 +29,7 @@
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
size_t csize, unsigned long offset,
- int userbuf)
+ bool userbuf)
{
void *vaddr;
@@ -40,14 +40,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
if (!vaddr)
return -ENOMEM;
- if (userbuf) {
- if (copy_to_user(buf, vaddr + offset, csize)) {
- iounmap(vaddr);
- return -EFAULT;
- }
- } else {
- memcpy(buf, vaddr + offset, csize);
- }
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
iounmap(vaddr);
return csize;
diff --git a/arch/arm64/kernel/crash_dump.c b/arch/arm64/kernel/crash_dump.c
index 58303a9..d22988f 100644
--- a/arch/arm64/kernel/crash_dump.c
+++ b/arch/arm64/kernel/crash_dump.c
@@ -27,7 +27,7 @@
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
size_t csize, unsigned long offset,
- int userbuf)
+ bool userbuf)
{
void *vaddr;
@@ -38,14 +38,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
if (!vaddr)
return -ENOMEM;
- if (userbuf) {
- if (copy_to_user((char __user *)buf, vaddr + offset, csize)) {
- memunmap(vaddr);
- return -EFAULT;
- }
- } else {
- memcpy(buf, vaddr + offset, csize);
- }
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
memunmap(vaddr);
diff --git a/arch/ia64/kernel/crash_dump.c b/arch/ia64/kernel/crash_dump.c
index 0ed3c3d..12128f8 100644
--- a/arch/ia64/kernel/crash_dump.c
+++ b/arch/ia64/kernel/crash_dump.c
@@ -33,19 +33,17 @@
*/
ssize_t
copy_oldmem_page(unsigned long pfn, char *buf,
- size_t csize, unsigned long offset, int userbuf)
+ size_t csize, unsigned long offset, bool userbuf)
{
void *vaddr;
if (!csize)
return 0;
+
vaddr = __va(pfn<<PAGE_SHIFT);
- if (userbuf) {
- if (copy_to_user(buf, (vaddr + offset), csize)) {
- return -EFAULT;
- }
- } else
- memcpy(buf, (vaddr + offset), csize);
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ return -EFAULT;
+
return csize;
}
diff --git a/arch/mips/kernel/crash_dump.c b/arch/mips/kernel/crash_dump.c
index 2e50f551..7670915 100644
--- a/arch/mips/kernel/crash_dump.c
+++ b/arch/mips/kernel/crash_dump.c
@@ -16,7 +16,7 @@
* in the current kernel.
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
- size_t csize, unsigned long offset, int userbuf)
+ size_t csize, unsigned long offset, bool userbuf)
{
void *vaddr;
@@ -24,13 +24,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
return 0;
vaddr = kmap_local_pfn(pfn);
-
- if (!userbuf) {
- memcpy(buf, vaddr + offset, csize);
- } else {
- if (copy_to_user(buf, vaddr + offset, csize))
- csize = -EFAULT;
- }
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
kunmap_local(vaddr);
diff --git a/arch/powerpc/kernel/crash_dump.c b/arch/powerpc/kernel/crash_dump.c
index 5693e1c67..e2e9612 100644
--- a/arch/powerpc/kernel/crash_dump.c
+++ b/arch/powerpc/kernel/crash_dump.c
@@ -69,13 +69,10 @@ void __init setup_kdump_trampoline(void)
#endif /* CONFIG_NONSTATIC_KERNEL */
static size_t copy_oldmem_vaddr(void *vaddr, char *buf, size_t csize,
- unsigned long offset, int userbuf)
+ unsigned long offset, bool userbuf)
{
- if (userbuf) {
- if (copy_to_user((char __user *)buf, (vaddr + offset), csize))
- return -EFAULT;
- } else
- memcpy(buf, (vaddr + offset), csize);
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ return -EFAULT;
return csize;
}
@@ -94,7 +91,7 @@ static size_t copy_oldmem_vaddr(void *vaddr, char *buf, size_t csize,
* in the current kernel. We stitch up a pte, similar to kmap_atomic.
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
- size_t csize, unsigned long offset, int userbuf)
+ size_t csize, unsigned long offset, bool userbuf)
{
void *vaddr;
phys_addr_t paddr;
diff --git a/arch/riscv/kernel/crash_dump.c b/arch/riscv/kernel/crash_dump.c
index 86cc0ad..4167437 100644
--- a/arch/riscv/kernel/crash_dump.c
+++ b/arch/riscv/kernel/crash_dump.c
@@ -22,7 +22,7 @@
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
size_t csize, unsigned long offset,
- int userbuf)
+ bool userbuf)
{
void *vaddr;
@@ -33,13 +33,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
if (!vaddr)
return -ENOMEM;
- if (userbuf) {
- if (copy_to_user((char __user *)buf, vaddr + offset, csize)) {
- memunmap(vaddr);
- return -EFAULT;
- }
- } else
- memcpy(buf, vaddr + offset, csize);
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
memunmap(vaddr);
return csize;
diff --git a/arch/sh/kernel/crash_dump.c b/arch/sh/kernel/crash_dump.c
index 5b41b59..4bc071a 100644
--- a/arch/sh/kernel/crash_dump.c
+++ b/arch/sh/kernel/crash_dump.c
@@ -24,7 +24,7 @@
* in the current kernel. We stitch up a pte, similar to kmap_atomic.
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
- size_t csize, unsigned long offset, int userbuf)
+ size_t csize, unsigned long offset, bool userbuf)
{
void __iomem *vaddr;
@@ -33,13 +33,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
vaddr = ioremap(pfn << PAGE_SHIFT, PAGE_SIZE);
- if (userbuf) {
- if (copy_to_user((void __user *)buf, (vaddr + offset), csize)) {
- iounmap(vaddr);
- return -EFAULT;
- }
- } else
- memcpy(buf, (vaddr + offset), csize);
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
iounmap(vaddr);
return csize;
diff --git a/arch/x86/kernel/crash_dump_32.c b/arch/x86/kernel/crash_dump_32.c
index 5fcac46..3eff124 100644
--- a/arch/x86/kernel/crash_dump_32.c
+++ b/arch/x86/kernel/crash_dump_32.c
@@ -43,7 +43,7 @@ static inline bool is_crashed_pfn_valid(unsigned long pfn)
* in the current kernel.
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
- unsigned long offset, int userbuf)
+ unsigned long offset, bool userbuf)
{
void *vaddr;
@@ -54,13 +54,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
return -EFAULT;
vaddr = kmap_local_pfn(pfn);
-
- if (!userbuf) {
- memcpy(buf, vaddr + offset, csize);
- } else {
- if (copy_to_user(buf, vaddr + offset, csize))
- csize = -EFAULT;
- }
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
kunmap_local(vaddr);
diff --git a/arch/x86/kernel/crash_dump_64.c b/arch/x86/kernel/crash_dump_64.c
index a7f617a..e8fffdf 100644
--- a/arch/x86/kernel/crash_dump_64.c
+++ b/arch/x86/kernel/crash_dump_64.c
@@ -13,7 +13,7 @@
#include <linux/cc_platform.h>
static ssize_t __copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
- unsigned long offset, int userbuf,
+ unsigned long offset, bool userbuf,
bool encrypted)
{
void *vaddr;
@@ -29,13 +29,8 @@ static ssize_t __copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
if (!vaddr)
return -ENOMEM;
- if (userbuf) {
- if (copy_to_user((void __user *)buf, vaddr + offset, csize)) {
- iounmap((void __iomem *)vaddr);
- return -EFAULT;
- }
- } else
- memcpy(buf, vaddr + offset, csize);
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
set_iounmap_nonlazy();
iounmap((void __iomem *)vaddr);
@@ -56,7 +51,7 @@ static ssize_t __copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
* mapped in the current kernel. We stitch up a pte, similar to kmap_atomic.
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
- unsigned long offset, int userbuf)
+ unsigned long offset, bool userbuf)
{
return __copy_oldmem_page(pfn, buf, csize, offset, userbuf, false);
}
@@ -67,7 +62,7 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
* machines.
*/
ssize_t copy_oldmem_page_encrypted(unsigned long pfn, char *buf, size_t csize,
- unsigned long offset, int userbuf)
+ unsigned long offset, bool userbuf)
{
return __copy_oldmem_page(pfn, buf, csize, offset, userbuf, true);
}
diff --git a/fs/proc/vmcore.c b/fs/proc/vmcore.c
index f67fd77..bba52aa 100644
--- a/fs/proc/vmcore.c
+++ b/fs/proc/vmcore.c
@@ -133,7 +133,7 @@ static int open_vmcore(struct inode *inode, struct file *file)
/* Reads a page from the oldmem device from given offset. */
ssize_t read_from_oldmem(char *buf, size_t count,
- u64 *ppos, int userbuf,
+ u64 *ppos, bool userbuf,
bool encrypted)
{
unsigned long pfn, offset;
@@ -233,7 +233,7 @@ int __weak remap_oldmem_pfn_range(struct vm_area_struct *vma,
*/
ssize_t __weak
copy_oldmem_page_encrypted(unsigned long pfn, char *buf, size_t csize,
- unsigned long offset, int userbuf)
+ unsigned long offset, bool userbuf)
{
return copy_oldmem_page(pfn, buf, csize, offset, userbuf);
}
diff --git a/include/linux/crash_dump.h b/include/linux/crash_dump.h
index 6208215..033448b 100644
--- a/include/linux/crash_dump.h
+++ b/include/linux/crash_dump.h
@@ -25,10 +25,10 @@ extern int remap_oldmem_pfn_range(struct vm_area_struct *vma,
unsigned long size, pgprot_t prot);
extern ssize_t copy_oldmem_page(unsigned long, char *, size_t,
- unsigned long, int);
+ unsigned long, bool);
extern ssize_t copy_oldmem_page_encrypted(unsigned long pfn, char *buf,
size_t csize, unsigned long offset,
- int userbuf);
+ bool userbuf);
void vmcore_cleanup(void);
@@ -136,11 +136,11 @@ static inline int vmcore_add_device_dump(struct vmcoredd_data *data)
#ifdef CONFIG_PROC_VMCORE
ssize_t read_from_oldmem(char *buf, size_t count,
- u64 *ppos, int userbuf,
+ u64 *ppos, bool userbuf,
bool encrypted);
#else
static inline ssize_t read_from_oldmem(char *buf, size_t count,
- u64 *ppos, int userbuf,
+ u64 *ppos, bool userbuf,
bool encrypted)
{
return -EOPNOTSUPP;
--
2.1.0
^ permalink raw reply related [flat|nested] 61+ messages in thread
* [PATCH v2 2/2] kdump: crashdump: use copy_to_user_or_kernel() to simplify code
@ 2021-12-11 3:33 ` Tiezhu Yang
0 siblings, 0 replies; 61+ messages in thread
From: Tiezhu Yang @ 2021-12-11 3:33 UTC (permalink / raw)
To: Dave Young, Baoquan He, Vivek Goyal, Andrew Morton
Cc: linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li
Use copy_to_user_or_kernel() to simplify the related code about
copy_oldmem_page() in arch/*/kernel/crash_dump*.c files.
Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
---
arch/arm/kernel/crash_dump.c | 12 +++---------
arch/arm64/kernel/crash_dump.c | 12 +++---------
arch/ia64/kernel/crash_dump.c | 12 +++++-------
arch/mips/kernel/crash_dump.c | 11 +++--------
arch/powerpc/kernel/crash_dump.c | 11 ++++-------
arch/riscv/kernel/crash_dump.c | 11 +++--------
arch/sh/kernel/crash_dump.c | 11 +++--------
arch/x86/kernel/crash_dump_32.c | 11 +++--------
arch/x86/kernel/crash_dump_64.c | 15 +++++----------
fs/proc/vmcore.c | 4 ++--
include/linux/crash_dump.h | 8 ++++----
11 files changed, 38 insertions(+), 80 deletions(-)
diff --git a/arch/arm/kernel/crash_dump.c b/arch/arm/kernel/crash_dump.c
index 53cb924..a27c5df 100644
--- a/arch/arm/kernel/crash_dump.c
+++ b/arch/arm/kernel/crash_dump.c
@@ -29,7 +29,7 @@
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
size_t csize, unsigned long offset,
- int userbuf)
+ bool userbuf)
{
void *vaddr;
@@ -40,14 +40,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
if (!vaddr)
return -ENOMEM;
- if (userbuf) {
- if (copy_to_user(buf, vaddr + offset, csize)) {
- iounmap(vaddr);
- return -EFAULT;
- }
- } else {
- memcpy(buf, vaddr + offset, csize);
- }
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
iounmap(vaddr);
return csize;
diff --git a/arch/arm64/kernel/crash_dump.c b/arch/arm64/kernel/crash_dump.c
index 58303a9..d22988f 100644
--- a/arch/arm64/kernel/crash_dump.c
+++ b/arch/arm64/kernel/crash_dump.c
@@ -27,7 +27,7 @@
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
size_t csize, unsigned long offset,
- int userbuf)
+ bool userbuf)
{
void *vaddr;
@@ -38,14 +38,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
if (!vaddr)
return -ENOMEM;
- if (userbuf) {
- if (copy_to_user((char __user *)buf, vaddr + offset, csize)) {
- memunmap(vaddr);
- return -EFAULT;
- }
- } else {
- memcpy(buf, vaddr + offset, csize);
- }
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
memunmap(vaddr);
diff --git a/arch/ia64/kernel/crash_dump.c b/arch/ia64/kernel/crash_dump.c
index 0ed3c3d..12128f8 100644
--- a/arch/ia64/kernel/crash_dump.c
+++ b/arch/ia64/kernel/crash_dump.c
@@ -33,19 +33,17 @@
*/
ssize_t
copy_oldmem_page(unsigned long pfn, char *buf,
- size_t csize, unsigned long offset, int userbuf)
+ size_t csize, unsigned long offset, bool userbuf)
{
void *vaddr;
if (!csize)
return 0;
+
vaddr = __va(pfn<<PAGE_SHIFT);
- if (userbuf) {
- if (copy_to_user(buf, (vaddr + offset), csize)) {
- return -EFAULT;
- }
- } else
- memcpy(buf, (vaddr + offset), csize);
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ return -EFAULT;
+
return csize;
}
diff --git a/arch/mips/kernel/crash_dump.c b/arch/mips/kernel/crash_dump.c
index 2e50f551..7670915 100644
--- a/arch/mips/kernel/crash_dump.c
+++ b/arch/mips/kernel/crash_dump.c
@@ -16,7 +16,7 @@
* in the current kernel.
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
- size_t csize, unsigned long offset, int userbuf)
+ size_t csize, unsigned long offset, bool userbuf)
{
void *vaddr;
@@ -24,13 +24,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
return 0;
vaddr = kmap_local_pfn(pfn);
-
- if (!userbuf) {
- memcpy(buf, vaddr + offset, csize);
- } else {
- if (copy_to_user(buf, vaddr + offset, csize))
- csize = -EFAULT;
- }
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
kunmap_local(vaddr);
diff --git a/arch/powerpc/kernel/crash_dump.c b/arch/powerpc/kernel/crash_dump.c
index 5693e1c67..e2e9612 100644
--- a/arch/powerpc/kernel/crash_dump.c
+++ b/arch/powerpc/kernel/crash_dump.c
@@ -69,13 +69,10 @@ void __init setup_kdump_trampoline(void)
#endif /* CONFIG_NONSTATIC_KERNEL */
static size_t copy_oldmem_vaddr(void *vaddr, char *buf, size_t csize,
- unsigned long offset, int userbuf)
+ unsigned long offset, bool userbuf)
{
- if (userbuf) {
- if (copy_to_user((char __user *)buf, (vaddr + offset), csize))
- return -EFAULT;
- } else
- memcpy(buf, (vaddr + offset), csize);
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ return -EFAULT;
return csize;
}
@@ -94,7 +91,7 @@ static size_t copy_oldmem_vaddr(void *vaddr, char *buf, size_t csize,
* in the current kernel. We stitch up a pte, similar to kmap_atomic.
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
- size_t csize, unsigned long offset, int userbuf)
+ size_t csize, unsigned long offset, bool userbuf)
{
void *vaddr;
phys_addr_t paddr;
diff --git a/arch/riscv/kernel/crash_dump.c b/arch/riscv/kernel/crash_dump.c
index 86cc0ad..4167437 100644
--- a/arch/riscv/kernel/crash_dump.c
+++ b/arch/riscv/kernel/crash_dump.c
@@ -22,7 +22,7 @@
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
size_t csize, unsigned long offset,
- int userbuf)
+ bool userbuf)
{
void *vaddr;
@@ -33,13 +33,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
if (!vaddr)
return -ENOMEM;
- if (userbuf) {
- if (copy_to_user((char __user *)buf, vaddr + offset, csize)) {
- memunmap(vaddr);
- return -EFAULT;
- }
- } else
- memcpy(buf, vaddr + offset, csize);
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
memunmap(vaddr);
return csize;
diff --git a/arch/sh/kernel/crash_dump.c b/arch/sh/kernel/crash_dump.c
index 5b41b59..4bc071a 100644
--- a/arch/sh/kernel/crash_dump.c
+++ b/arch/sh/kernel/crash_dump.c
@@ -24,7 +24,7 @@
* in the current kernel. We stitch up a pte, similar to kmap_atomic.
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
- size_t csize, unsigned long offset, int userbuf)
+ size_t csize, unsigned long offset, bool userbuf)
{
void __iomem *vaddr;
@@ -33,13 +33,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
vaddr = ioremap(pfn << PAGE_SHIFT, PAGE_SIZE);
- if (userbuf) {
- if (copy_to_user((void __user *)buf, (vaddr + offset), csize)) {
- iounmap(vaddr);
- return -EFAULT;
- }
- } else
- memcpy(buf, (vaddr + offset), csize);
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
iounmap(vaddr);
return csize;
diff --git a/arch/x86/kernel/crash_dump_32.c b/arch/x86/kernel/crash_dump_32.c
index 5fcac46..3eff124 100644
--- a/arch/x86/kernel/crash_dump_32.c
+++ b/arch/x86/kernel/crash_dump_32.c
@@ -43,7 +43,7 @@ static inline bool is_crashed_pfn_valid(unsigned long pfn)
* in the current kernel.
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
- unsigned long offset, int userbuf)
+ unsigned long offset, bool userbuf)
{
void *vaddr;
@@ -54,13 +54,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
return -EFAULT;
vaddr = kmap_local_pfn(pfn);
-
- if (!userbuf) {
- memcpy(buf, vaddr + offset, csize);
- } else {
- if (copy_to_user(buf, vaddr + offset, csize))
- csize = -EFAULT;
- }
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
kunmap_local(vaddr);
diff --git a/arch/x86/kernel/crash_dump_64.c b/arch/x86/kernel/crash_dump_64.c
index a7f617a..e8fffdf 100644
--- a/arch/x86/kernel/crash_dump_64.c
+++ b/arch/x86/kernel/crash_dump_64.c
@@ -13,7 +13,7 @@
#include <linux/cc_platform.h>
static ssize_t __copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
- unsigned long offset, int userbuf,
+ unsigned long offset, bool userbuf,
bool encrypted)
{
void *vaddr;
@@ -29,13 +29,8 @@ static ssize_t __copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
if (!vaddr)
return -ENOMEM;
- if (userbuf) {
- if (copy_to_user((void __user *)buf, vaddr + offset, csize)) {
- iounmap((void __iomem *)vaddr);
- return -EFAULT;
- }
- } else
- memcpy(buf, vaddr + offset, csize);
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
set_iounmap_nonlazy();
iounmap((void __iomem *)vaddr);
@@ -56,7 +51,7 @@ static ssize_t __copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
* mapped in the current kernel. We stitch up a pte, similar to kmap_atomic.
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
- unsigned long offset, int userbuf)
+ unsigned long offset, bool userbuf)
{
return __copy_oldmem_page(pfn, buf, csize, offset, userbuf, false);
}
@@ -67,7 +62,7 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
* machines.
*/
ssize_t copy_oldmem_page_encrypted(unsigned long pfn, char *buf, size_t csize,
- unsigned long offset, int userbuf)
+ unsigned long offset, bool userbuf)
{
return __copy_oldmem_page(pfn, buf, csize, offset, userbuf, true);
}
diff --git a/fs/proc/vmcore.c b/fs/proc/vmcore.c
index f67fd77..bba52aa 100644
--- a/fs/proc/vmcore.c
+++ b/fs/proc/vmcore.c
@@ -133,7 +133,7 @@ static int open_vmcore(struct inode *inode, struct file *file)
/* Reads a page from the oldmem device from given offset. */
ssize_t read_from_oldmem(char *buf, size_t count,
- u64 *ppos, int userbuf,
+ u64 *ppos, bool userbuf,
bool encrypted)
{
unsigned long pfn, offset;
@@ -233,7 +233,7 @@ int __weak remap_oldmem_pfn_range(struct vm_area_struct *vma,
*/
ssize_t __weak
copy_oldmem_page_encrypted(unsigned long pfn, char *buf, size_t csize,
- unsigned long offset, int userbuf)
+ unsigned long offset, bool userbuf)
{
return copy_oldmem_page(pfn, buf, csize, offset, userbuf);
}
diff --git a/include/linux/crash_dump.h b/include/linux/crash_dump.h
index 6208215..033448b 100644
--- a/include/linux/crash_dump.h
+++ b/include/linux/crash_dump.h
@@ -25,10 +25,10 @@ extern int remap_oldmem_pfn_range(struct vm_area_struct *vma,
unsigned long size, pgprot_t prot);
extern ssize_t copy_oldmem_page(unsigned long, char *, size_t,
- unsigned long, int);
+ unsigned long, bool);
extern ssize_t copy_oldmem_page_encrypted(unsigned long pfn, char *buf,
size_t csize, unsigned long offset,
- int userbuf);
+ bool userbuf);
void vmcore_cleanup(void);
@@ -136,11 +136,11 @@ static inline int vmcore_add_device_dump(struct vmcoredd_data *data)
#ifdef CONFIG_PROC_VMCORE
ssize_t read_from_oldmem(char *buf, size_t count,
- u64 *ppos, int userbuf,
+ u64 *ppos, bool userbuf,
bool encrypted);
#else
static inline ssize_t read_from_oldmem(char *buf, size_t count,
- u64 *ppos, int userbuf,
+ u64 *ppos, bool userbuf,
bool encrypted)
{
return -EOPNOTSUPP;
--
2.1.0
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply related [flat|nested] 61+ messages in thread
* [PATCH v2 2/2] kdump: crashdump: use copy_to_user_or_kernel() to simplify code
@ 2021-12-11 3:33 ` Tiezhu Yang
0 siblings, 0 replies; 61+ messages in thread
From: Tiezhu Yang @ 2021-12-11 3:33 UTC (permalink / raw)
To: Dave Young, Baoquan He, Vivek Goyal, Andrew Morton
Cc: linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li
Use copy_to_user_or_kernel() to simplify the related code about
copy_oldmem_page() in arch/*/kernel/crash_dump*.c files.
Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
---
arch/arm/kernel/crash_dump.c | 12 +++---------
arch/arm64/kernel/crash_dump.c | 12 +++---------
arch/ia64/kernel/crash_dump.c | 12 +++++-------
arch/mips/kernel/crash_dump.c | 11 +++--------
arch/powerpc/kernel/crash_dump.c | 11 ++++-------
arch/riscv/kernel/crash_dump.c | 11 +++--------
arch/sh/kernel/crash_dump.c | 11 +++--------
arch/x86/kernel/crash_dump_32.c | 11 +++--------
arch/x86/kernel/crash_dump_64.c | 15 +++++----------
fs/proc/vmcore.c | 4 ++--
include/linux/crash_dump.h | 8 ++++----
11 files changed, 38 insertions(+), 80 deletions(-)
diff --git a/arch/arm/kernel/crash_dump.c b/arch/arm/kernel/crash_dump.c
index 53cb924..a27c5df 100644
--- a/arch/arm/kernel/crash_dump.c
+++ b/arch/arm/kernel/crash_dump.c
@@ -29,7 +29,7 @@
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
size_t csize, unsigned long offset,
- int userbuf)
+ bool userbuf)
{
void *vaddr;
@@ -40,14 +40,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
if (!vaddr)
return -ENOMEM;
- if (userbuf) {
- if (copy_to_user(buf, vaddr + offset, csize)) {
- iounmap(vaddr);
- return -EFAULT;
- }
- } else {
- memcpy(buf, vaddr + offset, csize);
- }
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
iounmap(vaddr);
return csize;
diff --git a/arch/arm64/kernel/crash_dump.c b/arch/arm64/kernel/crash_dump.c
index 58303a9..d22988f 100644
--- a/arch/arm64/kernel/crash_dump.c
+++ b/arch/arm64/kernel/crash_dump.c
@@ -27,7 +27,7 @@
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
size_t csize, unsigned long offset,
- int userbuf)
+ bool userbuf)
{
void *vaddr;
@@ -38,14 +38,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
if (!vaddr)
return -ENOMEM;
- if (userbuf) {
- if (copy_to_user((char __user *)buf, vaddr + offset, csize)) {
- memunmap(vaddr);
- return -EFAULT;
- }
- } else {
- memcpy(buf, vaddr + offset, csize);
- }
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
memunmap(vaddr);
diff --git a/arch/ia64/kernel/crash_dump.c b/arch/ia64/kernel/crash_dump.c
index 0ed3c3d..12128f8 100644
--- a/arch/ia64/kernel/crash_dump.c
+++ b/arch/ia64/kernel/crash_dump.c
@@ -33,19 +33,17 @@
*/
ssize_t
copy_oldmem_page(unsigned long pfn, char *buf,
- size_t csize, unsigned long offset, int userbuf)
+ size_t csize, unsigned long offset, bool userbuf)
{
void *vaddr;
if (!csize)
return 0;
+
vaddr = __va(pfn<<PAGE_SHIFT);
- if (userbuf) {
- if (copy_to_user(buf, (vaddr + offset), csize)) {
- return -EFAULT;
- }
- } else
- memcpy(buf, (vaddr + offset), csize);
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ return -EFAULT;
+
return csize;
}
diff --git a/arch/mips/kernel/crash_dump.c b/arch/mips/kernel/crash_dump.c
index 2e50f551..7670915 100644
--- a/arch/mips/kernel/crash_dump.c
+++ b/arch/mips/kernel/crash_dump.c
@@ -16,7 +16,7 @@
* in the current kernel.
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
- size_t csize, unsigned long offset, int userbuf)
+ size_t csize, unsigned long offset, bool userbuf)
{
void *vaddr;
@@ -24,13 +24,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
return 0;
vaddr = kmap_local_pfn(pfn);
-
- if (!userbuf) {
- memcpy(buf, vaddr + offset, csize);
- } else {
- if (copy_to_user(buf, vaddr + offset, csize))
- csize = -EFAULT;
- }
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
kunmap_local(vaddr);
diff --git a/arch/powerpc/kernel/crash_dump.c b/arch/powerpc/kernel/crash_dump.c
index 5693e1c67..e2e9612 100644
--- a/arch/powerpc/kernel/crash_dump.c
+++ b/arch/powerpc/kernel/crash_dump.c
@@ -69,13 +69,10 @@ void __init setup_kdump_trampoline(void)
#endif /* CONFIG_NONSTATIC_KERNEL */
static size_t copy_oldmem_vaddr(void *vaddr, char *buf, size_t csize,
- unsigned long offset, int userbuf)
+ unsigned long offset, bool userbuf)
{
- if (userbuf) {
- if (copy_to_user((char __user *)buf, (vaddr + offset), csize))
- return -EFAULT;
- } else
- memcpy(buf, (vaddr + offset), csize);
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ return -EFAULT;
return csize;
}
@@ -94,7 +91,7 @@ static size_t copy_oldmem_vaddr(void *vaddr, char *buf, size_t csize,
* in the current kernel. We stitch up a pte, similar to kmap_atomic.
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
- size_t csize, unsigned long offset, int userbuf)
+ size_t csize, unsigned long offset, bool userbuf)
{
void *vaddr;
phys_addr_t paddr;
diff --git a/arch/riscv/kernel/crash_dump.c b/arch/riscv/kernel/crash_dump.c
index 86cc0ad..4167437 100644
--- a/arch/riscv/kernel/crash_dump.c
+++ b/arch/riscv/kernel/crash_dump.c
@@ -22,7 +22,7 @@
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
size_t csize, unsigned long offset,
- int userbuf)
+ bool userbuf)
{
void *vaddr;
@@ -33,13 +33,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
if (!vaddr)
return -ENOMEM;
- if (userbuf) {
- if (copy_to_user((char __user *)buf, vaddr + offset, csize)) {
- memunmap(vaddr);
- return -EFAULT;
- }
- } else
- memcpy(buf, vaddr + offset, csize);
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
memunmap(vaddr);
return csize;
diff --git a/arch/sh/kernel/crash_dump.c b/arch/sh/kernel/crash_dump.c
index 5b41b59..4bc071a 100644
--- a/arch/sh/kernel/crash_dump.c
+++ b/arch/sh/kernel/crash_dump.c
@@ -24,7 +24,7 @@
* in the current kernel. We stitch up a pte, similar to kmap_atomic.
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
- size_t csize, unsigned long offset, int userbuf)
+ size_t csize, unsigned long offset, bool userbuf)
{
void __iomem *vaddr;
@@ -33,13 +33,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
vaddr = ioremap(pfn << PAGE_SHIFT, PAGE_SIZE);
- if (userbuf) {
- if (copy_to_user((void __user *)buf, (vaddr + offset), csize)) {
- iounmap(vaddr);
- return -EFAULT;
- }
- } else
- memcpy(buf, (vaddr + offset), csize);
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
iounmap(vaddr);
return csize;
diff --git a/arch/x86/kernel/crash_dump_32.c b/arch/x86/kernel/crash_dump_32.c
index 5fcac46..3eff124 100644
--- a/arch/x86/kernel/crash_dump_32.c
+++ b/arch/x86/kernel/crash_dump_32.c
@@ -43,7 +43,7 @@ static inline bool is_crashed_pfn_valid(unsigned long pfn)
* in the current kernel.
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
- unsigned long offset, int userbuf)
+ unsigned long offset, bool userbuf)
{
void *vaddr;
@@ -54,13 +54,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
return -EFAULT;
vaddr = kmap_local_pfn(pfn);
-
- if (!userbuf) {
- memcpy(buf, vaddr + offset, csize);
- } else {
- if (copy_to_user(buf, vaddr + offset, csize))
- csize = -EFAULT;
- }
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
kunmap_local(vaddr);
diff --git a/arch/x86/kernel/crash_dump_64.c b/arch/x86/kernel/crash_dump_64.c
index a7f617a..e8fffdf 100644
--- a/arch/x86/kernel/crash_dump_64.c
+++ b/arch/x86/kernel/crash_dump_64.c
@@ -13,7 +13,7 @@
#include <linux/cc_platform.h>
static ssize_t __copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
- unsigned long offset, int userbuf,
+ unsigned long offset, bool userbuf,
bool encrypted)
{
void *vaddr;
@@ -29,13 +29,8 @@ static ssize_t __copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
if (!vaddr)
return -ENOMEM;
- if (userbuf) {
- if (copy_to_user((void __user *)buf, vaddr + offset, csize)) {
- iounmap((void __iomem *)vaddr);
- return -EFAULT;
- }
- } else
- memcpy(buf, vaddr + offset, csize);
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
set_iounmap_nonlazy();
iounmap((void __iomem *)vaddr);
@@ -56,7 +51,7 @@ static ssize_t __copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
* mapped in the current kernel. We stitch up a pte, similar to kmap_atomic.
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
- unsigned long offset, int userbuf)
+ unsigned long offset, bool userbuf)
{
return __copy_oldmem_page(pfn, buf, csize, offset, userbuf, false);
}
@@ -67,7 +62,7 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
* machines.
*/
ssize_t copy_oldmem_page_encrypted(unsigned long pfn, char *buf, size_t csize,
- unsigned long offset, int userbuf)
+ unsigned long offset, bool userbuf)
{
return __copy_oldmem_page(pfn, buf, csize, offset, userbuf, true);
}
diff --git a/fs/proc/vmcore.c b/fs/proc/vmcore.c
index f67fd77..bba52aa 100644
--- a/fs/proc/vmcore.c
+++ b/fs/proc/vmcore.c
@@ -133,7 +133,7 @@ static int open_vmcore(struct inode *inode, struct file *file)
/* Reads a page from the oldmem device from given offset. */
ssize_t read_from_oldmem(char *buf, size_t count,
- u64 *ppos, int userbuf,
+ u64 *ppos, bool userbuf,
bool encrypted)
{
unsigned long pfn, offset;
@@ -233,7 +233,7 @@ int __weak remap_oldmem_pfn_range(struct vm_area_struct *vma,
*/
ssize_t __weak
copy_oldmem_page_encrypted(unsigned long pfn, char *buf, size_t csize,
- unsigned long offset, int userbuf)
+ unsigned long offset, bool userbuf)
{
return copy_oldmem_page(pfn, buf, csize, offset, userbuf);
}
diff --git a/include/linux/crash_dump.h b/include/linux/crash_dump.h
index 6208215..033448b 100644
--- a/include/linux/crash_dump.h
+++ b/include/linux/crash_dump.h
@@ -25,10 +25,10 @@ extern int remap_oldmem_pfn_range(struct vm_area_struct *vma,
unsigned long size, pgprot_t prot);
extern ssize_t copy_oldmem_page(unsigned long, char *, size_t,
- unsigned long, int);
+ unsigned long, bool);
extern ssize_t copy_oldmem_page_encrypted(unsigned long pfn, char *buf,
size_t csize, unsigned long offset,
- int userbuf);
+ bool userbuf);
void vmcore_cleanup(void);
@@ -136,11 +136,11 @@ static inline int vmcore_add_device_dump(struct vmcoredd_data *data)
#ifdef CONFIG_PROC_VMCORE
ssize_t read_from_oldmem(char *buf, size_t count,
- u64 *ppos, int userbuf,
+ u64 *ppos, bool userbuf,
bool encrypted);
#else
static inline ssize_t read_from_oldmem(char *buf, size_t count,
- u64 *ppos, int userbuf,
+ u64 *ppos, bool userbuf,
bool encrypted)
{
return -EOPNOTSUPP;
--
2.1.0
_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec
^ permalink raw reply related [flat|nested] 61+ messages in thread
* [PATCH v2 2/2] kdump: crashdump: use copy_to_user_or_kernel() to simplify code
@ 2021-12-11 3:33 ` Tiezhu Yang
0 siblings, 0 replies; 61+ messages in thread
From: Tiezhu Yang @ 2021-12-11 3:33 UTC (permalink / raw)
To: Dave Young, Baoquan He, Vivek Goyal, Andrew Morton
Cc: linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li
Use copy_to_user_or_kernel() to simplify the related code about
copy_oldmem_page() in arch/*/kernel/crash_dump*.c files.
Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
---
arch/arm/kernel/crash_dump.c | 12 +++---------
arch/arm64/kernel/crash_dump.c | 12 +++---------
arch/ia64/kernel/crash_dump.c | 12 +++++-------
arch/mips/kernel/crash_dump.c | 11 +++--------
arch/powerpc/kernel/crash_dump.c | 11 ++++-------
arch/riscv/kernel/crash_dump.c | 11 +++--------
arch/sh/kernel/crash_dump.c | 11 +++--------
arch/x86/kernel/crash_dump_32.c | 11 +++--------
arch/x86/kernel/crash_dump_64.c | 15 +++++----------
fs/proc/vmcore.c | 4 ++--
include/linux/crash_dump.h | 8 ++++----
11 files changed, 38 insertions(+), 80 deletions(-)
diff --git a/arch/arm/kernel/crash_dump.c b/arch/arm/kernel/crash_dump.c
index 53cb924..a27c5df 100644
--- a/arch/arm/kernel/crash_dump.c
+++ b/arch/arm/kernel/crash_dump.c
@@ -29,7 +29,7 @@
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
size_t csize, unsigned long offset,
- int userbuf)
+ bool userbuf)
{
void *vaddr;
@@ -40,14 +40,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
if (!vaddr)
return -ENOMEM;
- if (userbuf) {
- if (copy_to_user(buf, vaddr + offset, csize)) {
- iounmap(vaddr);
- return -EFAULT;
- }
- } else {
- memcpy(buf, vaddr + offset, csize);
- }
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
iounmap(vaddr);
return csize;
diff --git a/arch/arm64/kernel/crash_dump.c b/arch/arm64/kernel/crash_dump.c
index 58303a9..d22988f 100644
--- a/arch/arm64/kernel/crash_dump.c
+++ b/arch/arm64/kernel/crash_dump.c
@@ -27,7 +27,7 @@
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
size_t csize, unsigned long offset,
- int userbuf)
+ bool userbuf)
{
void *vaddr;
@@ -38,14 +38,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
if (!vaddr)
return -ENOMEM;
- if (userbuf) {
- if (copy_to_user((char __user *)buf, vaddr + offset, csize)) {
- memunmap(vaddr);
- return -EFAULT;
- }
- } else {
- memcpy(buf, vaddr + offset, csize);
- }
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
memunmap(vaddr);
diff --git a/arch/ia64/kernel/crash_dump.c b/arch/ia64/kernel/crash_dump.c
index 0ed3c3d..12128f8 100644
--- a/arch/ia64/kernel/crash_dump.c
+++ b/arch/ia64/kernel/crash_dump.c
@@ -33,19 +33,17 @@
*/
ssize_t
copy_oldmem_page(unsigned long pfn, char *buf,
- size_t csize, unsigned long offset, int userbuf)
+ size_t csize, unsigned long offset, bool userbuf)
{
void *vaddr;
if (!csize)
return 0;
+
vaddr = __va(pfn<<PAGE_SHIFT);
- if (userbuf) {
- if (copy_to_user(buf, (vaddr + offset), csize)) {
- return -EFAULT;
- }
- } else
- memcpy(buf, (vaddr + offset), csize);
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ return -EFAULT;
+
return csize;
}
diff --git a/arch/mips/kernel/crash_dump.c b/arch/mips/kernel/crash_dump.c
index 2e50f551..7670915 100644
--- a/arch/mips/kernel/crash_dump.c
+++ b/arch/mips/kernel/crash_dump.c
@@ -16,7 +16,7 @@
* in the current kernel.
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
- size_t csize, unsigned long offset, int userbuf)
+ size_t csize, unsigned long offset, bool userbuf)
{
void *vaddr;
@@ -24,13 +24,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
return 0;
vaddr = kmap_local_pfn(pfn);
-
- if (!userbuf) {
- memcpy(buf, vaddr + offset, csize);
- } else {
- if (copy_to_user(buf, vaddr + offset, csize))
- csize = -EFAULT;
- }
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
kunmap_local(vaddr);
diff --git a/arch/powerpc/kernel/crash_dump.c b/arch/powerpc/kernel/crash_dump.c
index 5693e1c67..e2e9612 100644
--- a/arch/powerpc/kernel/crash_dump.c
+++ b/arch/powerpc/kernel/crash_dump.c
@@ -69,13 +69,10 @@ void __init setup_kdump_trampoline(void)
#endif /* CONFIG_NONSTATIC_KERNEL */
static size_t copy_oldmem_vaddr(void *vaddr, char *buf, size_t csize,
- unsigned long offset, int userbuf)
+ unsigned long offset, bool userbuf)
{
- if (userbuf) {
- if (copy_to_user((char __user *)buf, (vaddr + offset), csize))
- return -EFAULT;
- } else
- memcpy(buf, (vaddr + offset), csize);
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ return -EFAULT;
return csize;
}
@@ -94,7 +91,7 @@ static size_t copy_oldmem_vaddr(void *vaddr, char *buf, size_t csize,
* in the current kernel. We stitch up a pte, similar to kmap_atomic.
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
- size_t csize, unsigned long offset, int userbuf)
+ size_t csize, unsigned long offset, bool userbuf)
{
void *vaddr;
phys_addr_t paddr;
diff --git a/arch/riscv/kernel/crash_dump.c b/arch/riscv/kernel/crash_dump.c
index 86cc0ad..4167437 100644
--- a/arch/riscv/kernel/crash_dump.c
+++ b/arch/riscv/kernel/crash_dump.c
@@ -22,7 +22,7 @@
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
size_t csize, unsigned long offset,
- int userbuf)
+ bool userbuf)
{
void *vaddr;
@@ -33,13 +33,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
if (!vaddr)
return -ENOMEM;
- if (userbuf) {
- if (copy_to_user((char __user *)buf, vaddr + offset, csize)) {
- memunmap(vaddr);
- return -EFAULT;
- }
- } else
- memcpy(buf, vaddr + offset, csize);
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
memunmap(vaddr);
return csize;
diff --git a/arch/sh/kernel/crash_dump.c b/arch/sh/kernel/crash_dump.c
index 5b41b59..4bc071a 100644
--- a/arch/sh/kernel/crash_dump.c
+++ b/arch/sh/kernel/crash_dump.c
@@ -24,7 +24,7 @@
* in the current kernel. We stitch up a pte, similar to kmap_atomic.
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
- size_t csize, unsigned long offset, int userbuf)
+ size_t csize, unsigned long offset, bool userbuf)
{
void __iomem *vaddr;
@@ -33,13 +33,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
vaddr = ioremap(pfn << PAGE_SHIFT, PAGE_SIZE);
- if (userbuf) {
- if (copy_to_user((void __user *)buf, (vaddr + offset), csize)) {
- iounmap(vaddr);
- return -EFAULT;
- }
- } else
- memcpy(buf, (vaddr + offset), csize);
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
iounmap(vaddr);
return csize;
diff --git a/arch/x86/kernel/crash_dump_32.c b/arch/x86/kernel/crash_dump_32.c
index 5fcac46..3eff124 100644
--- a/arch/x86/kernel/crash_dump_32.c
+++ b/arch/x86/kernel/crash_dump_32.c
@@ -43,7 +43,7 @@ static inline bool is_crashed_pfn_valid(unsigned long pfn)
* in the current kernel.
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
- unsigned long offset, int userbuf)
+ unsigned long offset, bool userbuf)
{
void *vaddr;
@@ -54,13 +54,8 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
return -EFAULT;
vaddr = kmap_local_pfn(pfn);
-
- if (!userbuf) {
- memcpy(buf, vaddr + offset, csize);
- } else {
- if (copy_to_user(buf, vaddr + offset, csize))
- csize = -EFAULT;
- }
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
kunmap_local(vaddr);
diff --git a/arch/x86/kernel/crash_dump_64.c b/arch/x86/kernel/crash_dump_64.c
index a7f617a..e8fffdf 100644
--- a/arch/x86/kernel/crash_dump_64.c
+++ b/arch/x86/kernel/crash_dump_64.c
@@ -13,7 +13,7 @@
#include <linux/cc_platform.h>
static ssize_t __copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
- unsigned long offset, int userbuf,
+ unsigned long offset, bool userbuf,
bool encrypted)
{
void *vaddr;
@@ -29,13 +29,8 @@ static ssize_t __copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
if (!vaddr)
return -ENOMEM;
- if (userbuf) {
- if (copy_to_user((void __user *)buf, vaddr + offset, csize)) {
- iounmap((void __iomem *)vaddr);
- return -EFAULT;
- }
- } else
- memcpy(buf, vaddr + offset, csize);
+ if (copy_to_user_or_kernel(buf, vaddr + offset, csize, userbuf))
+ csize = -EFAULT;
set_iounmap_nonlazy();
iounmap((void __iomem *)vaddr);
@@ -56,7 +51,7 @@ static ssize_t __copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
* mapped in the current kernel. We stitch up a pte, similar to kmap_atomic.
*/
ssize_t copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
- unsigned long offset, int userbuf)
+ unsigned long offset, bool userbuf)
{
return __copy_oldmem_page(pfn, buf, csize, offset, userbuf, false);
}
@@ -67,7 +62,7 @@ ssize_t copy_oldmem_page(unsigned long pfn, char *buf, size_t csize,
* machines.
*/
ssize_t copy_oldmem_page_encrypted(unsigned long pfn, char *buf, size_t csize,
- unsigned long offset, int userbuf)
+ unsigned long offset, bool userbuf)
{
return __copy_oldmem_page(pfn, buf, csize, offset, userbuf, true);
}
diff --git a/fs/proc/vmcore.c b/fs/proc/vmcore.c
index f67fd77..bba52aa 100644
--- a/fs/proc/vmcore.c
+++ b/fs/proc/vmcore.c
@@ -133,7 +133,7 @@ static int open_vmcore(struct inode *inode, struct file *file)
/* Reads a page from the oldmem device from given offset. */
ssize_t read_from_oldmem(char *buf, size_t count,
- u64 *ppos, int userbuf,
+ u64 *ppos, bool userbuf,
bool encrypted)
{
unsigned long pfn, offset;
@@ -233,7 +233,7 @@ int __weak remap_oldmem_pfn_range(struct vm_area_struct *vma,
*/
ssize_t __weak
copy_oldmem_page_encrypted(unsigned long pfn, char *buf, size_t csize,
- unsigned long offset, int userbuf)
+ unsigned long offset, bool userbuf)
{
return copy_oldmem_page(pfn, buf, csize, offset, userbuf);
}
diff --git a/include/linux/crash_dump.h b/include/linux/crash_dump.h
index 6208215..033448b 100644
--- a/include/linux/crash_dump.h
+++ b/include/linux/crash_dump.h
@@ -25,10 +25,10 @@ extern int remap_oldmem_pfn_range(struct vm_area_struct *vma,
unsigned long size, pgprot_t prot);
extern ssize_t copy_oldmem_page(unsigned long, char *, size_t,
- unsigned long, int);
+ unsigned long, bool);
extern ssize_t copy_oldmem_page_encrypted(unsigned long pfn, char *buf,
size_t csize, unsigned long offset,
- int userbuf);
+ bool userbuf);
void vmcore_cleanup(void);
@@ -136,11 +136,11 @@ static inline int vmcore_add_device_dump(struct vmcoredd_data *data)
#ifdef CONFIG_PROC_VMCORE
ssize_t read_from_oldmem(char *buf, size_t count,
- u64 *ppos, int userbuf,
+ u64 *ppos, bool userbuf,
bool encrypted);
#else
static inline ssize_t read_from_oldmem(char *buf, size_t count,
- u64 *ppos, int userbuf,
+ u64 *ppos, bool userbuf,
bool encrypted)
{
return -EOPNOTSUPP;
--
2.1.0
^ permalink raw reply related [flat|nested] 61+ messages in thread
* Re: [PATCH v2 1/2] kdump: vmcore: remove copy_to() and add copy_to_user_or_kernel()
2021-12-11 3:33 ` Tiezhu Yang
` (3 preceding siblings ...)
(?)
@ 2021-12-11 10:32 ` Christophe Leroy
-1 siblings, 0 replies; 61+ messages in thread
From: Christophe Leroy @ 2021-12-11 10:32 UTC (permalink / raw)
To: Tiezhu Yang, Dave Young, Baoquan He, Vivek Goyal, Andrew Morton
Cc: linux-ia64, linux-sh, Xuefeng Li, x86, kexec, linux-mips,
linux-kernel, linux-fsdevel, linux-riscv, linuxppc-dev,
linux-arm-kernel
Le 11/12/2021 à 04:33, Tiezhu Yang a écrit :
> In arch/*/kernel/crash_dump*.c, there exist many similar code
> about copy_oldmem_page(), remove copy_to() in fs/proc/vmcore.c
> and add copy_to_user_or_kernel() in lib/usercopy.c, then we can
> use copy_to_user_or_kernel() to simplify the related code.
It should be an inline function in uaccess.h, see below why.
>
> Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
> ---
> fs/proc/vmcore.c | 28 +++++++---------------------
> include/linux/uaccess.h | 1 +
> lib/usercopy.c | 15 +++++++++++++++
> 3 files changed, 23 insertions(+), 21 deletions(-)
>
> diff --git a/fs/proc/vmcore.c b/fs/proc/vmcore.c
> index 509f851..f67fd77 100644
> --- a/fs/proc/vmcore.c
> +++ b/fs/proc/vmcore.c
> @@ -238,22 +238,8 @@ copy_oldmem_page_encrypted(unsigned long pfn, char *buf, size_t csize,
> return copy_oldmem_page(pfn, buf, csize, offset, userbuf);
> }
>
> -/*
> - * Copy to either kernel or user space
> - */
> -static int copy_to(void *target, void *src, size_t size, int userbuf)
> -{
> - if (userbuf) {
> - if (copy_to_user((char __user *) target, src, size))
> - return -EFAULT;
> - } else {
> - memcpy(target, src, size);
> - }
> - return 0;
> -}
> -
> #ifdef CONFIG_PROC_VMCORE_DEVICE_DUMP
> -static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, int userbuf)
> +static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, bool userbuf)
Changing int to bool in all the callers should be another patch. You can
have copy_to_user_or_kernel() take a bool in the patch while still
having all the callers using an int.
> {
> struct vmcoredd_node *dump;
> u64 offset = 0;
> @@ -266,7 +252,7 @@ static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, int userbuf)
> if (start < offset + dump->size) {
> tsz = min(offset + (u64)dump->size - start, (u64)size);
> buf = dump->buf + start - offset;
> - if (copy_to(dst, buf, tsz, userbuf)) {
> + if (copy_to_user_or_kernel(dst, buf, tsz, userbuf)) {
> ret = -EFAULT;
> goto out_unlock;
> }
> @@ -330,7 +316,7 @@ static int vmcoredd_mmap_dumps(struct vm_area_struct *vma, unsigned long dst,
> * returned otherwise number of bytes read are returned.
> */
> static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
> - int userbuf)
> + bool userbuf)
> {
> ssize_t acc = 0, tmp;
> size_t tsz;
> @@ -347,7 +333,7 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
> /* Read ELF core header */
> if (*fpos < elfcorebuf_sz) {
> tsz = min(elfcorebuf_sz - (size_t)*fpos, buflen);
> - if (copy_to(buffer, elfcorebuf + *fpos, tsz, userbuf))
> + if (copy_to_user_or_kernel(buffer, elfcorebuf + *fpos, tsz, userbuf))
> return -EFAULT;
> buflen -= tsz;
> *fpos += tsz;
> @@ -395,7 +381,7 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
> /* Read remaining elf notes */
> tsz = min(elfcorebuf_sz + elfnotes_sz - (size_t)*fpos, buflen);
> kaddr = elfnotes_buf + *fpos - elfcorebuf_sz - vmcoredd_orig_sz;
> - if (copy_to(buffer, kaddr, tsz, userbuf))
> + if (copy_to_user_or_kernel(buffer, kaddr, tsz, userbuf))
> return -EFAULT;
>
> buflen -= tsz;
> @@ -435,7 +421,7 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
> static ssize_t read_vmcore(struct file *file, char __user *buffer,
> size_t buflen, loff_t *fpos)
> {
> - return __read_vmcore((__force char *) buffer, buflen, fpos, 1);
> + return __read_vmcore((__force char *) buffer, buflen, fpos, true);
> }
>
> /*
> @@ -461,7 +447,7 @@ static vm_fault_t mmap_vmcore_fault(struct vm_fault *vmf)
> if (!PageUptodate(page)) {
> offset = (loff_t) index << PAGE_SHIFT;
> buf = __va((page_to_pfn(page) << PAGE_SHIFT));
> - rc = __read_vmcore(buf, PAGE_SIZE, &offset, 0);
> + rc = __read_vmcore(buf, PAGE_SIZE, &offset, false);
> if (rc < 0) {
> unlock_page(page);
> put_page(page);
> diff --git a/include/linux/uaccess.h b/include/linux/uaccess.h
> index ac03940..a25e682e 100644
> --- a/include/linux/uaccess.h
> +++ b/include/linux/uaccess.h
> @@ -283,6 +283,7 @@ __copy_from_user_inatomic_nocache(void *to, const void __user *from,
> #endif /* ARCH_HAS_NOCACHE_UACCESS */
>
> extern __must_check int check_zeroed_user(const void __user *from, size_t size);
> +extern __must_check int copy_to_user_or_kernel(void *target, void *src, size_t size, bool userbuf);
extern keyword is pointless for function prototypes, please don't add
new ones.
>
> /**
> * copy_struct_from_user: copy a struct from userspace
> diff --git a/lib/usercopy.c b/lib/usercopy.c
> index 7413dd3..7431b1b 100644
> --- a/lib/usercopy.c
> +++ b/lib/usercopy.c
> @@ -90,3 +90,18 @@ int check_zeroed_user(const void __user *from, size_t size)
> return -EFAULT;
> }
> EXPORT_SYMBOL(check_zeroed_user);
> +
> +/*
> + * Copy to either user or kernel space
> + */
> +int copy_to_user_or_kernel(void *target, void *src, size_t size, bool userbuf)
> +{
> + if (userbuf) {
> + if (copy_to_user((char __user *) target, src, size))
> + return -EFAULT;
> + } else {
> + memcpy(target, src, size);
> + }
> + return 0;
> +}
> +EXPORT_SYMBOL(copy_to_user_or_kernel);
>
Ref my answer to Andrew, I don't think outlining this fonction is a
worth it. As shown in that mail, the size of the caller is increased by
4 instructions (which is in the noise) but also this new function is not
small. So I see no benefit in term of size, and I don't think there is
any benefit in terms of performance either.
In this patch that's the same. Before the patch, read_vmcore() has a
size of 0x338.
With this patch, read_vmcore() has a size of 0x340. So that's 2
instructions more, so no benefit either.
So I think this should remain an inline function like in your first
patch (but with the new name).
000001a4 <copy_to_user_or_kernel>:
1a4: 2c 06 00 00 cmpwi r6,0
1a8: 94 21 ff f0 stwu r1,-16(r1)
1ac: 41 82 00 50 beq 1fc <copy_to_user_or_kernel+0x58>
1b0: 2c 05 00 00 cmpwi r5,0
1b4: 41 80 00 7c blt 230 <copy_to_user_or_kernel+0x8c>
1b8: 3d 00 b0 00 lis r8,-20480
1bc: 7f 83 40 40 cmplw cr7,r3,r8
1c0: 41 9c 00 14 blt cr7,1d4 <copy_to_user_or_kernel+0x30>
1c4: 40 82 00 64 bne 228 <copy_to_user_or_kernel+0x84>
1c8: 38 60 00 00 li r3,0
1cc: 38 21 00 10 addi r1,r1,16
1d0: 4e 80 00 20 blr
1d4: 7d 23 40 50 subf r9,r3,r8
1d8: 7f 85 48 40 cmplw cr7,r5,r9
1dc: 7c 08 02 a6 mflr r0
1e0: 90 01 00 14 stw r0,20(r1)
1e4: 41 9d 00 38 bgt cr7,21c <copy_to_user_or_kernel+0x78>
1e8: 48 00 00 01 bl 1e8 <copy_to_user_or_kernel+0x44>
1e8: R_PPC_REL24 __copy_tofrom_user
1ec: 80 01 00 14 lwz r0,20(r1)
1f0: 2c 03 00 00 cmpwi r3,0
1f4: 7c 08 03 a6 mtlr r0
1f8: 4b ff ff cc b 1c4 <copy_to_user_or_kernel+0x20>
1fc: 7c 08 02 a6 mflr r0
200: 90 01 00 14 stw r0,20(r1)
204: 48 00 00 01 bl 204 <copy_to_user_or_kernel+0x60>
204: R_PPC_REL24 memcpy
208: 38 60 00 00 li r3,0
20c: 80 01 00 14 lwz r0,20(r1)
210: 38 21 00 10 addi r1,r1,16
214: 7c 08 03 a6 mtlr r0
218: 4e 80 00 20 blr
21c: 80 01 00 14 lwz r0,20(r1)
220: 7c 08 03 a6 mtlr r0
224: 4b ff ff a0 b 1c4 <copy_to_user_or_kernel+0x20>
228: 38 60 ff f2 li r3,-14
22c: 4b ff ff a0 b 1cc <copy_to_user_or_kernel+0x28>
230: 0f e0 00 00 twui r0,0
234: 7c 08 02 a6 mflr r0
238: 90 01 00 14 stw r0,20(r1)
Also note that checkpatch.pl provides the following on your patch:
CHECK: No space is necessary after a cast
#88: FILE: fs/proc/vmcore.c:424:
+ return __read_vmcore((__force char *) buffer, buflen, fpos, true);
CHECK: extern prototypes should be avoided in .h files
#109: FILE: include/linux/uaccess.h:286:
+extern __must_check int copy_to_user_or_kernel(void *target, void *src,
size_t size, bool userbuf);
CHECK: No space is necessary after a cast
#128: FILE: lib/usercopy.c:100:
+ if (copy_to_user((char __user *) target, src, size))
total: 0 errors, 0 warnings, 3 checks, 96 lines checked
NOTE: For some of the reported defects, checkpatch may be able to
mechanically convert to the typical style using --fix or
--fix-inplace.
Commit 2c94767fa768 ("kdump: vmcore: remove copy_to() and add
copy_to_user_or_kernel()") has style problems, please review.
NOTE: If any of the errors are false positives, please report
them to the maintainer, see CHECKPATCH in MAINTAINERS.
Christophe
^ permalink raw reply [flat|nested] 61+ messages in thread
* Re: [PATCH v2 1/2] kdump: vmcore: remove copy_to() and add copy_to_user_or_kernel()
@ 2021-12-11 10:32 ` Christophe Leroy
0 siblings, 0 replies; 61+ messages in thread
From: Christophe Leroy @ 2021-12-11 10:32 UTC (permalink / raw)
To: Tiezhu Yang, Dave Young, Baoquan He, Vivek Goyal, Andrew Morton
Cc: linux-ia64, linux-sh, Xuefeng Li, x86, kexec, linux-mips,
linux-kernel, linux-fsdevel, linux-riscv, linuxppc-dev,
linux-arm-kernel
Le 11/12/2021 à 04:33, Tiezhu Yang a écrit :
> In arch/*/kernel/crash_dump*.c, there exist many similar code
> about copy_oldmem_page(), remove copy_to() in fs/proc/vmcore.c
> and add copy_to_user_or_kernel() in lib/usercopy.c, then we can
> use copy_to_user_or_kernel() to simplify the related code.
It should be an inline function in uaccess.h, see below why.
>
> Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
> ---
> fs/proc/vmcore.c | 28 +++++++---------------------
> include/linux/uaccess.h | 1 +
> lib/usercopy.c | 15 +++++++++++++++
> 3 files changed, 23 insertions(+), 21 deletions(-)
>
> diff --git a/fs/proc/vmcore.c b/fs/proc/vmcore.c
> index 509f851..f67fd77 100644
> --- a/fs/proc/vmcore.c
> +++ b/fs/proc/vmcore.c
> @@ -238,22 +238,8 @@ copy_oldmem_page_encrypted(unsigned long pfn, char *buf, size_t csize,
> return copy_oldmem_page(pfn, buf, csize, offset, userbuf);
> }
>
> -/*
> - * Copy to either kernel or user space
> - */
> -static int copy_to(void *target, void *src, size_t size, int userbuf)
> -{
> - if (userbuf) {
> - if (copy_to_user((char __user *) target, src, size))
> - return -EFAULT;
> - } else {
> - memcpy(target, src, size);
> - }
> - return 0;
> -}
> -
> #ifdef CONFIG_PROC_VMCORE_DEVICE_DUMP
> -static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, int userbuf)
> +static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, bool userbuf)
Changing int to bool in all the callers should be another patch. You can
have copy_to_user_or_kernel() take a bool in the patch while still
having all the callers using an int.
> {
> struct vmcoredd_node *dump;
> u64 offset = 0;
> @@ -266,7 +252,7 @@ static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, int userbuf)
> if (start < offset + dump->size) {
> tsz = min(offset + (u64)dump->size - start, (u64)size);
> buf = dump->buf + start - offset;
> - if (copy_to(dst, buf, tsz, userbuf)) {
> + if (copy_to_user_or_kernel(dst, buf, tsz, userbuf)) {
> ret = -EFAULT;
> goto out_unlock;
> }
> @@ -330,7 +316,7 @@ static int vmcoredd_mmap_dumps(struct vm_area_struct *vma, unsigned long dst,
> * returned otherwise number of bytes read are returned.
> */
> static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
> - int userbuf)
> + bool userbuf)
> {
> ssize_t acc = 0, tmp;
> size_t tsz;
> @@ -347,7 +333,7 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
> /* Read ELF core header */
> if (*fpos < elfcorebuf_sz) {
> tsz = min(elfcorebuf_sz - (size_t)*fpos, buflen);
> - if (copy_to(buffer, elfcorebuf + *fpos, tsz, userbuf))
> + if (copy_to_user_or_kernel(buffer, elfcorebuf + *fpos, tsz, userbuf))
> return -EFAULT;
> buflen -= tsz;
> *fpos += tsz;
> @@ -395,7 +381,7 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
> /* Read remaining elf notes */
> tsz = min(elfcorebuf_sz + elfnotes_sz - (size_t)*fpos, buflen);
> kaddr = elfnotes_buf + *fpos - elfcorebuf_sz - vmcoredd_orig_sz;
> - if (copy_to(buffer, kaddr, tsz, userbuf))
> + if (copy_to_user_or_kernel(buffer, kaddr, tsz, userbuf))
> return -EFAULT;
>
> buflen -= tsz;
> @@ -435,7 +421,7 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
> static ssize_t read_vmcore(struct file *file, char __user *buffer,
> size_t buflen, loff_t *fpos)
> {
> - return __read_vmcore((__force char *) buffer, buflen, fpos, 1);
> + return __read_vmcore((__force char *) buffer, buflen, fpos, true);
> }
>
> /*
> @@ -461,7 +447,7 @@ static vm_fault_t mmap_vmcore_fault(struct vm_fault *vmf)
> if (!PageUptodate(page)) {
> offset = (loff_t) index << PAGE_SHIFT;
> buf = __va((page_to_pfn(page) << PAGE_SHIFT));
> - rc = __read_vmcore(buf, PAGE_SIZE, &offset, 0);
> + rc = __read_vmcore(buf, PAGE_SIZE, &offset, false);
> if (rc < 0) {
> unlock_page(page);
> put_page(page);
> diff --git a/include/linux/uaccess.h b/include/linux/uaccess.h
> index ac03940..a25e682e 100644
> --- a/include/linux/uaccess.h
> +++ b/include/linux/uaccess.h
> @@ -283,6 +283,7 @@ __copy_from_user_inatomic_nocache(void *to, const void __user *from,
> #endif /* ARCH_HAS_NOCACHE_UACCESS */
>
> extern __must_check int check_zeroed_user(const void __user *from, size_t size);
> +extern __must_check int copy_to_user_or_kernel(void *target, void *src, size_t size, bool userbuf);
extern keyword is pointless for function prototypes, please don't add
new ones.
>
> /**
> * copy_struct_from_user: copy a struct from userspace
> diff --git a/lib/usercopy.c b/lib/usercopy.c
> index 7413dd3..7431b1b 100644
> --- a/lib/usercopy.c
> +++ b/lib/usercopy.c
> @@ -90,3 +90,18 @@ int check_zeroed_user(const void __user *from, size_t size)
> return -EFAULT;
> }
> EXPORT_SYMBOL(check_zeroed_user);
> +
> +/*
> + * Copy to either user or kernel space
> + */
> +int copy_to_user_or_kernel(void *target, void *src, size_t size, bool userbuf)
> +{
> + if (userbuf) {
> + if (copy_to_user((char __user *) target, src, size))
> + return -EFAULT;
> + } else {
> + memcpy(target, src, size);
> + }
> + return 0;
> +}
> +EXPORT_SYMBOL(copy_to_user_or_kernel);
>
Ref my answer to Andrew, I don't think outlining this fonction is a
worth it. As shown in that mail, the size of the caller is increased by
4 instructions (which is in the noise) but also this new function is not
small. So I see no benefit in term of size, and I don't think there is
any benefit in terms of performance either.
In this patch that's the same. Before the patch, read_vmcore() has a
size of 0x338.
With this patch, read_vmcore() has a size of 0x340. So that's 2
instructions more, so no benefit either.
So I think this should remain an inline function like in your first
patch (but with the new name).
000001a4 <copy_to_user_or_kernel>:
1a4: 2c 06 00 00 cmpwi r6,0
1a8: 94 21 ff f0 stwu r1,-16(r1)
1ac: 41 82 00 50 beq 1fc <copy_to_user_or_kernel+0x58>
1b0: 2c 05 00 00 cmpwi r5,0
1b4: 41 80 00 7c blt 230 <copy_to_user_or_kernel+0x8c>
1b8: 3d 00 b0 00 lis r8,-20480
1bc: 7f 83 40 40 cmplw cr7,r3,r8
1c0: 41 9c 00 14 blt cr7,1d4 <copy_to_user_or_kernel+0x30>
1c4: 40 82 00 64 bne 228 <copy_to_user_or_kernel+0x84>
1c8: 38 60 00 00 li r3,0
1cc: 38 21 00 10 addi r1,r1,16
1d0: 4e 80 00 20 blr
1d4: 7d 23 40 50 subf r9,r3,r8
1d8: 7f 85 48 40 cmplw cr7,r5,r9
1dc: 7c 08 02 a6 mflr r0
1e0: 90 01 00 14 stw r0,20(r1)
1e4: 41 9d 00 38 bgt cr7,21c <copy_to_user_or_kernel+0x78>
1e8: 48 00 00 01 bl 1e8 <copy_to_user_or_kernel+0x44>
1e8: R_PPC_REL24 __copy_tofrom_user
1ec: 80 01 00 14 lwz r0,20(r1)
1f0: 2c 03 00 00 cmpwi r3,0
1f4: 7c 08 03 a6 mtlr r0
1f8: 4b ff ff cc b 1c4 <copy_to_user_or_kernel+0x20>
1fc: 7c 08 02 a6 mflr r0
200: 90 01 00 14 stw r0,20(r1)
204: 48 00 00 01 bl 204 <copy_to_user_or_kernel+0x60>
204: R_PPC_REL24 memcpy
208: 38 60 00 00 li r3,0
20c: 80 01 00 14 lwz r0,20(r1)
210: 38 21 00 10 addi r1,r1,16
214: 7c 08 03 a6 mtlr r0
218: 4e 80 00 20 blr
21c: 80 01 00 14 lwz r0,20(r1)
220: 7c 08 03 a6 mtlr r0
224: 4b ff ff a0 b 1c4 <copy_to_user_or_kernel+0x20>
228: 38 60 ff f2 li r3,-14
22c: 4b ff ff a0 b 1cc <copy_to_user_or_kernel+0x28>
230: 0f e0 00 00 twui r0,0
234: 7c 08 02 a6 mflr r0
238: 90 01 00 14 stw r0,20(r1)
Also note that checkpatch.pl provides the following on your patch:
CHECK: No space is necessary after a cast
#88: FILE: fs/proc/vmcore.c:424:
+ return __read_vmcore((__force char *) buffer, buflen, fpos, true);
CHECK: extern prototypes should be avoided in .h files
#109: FILE: include/linux/uaccess.h:286:
+extern __must_check int copy_to_user_or_kernel(void *target, void *src,
size_t size, bool userbuf);
CHECK: No space is necessary after a cast
#128: FILE: lib/usercopy.c:100:
+ if (copy_to_user((char __user *) target, src, size))
total: 0 errors, 0 warnings, 3 checks, 96 lines checked
NOTE: For some of the reported defects, checkpatch may be able to
mechanically convert to the typical style using --fix or
--fix-inplace.
Commit 2c94767fa768 ("kdump: vmcore: remove copy_to() and add
copy_to_user_or_kernel()") has style problems, please review.
NOTE: If any of the errors are false positives, please report
them to the maintainer, see CHECKPATCH in MAINTAINERS.
Christophe
_______________________________________________
linux-riscv mailing list
linux-riscv@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-riscv
^ permalink raw reply [flat|nested] 61+ messages in thread
* Re: [PATCH v2 1/2] kdump: vmcore: remove copy_to() and add copy_to_user_or_kernel()
@ 2021-12-11 10:32 ` Christophe Leroy
0 siblings, 0 replies; 61+ messages in thread
From: Christophe Leroy @ 2021-12-11 10:32 UTC (permalink / raw)
To: Tiezhu Yang, Dave Young, Baoquan He, Vivek Goyal, Andrew Morton
Cc: linux-ia64, linux-sh, linuxppc-dev, x86, kexec, linux-mips,
linux-kernel, linux-fsdevel, linux-riscv, Xuefeng Li,
linux-arm-kernel
Le 11/12/2021 à 04:33, Tiezhu Yang a écrit :
> In arch/*/kernel/crash_dump*.c, there exist many similar code
> about copy_oldmem_page(), remove copy_to() in fs/proc/vmcore.c
> and add copy_to_user_or_kernel() in lib/usercopy.c, then we can
> use copy_to_user_or_kernel() to simplify the related code.
It should be an inline function in uaccess.h, see below why.
>
> Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
> ---
> fs/proc/vmcore.c | 28 +++++++---------------------
> include/linux/uaccess.h | 1 +
> lib/usercopy.c | 15 +++++++++++++++
> 3 files changed, 23 insertions(+), 21 deletions(-)
>
> diff --git a/fs/proc/vmcore.c b/fs/proc/vmcore.c
> index 509f851..f67fd77 100644
> --- a/fs/proc/vmcore.c
> +++ b/fs/proc/vmcore.c
> @@ -238,22 +238,8 @@ copy_oldmem_page_encrypted(unsigned long pfn, char *buf, size_t csize,
> return copy_oldmem_page(pfn, buf, csize, offset, userbuf);
> }
>
> -/*
> - * Copy to either kernel or user space
> - */
> -static int copy_to(void *target, void *src, size_t size, int userbuf)
> -{
> - if (userbuf) {
> - if (copy_to_user((char __user *) target, src, size))
> - return -EFAULT;
> - } else {
> - memcpy(target, src, size);
> - }
> - return 0;
> -}
> -
> #ifdef CONFIG_PROC_VMCORE_DEVICE_DUMP
> -static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, int userbuf)
> +static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, bool userbuf)
Changing int to bool in all the callers should be another patch. You can
have copy_to_user_or_kernel() take a bool in the patch while still
having all the callers using an int.
> {
> struct vmcoredd_node *dump;
> u64 offset = 0;
> @@ -266,7 +252,7 @@ static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, int userbuf)
> if (start < offset + dump->size) {
> tsz = min(offset + (u64)dump->size - start, (u64)size);
> buf = dump->buf + start - offset;
> - if (copy_to(dst, buf, tsz, userbuf)) {
> + if (copy_to_user_or_kernel(dst, buf, tsz, userbuf)) {
> ret = -EFAULT;
> goto out_unlock;
> }
> @@ -330,7 +316,7 @@ static int vmcoredd_mmap_dumps(struct vm_area_struct *vma, unsigned long dst,
> * returned otherwise number of bytes read are returned.
> */
> static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
> - int userbuf)
> + bool userbuf)
> {
> ssize_t acc = 0, tmp;
> size_t tsz;
> @@ -347,7 +333,7 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
> /* Read ELF core header */
> if (*fpos < elfcorebuf_sz) {
> tsz = min(elfcorebuf_sz - (size_t)*fpos, buflen);
> - if (copy_to(buffer, elfcorebuf + *fpos, tsz, userbuf))
> + if (copy_to_user_or_kernel(buffer, elfcorebuf + *fpos, tsz, userbuf))
> return -EFAULT;
> buflen -= tsz;
> *fpos += tsz;
> @@ -395,7 +381,7 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
> /* Read remaining elf notes */
> tsz = min(elfcorebuf_sz + elfnotes_sz - (size_t)*fpos, buflen);
> kaddr = elfnotes_buf + *fpos - elfcorebuf_sz - vmcoredd_orig_sz;
> - if (copy_to(buffer, kaddr, tsz, userbuf))
> + if (copy_to_user_or_kernel(buffer, kaddr, tsz, userbuf))
> return -EFAULT;
>
> buflen -= tsz;
> @@ -435,7 +421,7 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
> static ssize_t read_vmcore(struct file *file, char __user *buffer,
> size_t buflen, loff_t *fpos)
> {
> - return __read_vmcore((__force char *) buffer, buflen, fpos, 1);
> + return __read_vmcore((__force char *) buffer, buflen, fpos, true);
> }
>
> /*
> @@ -461,7 +447,7 @@ static vm_fault_t mmap_vmcore_fault(struct vm_fault *vmf)
> if (!PageUptodate(page)) {
> offset = (loff_t) index << PAGE_SHIFT;
> buf = __va((page_to_pfn(page) << PAGE_SHIFT));
> - rc = __read_vmcore(buf, PAGE_SIZE, &offset, 0);
> + rc = __read_vmcore(buf, PAGE_SIZE, &offset, false);
> if (rc < 0) {
> unlock_page(page);
> put_page(page);
> diff --git a/include/linux/uaccess.h b/include/linux/uaccess.h
> index ac03940..a25e682e 100644
> --- a/include/linux/uaccess.h
> +++ b/include/linux/uaccess.h
> @@ -283,6 +283,7 @@ __copy_from_user_inatomic_nocache(void *to, const void __user *from,
> #endif /* ARCH_HAS_NOCACHE_UACCESS */
>
> extern __must_check int check_zeroed_user(const void __user *from, size_t size);
> +extern __must_check int copy_to_user_or_kernel(void *target, void *src, size_t size, bool userbuf);
extern keyword is pointless for function prototypes, please don't add
new ones.
>
> /**
> * copy_struct_from_user: copy a struct from userspace
> diff --git a/lib/usercopy.c b/lib/usercopy.c
> index 7413dd3..7431b1b 100644
> --- a/lib/usercopy.c
> +++ b/lib/usercopy.c
> @@ -90,3 +90,18 @@ int check_zeroed_user(const void __user *from, size_t size)
> return -EFAULT;
> }
> EXPORT_SYMBOL(check_zeroed_user);
> +
> +/*
> + * Copy to either user or kernel space
> + */
> +int copy_to_user_or_kernel(void *target, void *src, size_t size, bool userbuf)
> +{
> + if (userbuf) {
> + if (copy_to_user((char __user *) target, src, size))
> + return -EFAULT;
> + } else {
> + memcpy(target, src, size);
> + }
> + return 0;
> +}
> +EXPORT_SYMBOL(copy_to_user_or_kernel);
>
Ref my answer to Andrew, I don't think outlining this fonction is a
worth it. As shown in that mail, the size of the caller is increased by
4 instructions (which is in the noise) but also this new function is not
small. So I see no benefit in term of size, and I don't think there is
any benefit in terms of performance either.
In this patch that's the same. Before the patch, read_vmcore() has a
size of 0x338.
With this patch, read_vmcore() has a size of 0x340. So that's 2
instructions more, so no benefit either.
So I think this should remain an inline function like in your first
patch (but with the new name).
000001a4 <copy_to_user_or_kernel>:
1a4: 2c 06 00 00 cmpwi r6,0
1a8: 94 21 ff f0 stwu r1,-16(r1)
1ac: 41 82 00 50 beq 1fc <copy_to_user_or_kernel+0x58>
1b0: 2c 05 00 00 cmpwi r5,0
1b4: 41 80 00 7c blt 230 <copy_to_user_or_kernel+0x8c>
1b8: 3d 00 b0 00 lis r8,-20480
1bc: 7f 83 40 40 cmplw cr7,r3,r8
1c0: 41 9c 00 14 blt cr7,1d4 <copy_to_user_or_kernel+0x30>
1c4: 40 82 00 64 bne 228 <copy_to_user_or_kernel+0x84>
1c8: 38 60 00 00 li r3,0
1cc: 38 21 00 10 addi r1,r1,16
1d0: 4e 80 00 20 blr
1d4: 7d 23 40 50 subf r9,r3,r8
1d8: 7f 85 48 40 cmplw cr7,r5,r9
1dc: 7c 08 02 a6 mflr r0
1e0: 90 01 00 14 stw r0,20(r1)
1e4: 41 9d 00 38 bgt cr7,21c <copy_to_user_or_kernel+0x78>
1e8: 48 00 00 01 bl 1e8 <copy_to_user_or_kernel+0x44>
1e8: R_PPC_REL24 __copy_tofrom_user
1ec: 80 01 00 14 lwz r0,20(r1)
1f0: 2c 03 00 00 cmpwi r3,0
1f4: 7c 08 03 a6 mtlr r0
1f8: 4b ff ff cc b 1c4 <copy_to_user_or_kernel+0x20>
1fc: 7c 08 02 a6 mflr r0
200: 90 01 00 14 stw r0,20(r1)
204: 48 00 00 01 bl 204 <copy_to_user_or_kernel+0x60>
204: R_PPC_REL24 memcpy
208: 38 60 00 00 li r3,0
20c: 80 01 00 14 lwz r0,20(r1)
210: 38 21 00 10 addi r1,r1,16
214: 7c 08 03 a6 mtlr r0
218: 4e 80 00 20 blr
21c: 80 01 00 14 lwz r0,20(r1)
220: 7c 08 03 a6 mtlr r0
224: 4b ff ff a0 b 1c4 <copy_to_user_or_kernel+0x20>
228: 38 60 ff f2 li r3,-14
22c: 4b ff ff a0 b 1cc <copy_to_user_or_kernel+0x28>
230: 0f e0 00 00 twui r0,0
234: 7c 08 02 a6 mflr r0
238: 90 01 00 14 stw r0,20(r1)
Also note that checkpatch.pl provides the following on your patch:
CHECK: No space is necessary after a cast
#88: FILE: fs/proc/vmcore.c:424:
+ return __read_vmcore((__force char *) buffer, buflen, fpos, true);
CHECK: extern prototypes should be avoided in .h files
#109: FILE: include/linux/uaccess.h:286:
+extern __must_check int copy_to_user_or_kernel(void *target, void *src,
size_t size, bool userbuf);
CHECK: No space is necessary after a cast
#128: FILE: lib/usercopy.c:100:
+ if (copy_to_user((char __user *) target, src, size))
total: 0 errors, 0 warnings, 3 checks, 96 lines checked
NOTE: For some of the reported defects, checkpatch may be able to
mechanically convert to the typical style using --fix or
--fix-inplace.
Commit 2c94767fa768 ("kdump: vmcore: remove copy_to() and add
copy_to_user_or_kernel()") has style problems, please review.
NOTE: If any of the errors are false positives, please report
them to the maintainer, see CHECKPATCH in MAINTAINERS.
Christophe
^ permalink raw reply [flat|nested] 61+ messages in thread
* Re: [PATCH v2 1/2] kdump: vmcore: remove copy_to() and add copy_to_user_or_kernel()
@ 2021-12-11 10:32 ` Christophe Leroy
0 siblings, 0 replies; 61+ messages in thread
From: Christophe Leroy @ 2021-12-11 10:32 UTC (permalink / raw)
To: Tiezhu Yang, Dave Young, Baoquan He, Vivek Goyal, Andrew Morton
Cc: linux-ia64, linux-sh, Xuefeng Li, x86, kexec, linux-mips,
linux-kernel, linux-fsdevel, linux-riscv, linuxppc-dev,
linux-arm-kernel
Le 11/12/2021 à 04:33, Tiezhu Yang a écrit :
> In arch/*/kernel/crash_dump*.c, there exist many similar code
> about copy_oldmem_page(), remove copy_to() in fs/proc/vmcore.c
> and add copy_to_user_or_kernel() in lib/usercopy.c, then we can
> use copy_to_user_or_kernel() to simplify the related code.
It should be an inline function in uaccess.h, see below why.
>
> Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
> ---
> fs/proc/vmcore.c | 28 +++++++---------------------
> include/linux/uaccess.h | 1 +
> lib/usercopy.c | 15 +++++++++++++++
> 3 files changed, 23 insertions(+), 21 deletions(-)
>
> diff --git a/fs/proc/vmcore.c b/fs/proc/vmcore.c
> index 509f851..f67fd77 100644
> --- a/fs/proc/vmcore.c
> +++ b/fs/proc/vmcore.c
> @@ -238,22 +238,8 @@ copy_oldmem_page_encrypted(unsigned long pfn, char *buf, size_t csize,
> return copy_oldmem_page(pfn, buf, csize, offset, userbuf);
> }
>
> -/*
> - * Copy to either kernel or user space
> - */
> -static int copy_to(void *target, void *src, size_t size, int userbuf)
> -{
> - if (userbuf) {
> - if (copy_to_user((char __user *) target, src, size))
> - return -EFAULT;
> - } else {
> - memcpy(target, src, size);
> - }
> - return 0;
> -}
> -
> #ifdef CONFIG_PROC_VMCORE_DEVICE_DUMP
> -static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, int userbuf)
> +static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, bool userbuf)
Changing int to bool in all the callers should be another patch. You can
have copy_to_user_or_kernel() take a bool in the patch while still
having all the callers using an int.
> {
> struct vmcoredd_node *dump;
> u64 offset = 0;
> @@ -266,7 +252,7 @@ static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, int userbuf)
> if (start < offset + dump->size) {
> tsz = min(offset + (u64)dump->size - start, (u64)size);
> buf = dump->buf + start - offset;
> - if (copy_to(dst, buf, tsz, userbuf)) {
> + if (copy_to_user_or_kernel(dst, buf, tsz, userbuf)) {
> ret = -EFAULT;
> goto out_unlock;
> }
> @@ -330,7 +316,7 @@ static int vmcoredd_mmap_dumps(struct vm_area_struct *vma, unsigned long dst,
> * returned otherwise number of bytes read are returned.
> */
> static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
> - int userbuf)
> + bool userbuf)
> {
> ssize_t acc = 0, tmp;
> size_t tsz;
> @@ -347,7 +333,7 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
> /* Read ELF core header */
> if (*fpos < elfcorebuf_sz) {
> tsz = min(elfcorebuf_sz - (size_t)*fpos, buflen);
> - if (copy_to(buffer, elfcorebuf + *fpos, tsz, userbuf))
> + if (copy_to_user_or_kernel(buffer, elfcorebuf + *fpos, tsz, userbuf))
> return -EFAULT;
> buflen -= tsz;
> *fpos += tsz;
> @@ -395,7 +381,7 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
> /* Read remaining elf notes */
> tsz = min(elfcorebuf_sz + elfnotes_sz - (size_t)*fpos, buflen);
> kaddr = elfnotes_buf + *fpos - elfcorebuf_sz - vmcoredd_orig_sz;
> - if (copy_to(buffer, kaddr, tsz, userbuf))
> + if (copy_to_user_or_kernel(buffer, kaddr, tsz, userbuf))
> return -EFAULT;
>
> buflen -= tsz;
> @@ -435,7 +421,7 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
> static ssize_t read_vmcore(struct file *file, char __user *buffer,
> size_t buflen, loff_t *fpos)
> {
> - return __read_vmcore((__force char *) buffer, buflen, fpos, 1);
> + return __read_vmcore((__force char *) buffer, buflen, fpos, true);
> }
>
> /*
> @@ -461,7 +447,7 @@ static vm_fault_t mmap_vmcore_fault(struct vm_fault *vmf)
> if (!PageUptodate(page)) {
> offset = (loff_t) index << PAGE_SHIFT;
> buf = __va((page_to_pfn(page) << PAGE_SHIFT));
> - rc = __read_vmcore(buf, PAGE_SIZE, &offset, 0);
> + rc = __read_vmcore(buf, PAGE_SIZE, &offset, false);
> if (rc < 0) {
> unlock_page(page);
> put_page(page);
> diff --git a/include/linux/uaccess.h b/include/linux/uaccess.h
> index ac03940..a25e682e 100644
> --- a/include/linux/uaccess.h
> +++ b/include/linux/uaccess.h
> @@ -283,6 +283,7 @@ __copy_from_user_inatomic_nocache(void *to, const void __user *from,
> #endif /* ARCH_HAS_NOCACHE_UACCESS */
>
> extern __must_check int check_zeroed_user(const void __user *from, size_t size);
> +extern __must_check int copy_to_user_or_kernel(void *target, void *src, size_t size, bool userbuf);
extern keyword is pointless for function prototypes, please don't add
new ones.
>
> /**
> * copy_struct_from_user: copy a struct from userspace
> diff --git a/lib/usercopy.c b/lib/usercopy.c
> index 7413dd3..7431b1b 100644
> --- a/lib/usercopy.c
> +++ b/lib/usercopy.c
> @@ -90,3 +90,18 @@ int check_zeroed_user(const void __user *from, size_t size)
> return -EFAULT;
> }
> EXPORT_SYMBOL(check_zeroed_user);
> +
> +/*
> + * Copy to either user or kernel space
> + */
> +int copy_to_user_or_kernel(void *target, void *src, size_t size, bool userbuf)
> +{
> + if (userbuf) {
> + if (copy_to_user((char __user *) target, src, size))
> + return -EFAULT;
> + } else {
> + memcpy(target, src, size);
> + }
> + return 0;
> +}
> +EXPORT_SYMBOL(copy_to_user_or_kernel);
>
Ref my answer to Andrew, I don't think outlining this fonction is a
worth it. As shown in that mail, the size of the caller is increased by
4 instructions (which is in the noise) but also this new function is not
small. So I see no benefit in term of size, and I don't think there is
any benefit in terms of performance either.
In this patch that's the same. Before the patch, read_vmcore() has a
size of 0x338.
With this patch, read_vmcore() has a size of 0x340. So that's 2
instructions more, so no benefit either.
So I think this should remain an inline function like in your first
patch (but with the new name).
000001a4 <copy_to_user_or_kernel>:
1a4: 2c 06 00 00 cmpwi r6,0
1a8: 94 21 ff f0 stwu r1,-16(r1)
1ac: 41 82 00 50 beq 1fc <copy_to_user_or_kernel+0x58>
1b0: 2c 05 00 00 cmpwi r5,0
1b4: 41 80 00 7c blt 230 <copy_to_user_or_kernel+0x8c>
1b8: 3d 00 b0 00 lis r8,-20480
1bc: 7f 83 40 40 cmplw cr7,r3,r8
1c0: 41 9c 00 14 blt cr7,1d4 <copy_to_user_or_kernel+0x30>
1c4: 40 82 00 64 bne 228 <copy_to_user_or_kernel+0x84>
1c8: 38 60 00 00 li r3,0
1cc: 38 21 00 10 addi r1,r1,16
1d0: 4e 80 00 20 blr
1d4: 7d 23 40 50 subf r9,r3,r8
1d8: 7f 85 48 40 cmplw cr7,r5,r9
1dc: 7c 08 02 a6 mflr r0
1e0: 90 01 00 14 stw r0,20(r1)
1e4: 41 9d 00 38 bgt cr7,21c <copy_to_user_or_kernel+0x78>
1e8: 48 00 00 01 bl 1e8 <copy_to_user_or_kernel+0x44>
1e8: R_PPC_REL24 __copy_tofrom_user
1ec: 80 01 00 14 lwz r0,20(r1)
1f0: 2c 03 00 00 cmpwi r3,0
1f4: 7c 08 03 a6 mtlr r0
1f8: 4b ff ff cc b 1c4 <copy_to_user_or_kernel+0x20>
1fc: 7c 08 02 a6 mflr r0
200: 90 01 00 14 stw r0,20(r1)
204: 48 00 00 01 bl 204 <copy_to_user_or_kernel+0x60>
204: R_PPC_REL24 memcpy
208: 38 60 00 00 li r3,0
20c: 80 01 00 14 lwz r0,20(r1)
210: 38 21 00 10 addi r1,r1,16
214: 7c 08 03 a6 mtlr r0
218: 4e 80 00 20 blr
21c: 80 01 00 14 lwz r0,20(r1)
220: 7c 08 03 a6 mtlr r0
224: 4b ff ff a0 b 1c4 <copy_to_user_or_kernel+0x20>
228: 38 60 ff f2 li r3,-14
22c: 4b ff ff a0 b 1cc <copy_to_user_or_kernel+0x28>
230: 0f e0 00 00 twui r0,0
234: 7c 08 02 a6 mflr r0
238: 90 01 00 14 stw r0,20(r1)
Also note that checkpatch.pl provides the following on your patch:
CHECK: No space is necessary after a cast
#88: FILE: fs/proc/vmcore.c:424:
+ return __read_vmcore((__force char *) buffer, buflen, fpos, true);
CHECK: extern prototypes should be avoided in .h files
#109: FILE: include/linux/uaccess.h:286:
+extern __must_check int copy_to_user_or_kernel(void *target, void *src,
size_t size, bool userbuf);
CHECK: No space is necessary after a cast
#128: FILE: lib/usercopy.c:100:
+ if (copy_to_user((char __user *) target, src, size))
total: 0 errors, 0 warnings, 3 checks, 96 lines checked
NOTE: For some of the reported defects, checkpatch may be able to
mechanically convert to the typical style using --fix or
--fix-inplace.
Commit 2c94767fa768 ("kdump: vmcore: remove copy_to() and add
copy_to_user_or_kernel()") has style problems, please review.
NOTE: If any of the errors are false positives, please report
them to the maintainer, see CHECKPATCH in MAINTAINERS.
Christophe
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 61+ messages in thread
* Re: [PATCH v2 1/2] kdump: vmcore: remove copy_to() and add copy_to_user_or_kernel()
@ 2021-12-11 10:32 ` Christophe Leroy
0 siblings, 0 replies; 61+ messages in thread
From: Christophe Leroy @ 2021-12-11 10:32 UTC (permalink / raw)
To: Tiezhu Yang, Dave Young, Baoquan He, Vivek Goyal, Andrew Morton
Cc: linux-ia64, linux-sh, Xuefeng Li, x86, kexec, linux-mips,
linux-kernel, linux-fsdevel, linux-riscv, linuxppc-dev,
linux-arm-kernel
Le 11/12/2021 à 04:33, Tiezhu Yang a écrit :
> In arch/*/kernel/crash_dump*.c, there exist many similar code
> about copy_oldmem_page(), remove copy_to() in fs/proc/vmcore.c
> and add copy_to_user_or_kernel() in lib/usercopy.c, then we can
> use copy_to_user_or_kernel() to simplify the related code.
It should be an inline function in uaccess.h, see below why.
>
> Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
> ---
> fs/proc/vmcore.c | 28 +++++++---------------------
> include/linux/uaccess.h | 1 +
> lib/usercopy.c | 15 +++++++++++++++
> 3 files changed, 23 insertions(+), 21 deletions(-)
>
> diff --git a/fs/proc/vmcore.c b/fs/proc/vmcore.c
> index 509f851..f67fd77 100644
> --- a/fs/proc/vmcore.c
> +++ b/fs/proc/vmcore.c
> @@ -238,22 +238,8 @@ copy_oldmem_page_encrypted(unsigned long pfn, char *buf, size_t csize,
> return copy_oldmem_page(pfn, buf, csize, offset, userbuf);
> }
>
> -/*
> - * Copy to either kernel or user space
> - */
> -static int copy_to(void *target, void *src, size_t size, int userbuf)
> -{
> - if (userbuf) {
> - if (copy_to_user((char __user *) target, src, size))
> - return -EFAULT;
> - } else {
> - memcpy(target, src, size);
> - }
> - return 0;
> -}
> -
> #ifdef CONFIG_PROC_VMCORE_DEVICE_DUMP
> -static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, int userbuf)
> +static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, bool userbuf)
Changing int to bool in all the callers should be another patch. You can
have copy_to_user_or_kernel() take a bool in the patch while still
having all the callers using an int.
> {
> struct vmcoredd_node *dump;
> u64 offset = 0;
> @@ -266,7 +252,7 @@ static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, int userbuf)
> if (start < offset + dump->size) {
> tsz = min(offset + (u64)dump->size - start, (u64)size);
> buf = dump->buf + start - offset;
> - if (copy_to(dst, buf, tsz, userbuf)) {
> + if (copy_to_user_or_kernel(dst, buf, tsz, userbuf)) {
> ret = -EFAULT;
> goto out_unlock;
> }
> @@ -330,7 +316,7 @@ static int vmcoredd_mmap_dumps(struct vm_area_struct *vma, unsigned long dst,
> * returned otherwise number of bytes read are returned.
> */
> static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
> - int userbuf)
> + bool userbuf)
> {
> ssize_t acc = 0, tmp;
> size_t tsz;
> @@ -347,7 +333,7 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
> /* Read ELF core header */
> if (*fpos < elfcorebuf_sz) {
> tsz = min(elfcorebuf_sz - (size_t)*fpos, buflen);
> - if (copy_to(buffer, elfcorebuf + *fpos, tsz, userbuf))
> + if (copy_to_user_or_kernel(buffer, elfcorebuf + *fpos, tsz, userbuf))
> return -EFAULT;
> buflen -= tsz;
> *fpos += tsz;
> @@ -395,7 +381,7 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
> /* Read remaining elf notes */
> tsz = min(elfcorebuf_sz + elfnotes_sz - (size_t)*fpos, buflen);
> kaddr = elfnotes_buf + *fpos - elfcorebuf_sz - vmcoredd_orig_sz;
> - if (copy_to(buffer, kaddr, tsz, userbuf))
> + if (copy_to_user_or_kernel(buffer, kaddr, tsz, userbuf))
> return -EFAULT;
>
> buflen -= tsz;
> @@ -435,7 +421,7 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
> static ssize_t read_vmcore(struct file *file, char __user *buffer,
> size_t buflen, loff_t *fpos)
> {
> - return __read_vmcore((__force char *) buffer, buflen, fpos, 1);
> + return __read_vmcore((__force char *) buffer, buflen, fpos, true);
> }
>
> /*
> @@ -461,7 +447,7 @@ static vm_fault_t mmap_vmcore_fault(struct vm_fault *vmf)
> if (!PageUptodate(page)) {
> offset = (loff_t) index << PAGE_SHIFT;
> buf = __va((page_to_pfn(page) << PAGE_SHIFT));
> - rc = __read_vmcore(buf, PAGE_SIZE, &offset, 0);
> + rc = __read_vmcore(buf, PAGE_SIZE, &offset, false);
> if (rc < 0) {
> unlock_page(page);
> put_page(page);
> diff --git a/include/linux/uaccess.h b/include/linux/uaccess.h
> index ac03940..a25e682e 100644
> --- a/include/linux/uaccess.h
> +++ b/include/linux/uaccess.h
> @@ -283,6 +283,7 @@ __copy_from_user_inatomic_nocache(void *to, const void __user *from,
> #endif /* ARCH_HAS_NOCACHE_UACCESS */
>
> extern __must_check int check_zeroed_user(const void __user *from, size_t size);
> +extern __must_check int copy_to_user_or_kernel(void *target, void *src, size_t size, bool userbuf);
extern keyword is pointless for function prototypes, please don't add
new ones.
>
> /**
> * copy_struct_from_user: copy a struct from userspace
> diff --git a/lib/usercopy.c b/lib/usercopy.c
> index 7413dd3..7431b1b 100644
> --- a/lib/usercopy.c
> +++ b/lib/usercopy.c
> @@ -90,3 +90,18 @@ int check_zeroed_user(const void __user *from, size_t size)
> return -EFAULT;
> }
> EXPORT_SYMBOL(check_zeroed_user);
> +
> +/*
> + * Copy to either user or kernel space
> + */
> +int copy_to_user_or_kernel(void *target, void *src, size_t size, bool userbuf)
> +{
> + if (userbuf) {
> + if (copy_to_user((char __user *) target, src, size))
> + return -EFAULT;
> + } else {
> + memcpy(target, src, size);
> + }
> + return 0;
> +}
> +EXPORT_SYMBOL(copy_to_user_or_kernel);
>
Ref my answer to Andrew, I don't think outlining this fonction is a
worth it. As shown in that mail, the size of the caller is increased by
4 instructions (which is in the noise) but also this new function is not
small. So I see no benefit in term of size, and I don't think there is
any benefit in terms of performance either.
In this patch that's the same. Before the patch, read_vmcore() has a
size of 0x338.
With this patch, read_vmcore() has a size of 0x340. So that's 2
instructions more, so no benefit either.
So I think this should remain an inline function like in your first
patch (but with the new name).
000001a4 <copy_to_user_or_kernel>:
1a4: 2c 06 00 00 cmpwi r6,0
1a8: 94 21 ff f0 stwu r1,-16(r1)
1ac: 41 82 00 50 beq 1fc <copy_to_user_or_kernel+0x58>
1b0: 2c 05 00 00 cmpwi r5,0
1b4: 41 80 00 7c blt 230 <copy_to_user_or_kernel+0x8c>
1b8: 3d 00 b0 00 lis r8,-20480
1bc: 7f 83 40 40 cmplw cr7,r3,r8
1c0: 41 9c 00 14 blt cr7,1d4 <copy_to_user_or_kernel+0x30>
1c4: 40 82 00 64 bne 228 <copy_to_user_or_kernel+0x84>
1c8: 38 60 00 00 li r3,0
1cc: 38 21 00 10 addi r1,r1,16
1d0: 4e 80 00 20 blr
1d4: 7d 23 40 50 subf r9,r3,r8
1d8: 7f 85 48 40 cmplw cr7,r5,r9
1dc: 7c 08 02 a6 mflr r0
1e0: 90 01 00 14 stw r0,20(r1)
1e4: 41 9d 00 38 bgt cr7,21c <copy_to_user_or_kernel+0x78>
1e8: 48 00 00 01 bl 1e8 <copy_to_user_or_kernel+0x44>
1e8: R_PPC_REL24 __copy_tofrom_user
1ec: 80 01 00 14 lwz r0,20(r1)
1f0: 2c 03 00 00 cmpwi r3,0
1f4: 7c 08 03 a6 mtlr r0
1f8: 4b ff ff cc b 1c4 <copy_to_user_or_kernel+0x20>
1fc: 7c 08 02 a6 mflr r0
200: 90 01 00 14 stw r0,20(r1)
204: 48 00 00 01 bl 204 <copy_to_user_or_kernel+0x60>
204: R_PPC_REL24 memcpy
208: 38 60 00 00 li r3,0
20c: 80 01 00 14 lwz r0,20(r1)
210: 38 21 00 10 addi r1,r1,16
214: 7c 08 03 a6 mtlr r0
218: 4e 80 00 20 blr
21c: 80 01 00 14 lwz r0,20(r1)
220: 7c 08 03 a6 mtlr r0
224: 4b ff ff a0 b 1c4 <copy_to_user_or_kernel+0x20>
228: 38 60 ff f2 li r3,-14
22c: 4b ff ff a0 b 1cc <copy_to_user_or_kernel+0x28>
230: 0f e0 00 00 twui r0,0
234: 7c 08 02 a6 mflr r0
238: 90 01 00 14 stw r0,20(r1)
Also note that checkpatch.pl provides the following on your patch:
CHECK: No space is necessary after a cast
#88: FILE: fs/proc/vmcore.c:424:
+ return __read_vmcore((__force char *) buffer, buflen, fpos, true);
CHECK: extern prototypes should be avoided in .h files
#109: FILE: include/linux/uaccess.h:286:
+extern __must_check int copy_to_user_or_kernel(void *target, void *src,
size_t size, bool userbuf);
CHECK: No space is necessary after a cast
#128: FILE: lib/usercopy.c:100:
+ if (copy_to_user((char __user *) target, src, size))
total: 0 errors, 0 warnings, 3 checks, 96 lines checked
NOTE: For some of the reported defects, checkpatch may be able to
mechanically convert to the typical style using --fix or
--fix-inplace.
Commit 2c94767fa768 ("kdump: vmcore: remove copy_to() and add
copy_to_user_or_kernel()") has style problems, please review.
NOTE: If any of the errors are false positives, please report
them to the maintainer, see CHECKPATCH in MAINTAINERS.
Christophe
_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec
^ permalink raw reply [flat|nested] 61+ messages in thread
* Re: [PATCH v2 1/2] kdump: vmcore: remove copy_to() and add copy_to_user_or_kernel()
@ 2021-12-11 10:32 ` Christophe Leroy
0 siblings, 0 replies; 61+ messages in thread
From: Christophe Leroy @ 2021-12-11 10:32 UTC (permalink / raw)
To: Tiezhu Yang, Dave Young, Baoquan He, Vivek Goyal, Andrew Morton
Cc: linux-ia64, linux-sh, Xuefeng Li, x86, kexec, linux-mips,
linux-kernel, linux-fsdevel, linux-riscv, linuxppc-dev,
linux-arm-kernel
DQoNCkxlIDExLzEyLzIwMjEgw6AgMDQ6MzMsIFRpZXpodSBZYW5nIGEgw6ljcml0wqA6DQo+IElu
IGFyY2gvKi9rZXJuZWwvY3Jhc2hfZHVtcCouYywgdGhlcmUgZXhpc3QgbWFueSBzaW1pbGFyIGNv
ZGUNCj4gYWJvdXQgY29weV9vbGRtZW1fcGFnZSgpLCByZW1vdmUgY29weV90bygpIGluIGZzL3By
b2Mvdm1jb3JlLmMNCj4gYW5kIGFkZCBjb3B5X3RvX3VzZXJfb3Jfa2VybmVsKCkgaW4gbGliL3Vz
ZXJjb3B5LmMsIHRoZW4gd2UgY2FuDQo+IHVzZSBjb3B5X3RvX3VzZXJfb3Jfa2VybmVsKCkgdG8g
c2ltcGxpZnkgdGhlIHJlbGF0ZWQgY29kZS4NCg0KSXQgc2hvdWxkIGJlIGFuIGlubGluZSBmdW5j
dGlvbiBpbiB1YWNjZXNzLmgsIHNlZSBiZWxvdyB3aHkuDQoNCj4gDQo+IFNpZ25lZC1vZmYtYnk6
IFRpZXpodSBZYW5nIDx5YW5ndGllemh1QGxvb25nc29uLmNuPg0KPiAtLS0NCj4gICBmcy9wcm9j
L3ZtY29yZS5jICAgICAgICB8IDI4ICsrKysrKystLS0tLS0tLS0tLS0tLS0tLS0tLS0NCj4gICBp
bmNsdWRlL2xpbnV4L3VhY2Nlc3MuaCB8ICAxICsNCj4gICBsaWIvdXNlcmNvcHkuYyAgICAgICAg
ICB8IDE1ICsrKysrKysrKysrKysrKw0KPiAgIDMgZmlsZXMgY2hhbmdlZCwgMjMgaW5zZXJ0aW9u
cygrKSwgMjEgZGVsZXRpb25zKC0pDQo+IA0KPiBkaWZmIC0tZ2l0IGEvZnMvcHJvYy92bWNvcmUu
YyBiL2ZzL3Byb2Mvdm1jb3JlLmMNCj4gaW5kZXggNTA5Zjg1MS4uZjY3ZmQ3NyAxMDA2NDQNCj4g
LS0tIGEvZnMvcHJvYy92bWNvcmUuYw0KPiArKysgYi9mcy9wcm9jL3ZtY29yZS5jDQo+IEBAIC0y
MzgsMjIgKzIzOCw4IEBAIGNvcHlfb2xkbWVtX3BhZ2VfZW5jcnlwdGVkKHVuc2lnbmVkIGxvbmcg
cGZuLCBjaGFyICpidWYsIHNpemVfdCBjc2l6ZSwNCj4gICAJcmV0dXJuIGNvcHlfb2xkbWVtX3Bh
Z2UocGZuLCBidWYsIGNzaXplLCBvZmZzZXQsIHVzZXJidWYpOw0KPiAgIH0NCj4gICANCj4gLS8q
DQo+IC0gKiBDb3B5IHRvIGVpdGhlciBrZXJuZWwgb3IgdXNlciBzcGFjZQ0KPiAtICovDQo+IC1z
dGF0aWMgaW50IGNvcHlfdG8odm9pZCAqdGFyZ2V0LCB2b2lkICpzcmMsIHNpemVfdCBzaXplLCBp
bnQgdXNlcmJ1ZikNCj4gLXsNCj4gLQlpZiAodXNlcmJ1Zikgew0KPiAtCQlpZiAoY29weV90b191
c2VyKChjaGFyIF9fdXNlciAqKSB0YXJnZXQsIHNyYywgc2l6ZSkpDQo+IC0JCQlyZXR1cm4gLUVG
QVVMVDsNCj4gLQl9IGVsc2Ugew0KPiAtCQltZW1jcHkodGFyZ2V0LCBzcmMsIHNpemUpOw0KPiAt
CX0NCj4gLQlyZXR1cm4gMDsNCj4gLX0NCj4gLQ0KPiAgICNpZmRlZiBDT05GSUdfUFJPQ19WTUNP
UkVfREVWSUNFX0RVTVANCj4gLXN0YXRpYyBpbnQgdm1jb3JlZGRfY29weV9kdW1wcyh2b2lkICpk
c3QsIHU2NCBzdGFydCwgc2l6ZV90IHNpemUsIGludCB1c2VyYnVmKQ0KPiArc3RhdGljIGludCB2
bWNvcmVkZF9jb3B5X2R1bXBzKHZvaWQgKmRzdCwgdTY0IHN0YXJ0LCBzaXplX3Qgc2l6ZSwgYm9v
bCB1c2VyYnVmKQ0KDQpDaGFuZ2luZyBpbnQgdG8gYm9vbCBpbiBhbGwgdGhlIGNhbGxlcnMgc2hv
dWxkIGJlIGFub3RoZXIgcGF0Y2guIFlvdSBjYW4gDQpoYXZlIGNvcHlfdG9fdXNlcl9vcl9rZXJu
ZWwoKSB0YWtlIGEgYm9vbCBpbiB0aGUgcGF0Y2ggd2hpbGUgc3RpbGwgDQpoYXZpbmcgYWxsIHRo
ZSBjYWxsZXJzIHVzaW5nIGFuIGludC4NCg0KPiAgIHsNCj4gICAJc3RydWN0IHZtY29yZWRkX25v
ZGUgKmR1bXA7DQo+ICAgCXU2NCBvZmZzZXQgPSAwOw0KPiBAQCAtMjY2LDcgKzI1Miw3IEBAIHN0
YXRpYyBpbnQgdm1jb3JlZGRfY29weV9kdW1wcyh2b2lkICpkc3QsIHU2NCBzdGFydCwgc2l6ZV90
IHNpemUsIGludCB1c2VyYnVmKQ0KPiAgIAkJaWYgKHN0YXJ0IDwgb2Zmc2V0ICsgZHVtcC0+c2l6
ZSkgew0KPiAgIAkJCXRzeiA9IG1pbihvZmZzZXQgKyAodTY0KWR1bXAtPnNpemUgLSBzdGFydCwg
KHU2NClzaXplKTsNCj4gICAJCQlidWYgPSBkdW1wLT5idWYgKyBzdGFydCAtIG9mZnNldDsNCj4g
LQkJCWlmIChjb3B5X3RvKGRzdCwgYnVmLCB0c3osIHVzZXJidWYpKSB7DQo+ICsJCQlpZiAoY29w
eV90b191c2VyX29yX2tlcm5lbChkc3QsIGJ1ZiwgdHN6LCB1c2VyYnVmKSkgew0KPiAgIAkJCQly
ZXQgPSAtRUZBVUxUOw0KPiAgIAkJCQlnb3RvIG91dF91bmxvY2s7DQo+ICAgCQkJfQ0KPiBAQCAt
MzMwLDcgKzMxNiw3IEBAIHN0YXRpYyBpbnQgdm1jb3JlZGRfbW1hcF9kdW1wcyhzdHJ1Y3Qgdm1f
YXJlYV9zdHJ1Y3QgKnZtYSwgdW5zaWduZWQgbG9uZyBkc3QsDQo+ICAgICogcmV0dXJuZWQgb3Ro
ZXJ3aXNlIG51bWJlciBvZiBieXRlcyByZWFkIGFyZSByZXR1cm5lZC4NCj4gICAgKi8NCj4gICBz
dGF0aWMgc3NpemVfdCBfX3JlYWRfdm1jb3JlKGNoYXIgKmJ1ZmZlciwgc2l6ZV90IGJ1Zmxlbiwg
bG9mZl90ICpmcG9zLA0KPiAtCQkJICAgICBpbnQgdXNlcmJ1ZikNCj4gKwkJCSAgICAgYm9vbCB1
c2VyYnVmKQ0KPiAgIHsNCj4gICAJc3NpemVfdCBhY2MgPSAwLCB0bXA7DQo+ICAgCXNpemVfdCB0
c3o7DQo+IEBAIC0zNDcsNyArMzMzLDcgQEAgc3RhdGljIHNzaXplX3QgX19yZWFkX3ZtY29yZShj
aGFyICpidWZmZXIsIHNpemVfdCBidWZsZW4sIGxvZmZfdCAqZnBvcywNCj4gICAJLyogUmVhZCBF
TEYgY29yZSBoZWFkZXIgKi8NCj4gICAJaWYgKCpmcG9zIDwgZWxmY29yZWJ1Zl9zeikgew0KPiAg
IAkJdHN6ID0gbWluKGVsZmNvcmVidWZfc3ogLSAoc2l6ZV90KSpmcG9zLCBidWZsZW4pOw0KPiAt
CQlpZiAoY29weV90byhidWZmZXIsIGVsZmNvcmVidWYgKyAqZnBvcywgdHN6LCB1c2VyYnVmKSkN
Cj4gKwkJaWYgKGNvcHlfdG9fdXNlcl9vcl9rZXJuZWwoYnVmZmVyLCBlbGZjb3JlYnVmICsgKmZw
b3MsIHRzeiwgdXNlcmJ1ZikpDQo+ICAgCQkJcmV0dXJuIC1FRkFVTFQ7DQo+ICAgCQlidWZsZW4g
LT0gdHN6Ow0KPiAgIAkJKmZwb3MgKz0gdHN6Ow0KPiBAQCAtMzk1LDcgKzM4MSw3IEBAIHN0YXRp
YyBzc2l6ZV90IF9fcmVhZF92bWNvcmUoY2hhciAqYnVmZmVyLCBzaXplX3QgYnVmbGVuLCBsb2Zm
X3QgKmZwb3MsDQo+ICAgCQkvKiBSZWFkIHJlbWFpbmluZyBlbGYgbm90ZXMgKi8NCj4gICAJCXRz
eiA9IG1pbihlbGZjb3JlYnVmX3N6ICsgZWxmbm90ZXNfc3ogLSAoc2l6ZV90KSpmcG9zLCBidWZs
ZW4pOw0KPiAgIAkJa2FkZHIgPSBlbGZub3Rlc19idWYgKyAqZnBvcyAtIGVsZmNvcmVidWZfc3og
LSB2bWNvcmVkZF9vcmlnX3N6Ow0KPiAtCQlpZiAoY29weV90byhidWZmZXIsIGthZGRyLCB0c3os
IHVzZXJidWYpKQ0KPiArCQlpZiAoY29weV90b191c2VyX29yX2tlcm5lbChidWZmZXIsIGthZGRy
LCB0c3osIHVzZXJidWYpKQ0KPiAgIAkJCXJldHVybiAtRUZBVUxUOw0KPiAgIA0KPiAgIAkJYnVm
bGVuIC09IHRzejsNCj4gQEAgLTQzNSw3ICs0MjEsNyBAQCBzdGF0aWMgc3NpemVfdCBfX3JlYWRf
dm1jb3JlKGNoYXIgKmJ1ZmZlciwgc2l6ZV90IGJ1ZmxlbiwgbG9mZl90ICpmcG9zLA0KPiAgIHN0
YXRpYyBzc2l6ZV90IHJlYWRfdm1jb3JlKHN0cnVjdCBmaWxlICpmaWxlLCBjaGFyIF9fdXNlciAq
YnVmZmVyLA0KPiAgIAkJCSAgIHNpemVfdCBidWZsZW4sIGxvZmZfdCAqZnBvcykNCj4gICB7DQo+
IC0JcmV0dXJuIF9fcmVhZF92bWNvcmUoKF9fZm9yY2UgY2hhciAqKSBidWZmZXIsIGJ1Zmxlbiwg
ZnBvcywgMSk7DQo+ICsJcmV0dXJuIF9fcmVhZF92bWNvcmUoKF9fZm9yY2UgY2hhciAqKSBidWZm
ZXIsIGJ1ZmxlbiwgZnBvcywgdHJ1ZSk7DQo+ICAgfQ0KPiAgIA0KPiAgIC8qDQo+IEBAIC00NjEs
NyArNDQ3LDcgQEAgc3RhdGljIHZtX2ZhdWx0X3QgbW1hcF92bWNvcmVfZmF1bHQoc3RydWN0IHZt
X2ZhdWx0ICp2bWYpDQo+ICAgCWlmICghUGFnZVVwdG9kYXRlKHBhZ2UpKSB7DQo+ICAgCQlvZmZz
ZXQgPSAobG9mZl90KSBpbmRleCA8PCBQQUdFX1NISUZUOw0KPiAgIAkJYnVmID0gX192YSgocGFn
ZV90b19wZm4ocGFnZSkgPDwgUEFHRV9TSElGVCkpOw0KPiAtCQlyYyA9IF9fcmVhZF92bWNvcmUo
YnVmLCBQQUdFX1NJWkUsICZvZmZzZXQsIDApOw0KPiArCQlyYyA9IF9fcmVhZF92bWNvcmUoYnVm
LCBQQUdFX1NJWkUsICZvZmZzZXQsIGZhbHNlKTsNCj4gICAJCWlmIChyYyA8IDApIHsNCj4gICAJ
CQl1bmxvY2tfcGFnZShwYWdlKTsNCj4gICAJCQlwdXRfcGFnZShwYWdlKTsNCj4gZGlmZiAtLWdp
dCBhL2luY2x1ZGUvbGludXgvdWFjY2Vzcy5oIGIvaW5jbHVkZS9saW51eC91YWNjZXNzLmgNCj4g
aW5kZXggYWMwMzk0MC4uYTI1ZTY4MmUgMTAwNjQ0DQo+IC0tLSBhL2luY2x1ZGUvbGludXgvdWFj
Y2Vzcy5oDQo+ICsrKyBiL2luY2x1ZGUvbGludXgvdWFjY2Vzcy5oDQo+IEBAIC0yODMsNiArMjgz
LDcgQEAgX19jb3B5X2Zyb21fdXNlcl9pbmF0b21pY19ub2NhY2hlKHZvaWQgKnRvLCBjb25zdCB2
b2lkIF9fdXNlciAqZnJvbSwNCj4gICAjZW5kaWYJCS8qIEFSQ0hfSEFTX05PQ0FDSEVfVUFDQ0VT
UyAqLw0KPiAgIA0KPiAgIGV4dGVybiBfX211c3RfY2hlY2sgaW50IGNoZWNrX3plcm9lZF91c2Vy
KGNvbnN0IHZvaWQgX191c2VyICpmcm9tLCBzaXplX3Qgc2l6ZSk7DQo+ICtleHRlcm4gX19tdXN0
X2NoZWNrIGludCBjb3B5X3RvX3VzZXJfb3Jfa2VybmVsKHZvaWQgKnRhcmdldCwgdm9pZCAqc3Jj
LCBzaXplX3Qgc2l6ZSwgYm9vbCB1c2VyYnVmKTsNCg0KZXh0ZXJuIGtleXdvcmQgaXMgcG9pbnRs
ZXNzIGZvciBmdW5jdGlvbiBwcm90b3R5cGVzLCBwbGVhc2UgZG9uJ3QgYWRkIA0KbmV3IG9uZXMu
DQoNCj4gICANCj4gICAvKioNCj4gICAgKiBjb3B5X3N0cnVjdF9mcm9tX3VzZXI6IGNvcHkgYSBz
dHJ1Y3QgZnJvbSB1c2Vyc3BhY2UNCj4gZGlmZiAtLWdpdCBhL2xpYi91c2VyY29weS5jIGIvbGli
L3VzZXJjb3B5LmMNCj4gaW5kZXggNzQxM2RkMy4uNzQzMWIxYiAxMDA2NDQNCj4gLS0tIGEvbGli
L3VzZXJjb3B5LmMNCj4gKysrIGIvbGliL3VzZXJjb3B5LmMNCj4gQEAgLTkwLDMgKzkwLDE4IEBA
IGludCBjaGVja196ZXJvZWRfdXNlcihjb25zdCB2b2lkIF9fdXNlciAqZnJvbSwgc2l6ZV90IHNp
emUpDQo+ICAgCXJldHVybiAtRUZBVUxUOw0KPiAgIH0NCj4gICBFWFBPUlRfU1lNQk9MKGNoZWNr
X3plcm9lZF91c2VyKTsNCj4gKw0KPiArLyoNCj4gKyAqIENvcHkgdG8gZWl0aGVyIHVzZXIgb3Ig
a2VybmVsIHNwYWNlDQo+ICsgKi8NCj4gK2ludCBjb3B5X3RvX3VzZXJfb3Jfa2VybmVsKHZvaWQg
KnRhcmdldCwgdm9pZCAqc3JjLCBzaXplX3Qgc2l6ZSwgYm9vbCB1c2VyYnVmKQ0KPiArew0KPiAr
CWlmICh1c2VyYnVmKSB7DQo+ICsJCWlmIChjb3B5X3RvX3VzZXIoKGNoYXIgX191c2VyICopIHRh
cmdldCwgc3JjLCBzaXplKSkNCj4gKwkJCXJldHVybiAtRUZBVUxUOw0KPiArCX0gZWxzZSB7DQo+
ICsJCW1lbWNweSh0YXJnZXQsIHNyYywgc2l6ZSk7DQo+ICsJfQ0KPiArCXJldHVybiAwOw0KPiAr
fQ0KPiArRVhQT1JUX1NZTUJPTChjb3B5X3RvX3VzZXJfb3Jfa2VybmVsKTsNCj4gDQoNClJlZiBt
eSBhbnN3ZXIgdG8gQW5kcmV3LCBJIGRvbid0IHRoaW5rIG91dGxpbmluZyB0aGlzIGZvbmN0aW9u
IGlzIGEgDQp3b3J0aCBpdC4gQXMgc2hvd24gaW4gdGhhdCBtYWlsLCB0aGUgc2l6ZSBvZiB0aGUg
Y2FsbGVyIGlzIGluY3JlYXNlZCBieSANCjQgaW5zdHJ1Y3Rpb25zICh3aGljaCBpcyBpbiB0aGUg
bm9pc2UpIGJ1dCBhbHNvIHRoaXMgbmV3IGZ1bmN0aW9uIGlzIG5vdCANCnNtYWxsLiBTbyBJIHNl
ZSBubyBiZW5lZml0IGluIHRlcm0gb2Ygc2l6ZSwgYW5kIEkgZG9uJ3QgdGhpbmsgdGhlcmUgaXMg
DQphbnkgYmVuZWZpdCBpbiB0ZXJtcyBvZiBwZXJmb3JtYW5jZSBlaXRoZXIuDQoNCkluIHRoaXMg
cGF0Y2ggdGhhdCdzIHRoZSBzYW1lLiBCZWZvcmUgdGhlIHBhdGNoLCByZWFkX3ZtY29yZSgpIGhh
cyBhIA0Kc2l6ZSBvZiAweDMzOC4NCldpdGggdGhpcyBwYXRjaCwgcmVhZF92bWNvcmUoKSBoYXMg
YSBzaXplIG9mIDB4MzQwLiBTbyB0aGF0J3MgMiANCmluc3RydWN0aW9ucyBtb3JlLCBzbyBubyBi
ZW5lZml0IGVpdGhlci4NCg0KU28gSSB0aGluayB0aGlzIHNob3VsZCByZW1haW4gYW4gaW5saW5l
IGZ1bmN0aW9uIGxpa2UgaW4geW91ciBmaXJzdCANCnBhdGNoIChidXQgd2l0aCB0aGUgbmV3IG5h
bWUpLg0KDQowMDAwMDFhNCA8Y29weV90b191c2VyX29yX2tlcm5lbD46DQogIDFhNDoJMmMgMDYg
MDAgMDAgCWNtcHdpICAgcjYsMA0KICAxYTg6CTk0IDIxIGZmIGYwIAlzdHd1ICAgIHIxLC0xNihy
MSkNCiAgMWFjOgk0MSA4MiAwMCA1MCAJYmVxICAgICAxZmMgPGNvcHlfdG9fdXNlcl9vcl9rZXJu
ZWwrMHg1OD4NCiAgMWIwOgkyYyAwNSAwMCAwMCAJY21wd2kgICByNSwwDQogIDFiNDoJNDEgODAg
MDAgN2MgCWJsdCAgICAgMjMwIDxjb3B5X3RvX3VzZXJfb3Jfa2VybmVsKzB4OGM+DQogIDFiODoJ
M2QgMDAgYjAgMDAgCWxpcyAgICAgcjgsLTIwNDgwDQogIDFiYzoJN2YgODMgNDAgNDAgCWNtcGx3
ICAgY3I3LHIzLHI4DQogIDFjMDoJNDEgOWMgMDAgMTQgCWJsdCAgICAgY3I3LDFkNCA8Y29weV90
b191c2VyX29yX2tlcm5lbCsweDMwPg0KICAxYzQ6CTQwIDgyIDAwIDY0IAlibmUgICAgIDIyOCA8
Y29weV90b191c2VyX29yX2tlcm5lbCsweDg0Pg0KICAxYzg6CTM4IDYwIDAwIDAwIAlsaSAgICAg
IHIzLDANCiAgMWNjOgkzOCAyMSAwMCAxMCAJYWRkaSAgICByMSxyMSwxNg0KICAxZDA6CTRlIDgw
IDAwIDIwIAlibHINCiAgMWQ0Ogk3ZCAyMyA0MCA1MCAJc3ViZiAgICByOSxyMyxyOA0KICAxZDg6
CTdmIDg1IDQ4IDQwIAljbXBsdyAgIGNyNyxyNSxyOQ0KICAxZGM6CTdjIDA4IDAyIGE2IAltZmxy
ICAgIHIwDQogIDFlMDoJOTAgMDEgMDAgMTQgCXN0dyAgICAgcjAsMjAocjEpDQogIDFlNDoJNDEg
OWQgMDAgMzggCWJndCAgICAgY3I3LDIxYyA8Y29weV90b191c2VyX29yX2tlcm5lbCsweDc4Pg0K
ICAxZTg6CTQ4IDAwIDAwIDAxIAlibCAgICAgIDFlOCA8Y29weV90b191c2VyX29yX2tlcm5lbCsw
eDQ0Pg0KCQkJMWU4OiBSX1BQQ19SRUwyNAlfX2NvcHlfdG9mcm9tX3VzZXINCiAgMWVjOgk4MCAw
MSAwMCAxNCAJbHd6ICAgICByMCwyMChyMSkNCiAgMWYwOgkyYyAwMyAwMCAwMCAJY21wd2kgICBy
MywwDQogIDFmNDoJN2MgMDggMDMgYTYgCW10bHIgICAgcjANCiAgMWY4Ogk0YiBmZiBmZiBjYyAJ
YiAgICAgICAxYzQgPGNvcHlfdG9fdXNlcl9vcl9rZXJuZWwrMHgyMD4NCiAgMWZjOgk3YyAwOCAw
MiBhNiAJbWZsciAgICByMA0KICAyMDA6CTkwIDAxIDAwIDE0IAlzdHcgICAgIHIwLDIwKHIxKQ0K
ICAyMDQ6CTQ4IDAwIDAwIDAxIAlibCAgICAgIDIwNCA8Y29weV90b191c2VyX29yX2tlcm5lbCsw
eDYwPg0KCQkJMjA0OiBSX1BQQ19SRUwyNAltZW1jcHkNCiAgMjA4OgkzOCA2MCAwMCAwMCAJbGkg
ICAgICByMywwDQogIDIwYzoJODAgMDEgMDAgMTQgCWx3eiAgICAgcjAsMjAocjEpDQogIDIxMDoJ
MzggMjEgMDAgMTAgCWFkZGkgICAgcjEscjEsMTYNCiAgMjE0Ogk3YyAwOCAwMyBhNiAJbXRsciAg
ICByMA0KICAyMTg6CTRlIDgwIDAwIDIwIAlibHINCiAgMjFjOgk4MCAwMSAwMCAxNCAJbHd6ICAg
ICByMCwyMChyMSkNCiAgMjIwOgk3YyAwOCAwMyBhNiAJbXRsciAgICByMA0KICAyMjQ6CTRiIGZm
IGZmIGEwIAliICAgICAgIDFjNCA8Y29weV90b191c2VyX29yX2tlcm5lbCsweDIwPg0KICAyMjg6
CTM4IDYwIGZmIGYyIAlsaSAgICAgIHIzLC0xNA0KICAyMmM6CTRiIGZmIGZmIGEwIAliICAgICAg
IDFjYyA8Y29weV90b191c2VyX29yX2tlcm5lbCsweDI4Pg0KICAyMzA6CTBmIGUwIDAwIDAwIAl0
d3VpICAgIHIwLDANCiAgMjM0Ogk3YyAwOCAwMiBhNiAJbWZsciAgICByMA0KICAyMzg6CTkwIDAx
IDAwIDE0IAlzdHcgICAgIHIwLDIwKHIxKQ0KDQoNCkFsc28gbm90ZSB0aGF0IGNoZWNrcGF0Y2gu
cGwgcHJvdmlkZXMgdGhlIGZvbGxvd2luZyBvbiB5b3VyIHBhdGNoOg0KDQpDSEVDSzogTm8gc3Bh
Y2UgaXMgbmVjZXNzYXJ5IGFmdGVyIGEgY2FzdA0KIzg4OiBGSUxFOiBmcy9wcm9jL3ZtY29yZS5j
OjQyNDoNCisJcmV0dXJuIF9fcmVhZF92bWNvcmUoKF9fZm9yY2UgY2hhciAqKSBidWZmZXIsIGJ1
ZmxlbiwgZnBvcywgdHJ1ZSk7DQoNCkNIRUNLOiBleHRlcm4gcHJvdG90eXBlcyBzaG91bGQgYmUg
YXZvaWRlZCBpbiAuaCBmaWxlcw0KIzEwOTogRklMRTogaW5jbHVkZS9saW51eC91YWNjZXNzLmg6
Mjg2Og0KK2V4dGVybiBfX211c3RfY2hlY2sgaW50IGNvcHlfdG9fdXNlcl9vcl9rZXJuZWwodm9p
ZCAqdGFyZ2V0LCB2b2lkICpzcmMsIA0Kc2l6ZV90IHNpemUsIGJvb2wgdXNlcmJ1Zik7DQoNCkNI
RUNLOiBObyBzcGFjZSBpcyBuZWNlc3NhcnkgYWZ0ZXIgYSBjYXN0DQojMTI4OiBGSUxFOiBsaWIv
dXNlcmNvcHkuYzoxMDA6DQorCQlpZiAoY29weV90b191c2VyKChjaGFyIF9fdXNlciAqKSB0YXJn
ZXQsIHNyYywgc2l6ZSkpDQoNCnRvdGFsOiAwIGVycm9ycywgMCB3YXJuaW5ncywgMyBjaGVja3Ms
IDk2IGxpbmVzIGNoZWNrZWQNCg0KTk9URTogRm9yIHNvbWUgb2YgdGhlIHJlcG9ydGVkIGRlZmVj
dHMsIGNoZWNrcGF0Y2ggbWF5IGJlIGFibGUgdG8NCiAgICAgICBtZWNoYW5pY2FsbHkgY29udmVy
dCB0byB0aGUgdHlwaWNhbCBzdHlsZSB1c2luZyAtLWZpeCBvciANCi0tZml4LWlucGxhY2UuDQoN
CkNvbW1pdCAyYzk0NzY3ZmE3NjggKCJrZHVtcDogdm1jb3JlOiByZW1vdmUgY29weV90bygpIGFu
ZCBhZGQgDQpjb3B5X3RvX3VzZXJfb3Jfa2VybmVsKCkiKSBoYXMgc3R5bGUgcHJvYmxlbXMsIHBs
ZWFzZSByZXZpZXcuDQoNCk5PVEU6IElmIGFueSBvZiB0aGUgZXJyb3JzIGFyZSBmYWxzZSBwb3Np
dGl2ZXMsIHBsZWFzZSByZXBvcnQNCiAgICAgICB0aGVtIHRvIHRoZSBtYWludGFpbmVyLCBzZWUg
Q0hFQ0tQQVRDSCBpbiBNQUlOVEFJTkVSUy4NCg0KDQpDaHJpc3RvcGhl
^ permalink raw reply [flat|nested] 61+ messages in thread
* RE: [PATCH v2 0/2] kdump: simplify code
2021-12-11 3:33 ` Tiezhu Yang
` (2 preceding siblings ...)
(?)
@ 2021-12-11 17:53 ` David Laight
-1 siblings, 0 replies; 61+ messages in thread
From: David Laight @ 2021-12-11 17:53 UTC (permalink / raw)
To: 'Tiezhu Yang',
Dave Young, Baoquan He, Vivek Goyal, Andrew Morton
Cc: linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li
From: Tiezhu Yang
> Sent: 11 December 2021 03:33
>
> v2:
> -- add copy_to_user_or_kernel() in lib/usercopy.c
> -- define userbuf as bool type
Instead of having a flag to indicate whether the buffer is user or kernel,
would it be better to have two separate buffer pointers.
One for a user space buffer, the other for a kernel space buffer.
Exactly one of the buffers should always be NULL.
That way the flag is never incorrectly set.
David
-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)
^ permalink raw reply [flat|nested] 61+ messages in thread
* RE: [PATCH v2 0/2] kdump: simplify code
@ 2021-12-11 17:53 ` David Laight
0 siblings, 0 replies; 61+ messages in thread
From: David Laight @ 2021-12-11 17:53 UTC (permalink / raw)
To: 'Tiezhu Yang',
Dave Young, Baoquan He, Vivek Goyal, Andrew Morton
Cc: linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li
From: Tiezhu Yang
> Sent: 11 December 2021 03:33
>
> v2:
> -- add copy_to_user_or_kernel() in lib/usercopy.c
> -- define userbuf as bool type
Instead of having a flag to indicate whether the buffer is user or kernel,
would it be better to have two separate buffer pointers.
One for a user space buffer, the other for a kernel space buffer.
Exactly one of the buffers should always be NULL.
That way the flag is never incorrectly set.
David
-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)
_______________________________________________
linux-riscv mailing list
linux-riscv@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-riscv
^ permalink raw reply [flat|nested] 61+ messages in thread
* RE: [PATCH v2 0/2] kdump: simplify code
@ 2021-12-11 17:53 ` David Laight
0 siblings, 0 replies; 61+ messages in thread
From: David Laight @ 2021-12-11 17:53 UTC (permalink / raw)
To: 'Tiezhu Yang',
Dave Young, Baoquan He, Vivek Goyal, Andrew Morton
Cc: linux-ia64, linux-sh, Xuefeng Li, x86, kexec, linux-mips,
linux-kernel, linux-fsdevel, linux-riscv, linuxppc-dev,
linux-arm-kernel
From: Tiezhu Yang
> Sent: 11 December 2021 03:33
>
> v2:
> -- add copy_to_user_or_kernel() in lib/usercopy.c
> -- define userbuf as bool type
Instead of having a flag to indicate whether the buffer is user or kernel,
would it be better to have two separate buffer pointers.
One for a user space buffer, the other for a kernel space buffer.
Exactly one of the buffers should always be NULL.
That way the flag is never incorrectly set.
David
-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)
^ permalink raw reply [flat|nested] 61+ messages in thread
* RE: [PATCH v2 0/2] kdump: simplify code
@ 2021-12-11 17:53 ` David Laight
0 siblings, 0 replies; 61+ messages in thread
From: David Laight @ 2021-12-11 17:53 UTC (permalink / raw)
To: 'Tiezhu Yang',
Dave Young, Baoquan He, Vivek Goyal, Andrew Morton
Cc: linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li
From: Tiezhu Yang
> Sent: 11 December 2021 03:33
>
> v2:
> -- add copy_to_user_or_kernel() in lib/usercopy.c
> -- define userbuf as bool type
Instead of having a flag to indicate whether the buffer is user or kernel,
would it be better to have two separate buffer pointers.
One for a user space buffer, the other for a kernel space buffer.
Exactly one of the buffers should always be NULL.
That way the flag is never incorrectly set.
David
-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 61+ messages in thread
* RE: [PATCH v2 0/2] kdump: simplify code
@ 2021-12-11 17:53 ` David Laight
0 siblings, 0 replies; 61+ messages in thread
From: David Laight @ 2021-12-11 17:53 UTC (permalink / raw)
To: 'Tiezhu Yang',
Dave Young, Baoquan He, Vivek Goyal, Andrew Morton
Cc: linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li
From: Tiezhu Yang
> Sent: 11 December 2021 03:33
>
> v2:
> -- add copy_to_user_or_kernel() in lib/usercopy.c
> -- define userbuf as bool type
Instead of having a flag to indicate whether the buffer is user or kernel,
would it be better to have two separate buffer pointers.
One for a user space buffer, the other for a kernel space buffer.
Exactly one of the buffers should always be NULL.
That way the flag is never incorrectly set.
David
-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)
_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec
^ permalink raw reply [flat|nested] 61+ messages in thread
* Re: [PATCH v2 0/2] kdump: simplify code
2021-12-11 17:53 ` David Laight
` (3 preceding siblings ...)
(?)
@ 2021-12-11 19:32 ` Christophe Leroy
-1 siblings, 0 replies; 61+ messages in thread
From: Christophe Leroy @ 2021-12-11 19:32 UTC (permalink / raw)
To: David Laight, 'Tiezhu Yang',
Dave Young, Baoquan He, Vivek Goyal, Andrew Morton
Cc: linux-ia64, linux-sh, Xuefeng Li, x86, kexec, linux-mips,
linux-kernel, linux-fsdevel, linux-riscv, linuxppc-dev,
linux-arm-kernel
Le 11/12/2021 à 18:53, David Laight a écrit :
> From: Tiezhu Yang
>> Sent: 11 December 2021 03:33
>>
>> v2:
>> -- add copy_to_user_or_kernel() in lib/usercopy.c
>> -- define userbuf as bool type
>
> Instead of having a flag to indicate whether the buffer is user or kernel,
> would it be better to have two separate buffer pointers.
> One for a user space buffer, the other for a kernel space buffer.
> Exactly one of the buffers should always be NULL.
>
> That way the flag is never incorrectly set.
>
It's a very good idea.
I was worried about the casts forcing the __user property away and back.
With that approach we will preserve the __user tags on user buffers and
enable sparse checking.
The only little drawback I see is that apparently GCC doesn't consider
the NULL value as a constant and therefore doesn't perform constant
folding on pointers. Not sure if this is a problem here.
Christophe
^ permalink raw reply [flat|nested] 61+ messages in thread
* Re: [PATCH v2 0/2] kdump: simplify code
@ 2021-12-11 19:32 ` Christophe Leroy
0 siblings, 0 replies; 61+ messages in thread
From: Christophe Leroy @ 2021-12-11 19:32 UTC (permalink / raw)
To: David Laight, 'Tiezhu Yang',
Dave Young, Baoquan He, Vivek Goyal, Andrew Morton
Cc: linux-ia64, linux-sh, Xuefeng Li, x86, kexec, linux-mips,
linux-kernel, linux-fsdevel, linux-riscv, linuxppc-dev,
linux-arm-kernel
Le 11/12/2021 à 18:53, David Laight a écrit :
> From: Tiezhu Yang
>> Sent: 11 December 2021 03:33
>>
>> v2:
>> -- add copy_to_user_or_kernel() in lib/usercopy.c
>> -- define userbuf as bool type
>
> Instead of having a flag to indicate whether the buffer is user or kernel,
> would it be better to have two separate buffer pointers.
> One for a user space buffer, the other for a kernel space buffer.
> Exactly one of the buffers should always be NULL.
>
> That way the flag is never incorrectly set.
>
It's a very good idea.
I was worried about the casts forcing the __user property away and back.
With that approach we will preserve the __user tags on user buffers and
enable sparse checking.
The only little drawback I see is that apparently GCC doesn't consider
the NULL value as a constant and therefore doesn't perform constant
folding on pointers. Not sure if this is a problem here.
Christophe
_______________________________________________
linux-riscv mailing list
linux-riscv@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-riscv
^ permalink raw reply [flat|nested] 61+ messages in thread
* Re: [PATCH v2 0/2] kdump: simplify code
@ 2021-12-11 19:32 ` Christophe Leroy
0 siblings, 0 replies; 61+ messages in thread
From: Christophe Leroy @ 2021-12-11 19:32 UTC (permalink / raw)
To: David Laight, 'Tiezhu Yang',
Dave Young, Baoquan He, Vivek Goyal, Andrew Morton
Cc: linux-ia64, linux-sh, linuxppc-dev, x86, kexec, linux-mips,
linux-kernel, linux-fsdevel, linux-riscv, Xuefeng Li,
linux-arm-kernel
Le 11/12/2021 à 18:53, David Laight a écrit :
> From: Tiezhu Yang
>> Sent: 11 December 2021 03:33
>>
>> v2:
>> -- add copy_to_user_or_kernel() in lib/usercopy.c
>> -- define userbuf as bool type
>
> Instead of having a flag to indicate whether the buffer is user or kernel,
> would it be better to have two separate buffer pointers.
> One for a user space buffer, the other for a kernel space buffer.
> Exactly one of the buffers should always be NULL.
>
> That way the flag is never incorrectly set.
>
It's a very good idea.
I was worried about the casts forcing the __user property away and back.
With that approach we will preserve the __user tags on user buffers and
enable sparse checking.
The only little drawback I see is that apparently GCC doesn't consider
the NULL value as a constant and therefore doesn't perform constant
folding on pointers. Not sure if this is a problem here.
Christophe
^ permalink raw reply [flat|nested] 61+ messages in thread
* Re: [PATCH v2 0/2] kdump: simplify code
@ 2021-12-11 19:32 ` Christophe Leroy
0 siblings, 0 replies; 61+ messages in thread
From: Christophe Leroy @ 2021-12-11 19:32 UTC (permalink / raw)
To: David Laight, 'Tiezhu Yang',
Dave Young, Baoquan He, Vivek Goyal, Andrew Morton
Cc: linux-ia64, linux-sh, Xuefeng Li, x86, kexec, linux-mips,
linux-kernel, linux-fsdevel, linux-riscv, linuxppc-dev,
linux-arm-kernel
Le 11/12/2021 à 18:53, David Laight a écrit :
> From: Tiezhu Yang
>> Sent: 11 December 2021 03:33
>>
>> v2:
>> -- add copy_to_user_or_kernel() in lib/usercopy.c
>> -- define userbuf as bool type
>
> Instead of having a flag to indicate whether the buffer is user or kernel,
> would it be better to have two separate buffer pointers.
> One for a user space buffer, the other for a kernel space buffer.
> Exactly one of the buffers should always be NULL.
>
> That way the flag is never incorrectly set.
>
It's a very good idea.
I was worried about the casts forcing the __user property away and back.
With that approach we will preserve the __user tags on user buffers and
enable sparse checking.
The only little drawback I see is that apparently GCC doesn't consider
the NULL value as a constant and therefore doesn't perform constant
folding on pointers. Not sure if this is a problem here.
Christophe
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 61+ messages in thread
* Re: [PATCH v2 0/2] kdump: simplify code
@ 2021-12-11 19:32 ` Christophe Leroy
0 siblings, 0 replies; 61+ messages in thread
From: Christophe Leroy @ 2021-12-11 19:32 UTC (permalink / raw)
To: David Laight, 'Tiezhu Yang',
Dave Young, Baoquan He, Vivek Goyal, Andrew Morton
Cc: linux-ia64, linux-sh, Xuefeng Li, x86, kexec, linux-mips,
linux-kernel, linux-fsdevel, linux-riscv, linuxppc-dev,
linux-arm-kernel
Le 11/12/2021 à 18:53, David Laight a écrit :
> From: Tiezhu Yang
>> Sent: 11 December 2021 03:33
>>
>> v2:
>> -- add copy_to_user_or_kernel() in lib/usercopy.c
>> -- define userbuf as bool type
>
> Instead of having a flag to indicate whether the buffer is user or kernel,
> would it be better to have two separate buffer pointers.
> One for a user space buffer, the other for a kernel space buffer.
> Exactly one of the buffers should always be NULL.
>
> That way the flag is never incorrectly set.
>
It's a very good idea.
I was worried about the casts forcing the __user property away and back.
With that approach we will preserve the __user tags on user buffers and
enable sparse checking.
The only little drawback I see is that apparently GCC doesn't consider
the NULL value as a constant and therefore doesn't perform constant
folding on pointers. Not sure if this is a problem here.
Christophe
_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec
^ permalink raw reply [flat|nested] 61+ messages in thread
* Re: [PATCH v2 0/2] kdump: simplify code
@ 2021-12-11 19:32 ` Christophe Leroy
0 siblings, 0 replies; 61+ messages in thread
From: Christophe Leroy @ 2021-12-11 19:32 UTC (permalink / raw)
To: David Laight, 'Tiezhu Yang',
Dave Young, Baoquan He, Vivek Goyal, Andrew Morton
Cc: linux-ia64, linux-sh, Xuefeng Li, x86, kexec, linux-mips,
linux-kernel, linux-fsdevel, linux-riscv, linuxppc-dev,
linux-arm-kernel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^ permalink raw reply [flat|nested] 61+ messages in thread
* Re: [PATCH v2 0/2] kdump: simplify code
2021-12-11 17:53 ` David Laight
` (2 preceding siblings ...)
(?)
@ 2021-12-12 11:47 ` Matthew Wilcox
-1 siblings, 0 replies; 61+ messages in thread
From: Matthew Wilcox @ 2021-12-12 11:47 UTC (permalink / raw)
To: David Laight
Cc: 'Tiezhu Yang',
Dave Young, Baoquan He, Vivek Goyal, Andrew Morton,
linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li
On Sat, Dec 11, 2021 at 05:53:46PM +0000, David Laight wrote:
> From: Tiezhu Yang
> > Sent: 11 December 2021 03:33
> >
> > v2:
> > -- add copy_to_user_or_kernel() in lib/usercopy.c
> > -- define userbuf as bool type
>
> Instead of having a flag to indicate whether the buffer is user or kernel,
> would it be better to have two separate buffer pointers.
> One for a user space buffer, the other for a kernel space buffer.
> Exactly one of the buffers should always be NULL.
No. You should be using an iov_iter instead. See
https://lore.kernel.org/all/Ya4bdB0UBJCZhUSo@casper.infradead.org/
for a start on this.
^ permalink raw reply [flat|nested] 61+ messages in thread
* Re: [PATCH v2 0/2] kdump: simplify code
@ 2021-12-12 11:47 ` Matthew Wilcox
0 siblings, 0 replies; 61+ messages in thread
From: Matthew Wilcox @ 2021-12-12 11:47 UTC (permalink / raw)
To: David Laight
Cc: 'Tiezhu Yang',
Dave Young, Baoquan He, Vivek Goyal, Andrew Morton,
linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li
On Sat, Dec 11, 2021 at 05:53:46PM +0000, David Laight wrote:
> From: Tiezhu Yang
> > Sent: 11 December 2021 03:33
> >
> > v2:
> > -- add copy_to_user_or_kernel() in lib/usercopy.c
> > -- define userbuf as bool type
>
> Instead of having a flag to indicate whether the buffer is user or kernel,
> would it be better to have two separate buffer pointers.
> One for a user space buffer, the other for a kernel space buffer.
> Exactly one of the buffers should always be NULL.
No. You should be using an iov_iter instead. See
https://lore.kernel.org/all/Ya4bdB0UBJCZhUSo@casper.infradead.org/
for a start on this.
_______________________________________________
linux-riscv mailing list
linux-riscv@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-riscv
^ permalink raw reply [flat|nested] 61+ messages in thread
* Re: [PATCH v2 0/2] kdump: simplify code
@ 2021-12-12 11:47 ` Matthew Wilcox
0 siblings, 0 replies; 61+ messages in thread
From: Matthew Wilcox @ 2021-12-12 11:47 UTC (permalink / raw)
To: David Laight
Cc: linux-ia64, Baoquan He, linux-sh, linuxppc-dev, x86, kexec,
linux-mips, linux-riscv, linux-kernel, linux-arm-kernel,
Xuefeng Li, linux-fsdevel, Andrew Morton, Dave Young,
'Tiezhu Yang',
Vivek Goyal
On Sat, Dec 11, 2021 at 05:53:46PM +0000, David Laight wrote:
> From: Tiezhu Yang
> > Sent: 11 December 2021 03:33
> >
> > v2:
> > -- add copy_to_user_or_kernel() in lib/usercopy.c
> > -- define userbuf as bool type
>
> Instead of having a flag to indicate whether the buffer is user or kernel,
> would it be better to have two separate buffer pointers.
> One for a user space buffer, the other for a kernel space buffer.
> Exactly one of the buffers should always be NULL.
No. You should be using an iov_iter instead. See
https://lore.kernel.org/all/Ya4bdB0UBJCZhUSo@casper.infradead.org/
for a start on this.
^ permalink raw reply [flat|nested] 61+ messages in thread
* Re: [PATCH v2 0/2] kdump: simplify code
@ 2021-12-12 11:47 ` Matthew Wilcox
0 siblings, 0 replies; 61+ messages in thread
From: Matthew Wilcox @ 2021-12-12 11:47 UTC (permalink / raw)
To: David Laight
Cc: 'Tiezhu Yang',
Dave Young, Baoquan He, Vivek Goyal, Andrew Morton,
linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li
On Sat, Dec 11, 2021 at 05:53:46PM +0000, David Laight wrote:
> From: Tiezhu Yang
> > Sent: 11 December 2021 03:33
> >
> > v2:
> > -- add copy_to_user_or_kernel() in lib/usercopy.c
> > -- define userbuf as bool type
>
> Instead of having a flag to indicate whether the buffer is user or kernel,
> would it be better to have two separate buffer pointers.
> One for a user space buffer, the other for a kernel space buffer.
> Exactly one of the buffers should always be NULL.
No. You should be using an iov_iter instead. See
https://lore.kernel.org/all/Ya4bdB0UBJCZhUSo@casper.infradead.org/
for a start on this.
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 61+ messages in thread
* Re: [PATCH v2 0/2] kdump: simplify code
@ 2021-12-12 11:47 ` Matthew Wilcox
0 siblings, 0 replies; 61+ messages in thread
From: Matthew Wilcox @ 2021-12-12 11:47 UTC (permalink / raw)
To: David Laight
Cc: 'Tiezhu Yang',
Dave Young, Baoquan He, Vivek Goyal, Andrew Morton,
linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li
On Sat, Dec 11, 2021 at 05:53:46PM +0000, David Laight wrote:
> From: Tiezhu Yang
> > Sent: 11 December 2021 03:33
> >
> > v2:
> > -- add copy_to_user_or_kernel() in lib/usercopy.c
> > -- define userbuf as bool type
>
> Instead of having a flag to indicate whether the buffer is user or kernel,
> would it be better to have two separate buffer pointers.
> One for a user space buffer, the other for a kernel space buffer.
> Exactly one of the buffers should always be NULL.
No. You should be using an iov_iter instead. See
https://lore.kernel.org/all/Ya4bdB0UBJCZhUSo@casper.infradead.org/
for a start on this.
_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec
^ permalink raw reply [flat|nested] 61+ messages in thread
* RE: [PATCH v2 0/2] kdump: simplify code
2021-12-12 11:47 ` Matthew Wilcox
` (2 preceding siblings ...)
(?)
@ 2021-12-13 8:30 ` David Laight
-1 siblings, 0 replies; 61+ messages in thread
From: David Laight @ 2021-12-13 8:30 UTC (permalink / raw)
To: 'Matthew Wilcox'
Cc: 'Tiezhu Yang',
Dave Young, Baoquan He, Vivek Goyal, Andrew Morton,
linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li
From: Matthew Wilcox
> Sent: 12 December 2021 11:48
>
> On Sat, Dec 11, 2021 at 05:53:46PM +0000, David Laight wrote:
> > From: Tiezhu Yang
> > > Sent: 11 December 2021 03:33
> > >
> > > v2:
> > > -- add copy_to_user_or_kernel() in lib/usercopy.c
> > > -- define userbuf as bool type
> >
> > Instead of having a flag to indicate whether the buffer is user or kernel,
> > would it be better to have two separate buffer pointers.
> > One for a user space buffer, the other for a kernel space buffer.
> > Exactly one of the buffers should always be NULL.
>
> No. You should be using an iov_iter instead. See
> https://lore.kernel.org/all/Ya4bdB0UBJCZhUSo@casper.infradead.org/
> for a start on this.
iov_iter gets horribly expensive...
David
-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)
^ permalink raw reply [flat|nested] 61+ messages in thread
* RE: [PATCH v2 0/2] kdump: simplify code
@ 2021-12-13 8:30 ` David Laight
0 siblings, 0 replies; 61+ messages in thread
From: David Laight @ 2021-12-13 8:30 UTC (permalink / raw)
To: 'Matthew Wilcox'
Cc: 'Tiezhu Yang',
Dave Young, Baoquan He, Vivek Goyal, Andrew Morton,
linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li
From: Matthew Wilcox
> Sent: 12 December 2021 11:48
>
> On Sat, Dec 11, 2021 at 05:53:46PM +0000, David Laight wrote:
> > From: Tiezhu Yang
> > > Sent: 11 December 2021 03:33
> > >
> > > v2:
> > > -- add copy_to_user_or_kernel() in lib/usercopy.c
> > > -- define userbuf as bool type
> >
> > Instead of having a flag to indicate whether the buffer is user or kernel,
> > would it be better to have two separate buffer pointers.
> > One for a user space buffer, the other for a kernel space buffer.
> > Exactly one of the buffers should always be NULL.
>
> No. You should be using an iov_iter instead. See
> https://lore.kernel.org/all/Ya4bdB0UBJCZhUSo@casper.infradead.org/
> for a start on this.
iov_iter gets horribly expensive...
David
-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)
_______________________________________________
linux-riscv mailing list
linux-riscv@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-riscv
^ permalink raw reply [flat|nested] 61+ messages in thread
* RE: [PATCH v2 0/2] kdump: simplify code
@ 2021-12-13 8:30 ` David Laight
0 siblings, 0 replies; 61+ messages in thread
From: David Laight @ 2021-12-13 8:30 UTC (permalink / raw)
To: 'Matthew Wilcox'
Cc: linux-ia64, Baoquan He, linux-sh, linuxppc-dev, x86, kexec,
linux-mips, linux-riscv, linux-kernel, linux-arm-kernel,
Xuefeng Li, linux-fsdevel, Andrew Morton, Dave Young,
'Tiezhu Yang',
Vivek Goyal
From: Matthew Wilcox
> Sent: 12 December 2021 11:48
>
> On Sat, Dec 11, 2021 at 05:53:46PM +0000, David Laight wrote:
> > From: Tiezhu Yang
> > > Sent: 11 December 2021 03:33
> > >
> > > v2:
> > > -- add copy_to_user_or_kernel() in lib/usercopy.c
> > > -- define userbuf as bool type
> >
> > Instead of having a flag to indicate whether the buffer is user or kernel,
> > would it be better to have two separate buffer pointers.
> > One for a user space buffer, the other for a kernel space buffer.
> > Exactly one of the buffers should always be NULL.
>
> No. You should be using an iov_iter instead. See
> https://lore.kernel.org/all/Ya4bdB0UBJCZhUSo@casper.infradead.org/
> for a start on this.
iov_iter gets horribly expensive...
David
-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)
^ permalink raw reply [flat|nested] 61+ messages in thread
* RE: [PATCH v2 0/2] kdump: simplify code
@ 2021-12-13 8:30 ` David Laight
0 siblings, 0 replies; 61+ messages in thread
From: David Laight @ 2021-12-13 8:30 UTC (permalink / raw)
To: 'Matthew Wilcox'
Cc: 'Tiezhu Yang',
Dave Young, Baoquan He, Vivek Goyal, Andrew Morton,
linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li
From: Matthew Wilcox
> Sent: 12 December 2021 11:48
>
> On Sat, Dec 11, 2021 at 05:53:46PM +0000, David Laight wrote:
> > From: Tiezhu Yang
> > > Sent: 11 December 2021 03:33
> > >
> > > v2:
> > > -- add copy_to_user_or_kernel() in lib/usercopy.c
> > > -- define userbuf as bool type
> >
> > Instead of having a flag to indicate whether the buffer is user or kernel,
> > would it be better to have two separate buffer pointers.
> > One for a user space buffer, the other for a kernel space buffer.
> > Exactly one of the buffers should always be NULL.
>
> No. You should be using an iov_iter instead. See
> https://lore.kernel.org/all/Ya4bdB0UBJCZhUSo@casper.infradead.org/
> for a start on this.
iov_iter gets horribly expensive...
David
-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 61+ messages in thread
* RE: [PATCH v2 0/2] kdump: simplify code
@ 2021-12-13 8:30 ` David Laight
0 siblings, 0 replies; 61+ messages in thread
From: David Laight @ 2021-12-13 8:30 UTC (permalink / raw)
To: 'Matthew Wilcox'
Cc: 'Tiezhu Yang',
Dave Young, Baoquan He, Vivek Goyal, Andrew Morton,
linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li
From: Matthew Wilcox
> Sent: 12 December 2021 11:48
>
> On Sat, Dec 11, 2021 at 05:53:46PM +0000, David Laight wrote:
> > From: Tiezhu Yang
> > > Sent: 11 December 2021 03:33
> > >
> > > v2:
> > > -- add copy_to_user_or_kernel() in lib/usercopy.c
> > > -- define userbuf as bool type
> >
> > Instead of having a flag to indicate whether the buffer is user or kernel,
> > would it be better to have two separate buffer pointers.
> > One for a user space buffer, the other for a kernel space buffer.
> > Exactly one of the buffers should always be NULL.
>
> No. You should be using an iov_iter instead. See
> https://lore.kernel.org/all/Ya4bdB0UBJCZhUSo@casper.infradead.org/
> for a start on this.
iov_iter gets horribly expensive...
David
-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)
_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec
^ permalink raw reply [flat|nested] 61+ messages in thread
* Re: [PATCH v2 0/2] kdump: simplify code
2021-12-13 8:30 ` David Laight
` (2 preceding siblings ...)
(?)
@ 2021-12-13 14:43 ` Matthew Wilcox
-1 siblings, 0 replies; 61+ messages in thread
From: Matthew Wilcox @ 2021-12-13 14:43 UTC (permalink / raw)
To: David Laight
Cc: 'Tiezhu Yang',
Dave Young, Baoquan He, Vivek Goyal, Andrew Morton,
linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li
On Mon, Dec 13, 2021 at 08:30:33AM +0000, David Laight wrote:
> From: Matthew Wilcox
> > Sent: 12 December 2021 11:48
> >
> > On Sat, Dec 11, 2021 at 05:53:46PM +0000, David Laight wrote:
> > > From: Tiezhu Yang
> > > > Sent: 11 December 2021 03:33
> > > >
> > > > v2:
> > > > -- add copy_to_user_or_kernel() in lib/usercopy.c
> > > > -- define userbuf as bool type
> > >
> > > Instead of having a flag to indicate whether the buffer is user or kernel,
> > > would it be better to have two separate buffer pointers.
> > > One for a user space buffer, the other for a kernel space buffer.
> > > Exactly one of the buffers should always be NULL.
> >
> > No. You should be using an iov_iter instead. See
> > https://lore.kernel.org/all/Ya4bdB0UBJCZhUSo@casper.infradead.org/
> > for a start on this.
>
> iov_iter gets horribly expensive...
Oh, right. Reading the kcore is a high-performance path, my mistake.
^ permalink raw reply [flat|nested] 61+ messages in thread
* Re: [PATCH v2 0/2] kdump: simplify code
@ 2021-12-13 14:43 ` Matthew Wilcox
0 siblings, 0 replies; 61+ messages in thread
From: Matthew Wilcox @ 2021-12-13 14:43 UTC (permalink / raw)
To: David Laight
Cc: linux-ia64, Baoquan He, linux-sh, linuxppc-dev, x86, kexec,
linux-mips, linux-riscv, linux-kernel, linux-arm-kernel,
Xuefeng Li, linux-fsdevel, Andrew Morton, Dave Young,
'Tiezhu Yang',
Vivek Goyal
On Mon, Dec 13, 2021 at 08:30:33AM +0000, David Laight wrote:
> From: Matthew Wilcox
> > Sent: 12 December 2021 11:48
> >
> > On Sat, Dec 11, 2021 at 05:53:46PM +0000, David Laight wrote:
> > > From: Tiezhu Yang
> > > > Sent: 11 December 2021 03:33
> > > >
> > > > v2:
> > > > -- add copy_to_user_or_kernel() in lib/usercopy.c
> > > > -- define userbuf as bool type
> > >
> > > Instead of having a flag to indicate whether the buffer is user or kernel,
> > > would it be better to have two separate buffer pointers.
> > > One for a user space buffer, the other for a kernel space buffer.
> > > Exactly one of the buffers should always be NULL.
> >
> > No. You should be using an iov_iter instead. See
> > https://lore.kernel.org/all/Ya4bdB0UBJCZhUSo@casper.infradead.org/
> > for a start on this.
>
> iov_iter gets horribly expensive...
Oh, right. Reading the kcore is a high-performance path, my mistake.
^ permalink raw reply [flat|nested] 61+ messages in thread
* Re: [PATCH v2 0/2] kdump: simplify code
@ 2021-12-13 14:43 ` Matthew Wilcox
0 siblings, 0 replies; 61+ messages in thread
From: Matthew Wilcox @ 2021-12-13 14:43 UTC (permalink / raw)
To: David Laight
Cc: 'Tiezhu Yang',
Dave Young, Baoquan He, Vivek Goyal, Andrew Morton,
linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li
On Mon, Dec 13, 2021 at 08:30:33AM +0000, David Laight wrote:
> From: Matthew Wilcox
> > Sent: 12 December 2021 11:48
> >
> > On Sat, Dec 11, 2021 at 05:53:46PM +0000, David Laight wrote:
> > > From: Tiezhu Yang
> > > > Sent: 11 December 2021 03:33
> > > >
> > > > v2:
> > > > -- add copy_to_user_or_kernel() in lib/usercopy.c
> > > > -- define userbuf as bool type
> > >
> > > Instead of having a flag to indicate whether the buffer is user or kernel,
> > > would it be better to have two separate buffer pointers.
> > > One for a user space buffer, the other for a kernel space buffer.
> > > Exactly one of the buffers should always be NULL.
> >
> > No. You should be using an iov_iter instead. See
> > https://lore.kernel.org/all/Ya4bdB0UBJCZhUSo@casper.infradead.org/
> > for a start on this.
>
> iov_iter gets horribly expensive...
Oh, right. Reading the kcore is a high-performance path, my mistake.
_______________________________________________
linux-riscv mailing list
linux-riscv@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-riscv
^ permalink raw reply [flat|nested] 61+ messages in thread
* Re: [PATCH v2 0/2] kdump: simplify code
@ 2021-12-13 14:43 ` Matthew Wilcox
0 siblings, 0 replies; 61+ messages in thread
From: Matthew Wilcox @ 2021-12-13 14:43 UTC (permalink / raw)
To: David Laight
Cc: 'Tiezhu Yang',
Dave Young, Baoquan He, Vivek Goyal, Andrew Morton,
linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li
On Mon, Dec 13, 2021 at 08:30:33AM +0000, David Laight wrote:
> From: Matthew Wilcox
> > Sent: 12 December 2021 11:48
> >
> > On Sat, Dec 11, 2021 at 05:53:46PM +0000, David Laight wrote:
> > > From: Tiezhu Yang
> > > > Sent: 11 December 2021 03:33
> > > >
> > > > v2:
> > > > -- add copy_to_user_or_kernel() in lib/usercopy.c
> > > > -- define userbuf as bool type
> > >
> > > Instead of having a flag to indicate whether the buffer is user or kernel,
> > > would it be better to have two separate buffer pointers.
> > > One for a user space buffer, the other for a kernel space buffer.
> > > Exactly one of the buffers should always be NULL.
> >
> > No. You should be using an iov_iter instead. See
> > https://lore.kernel.org/all/Ya4bdB0UBJCZhUSo@casper.infradead.org/
> > for a start on this.
>
> iov_iter gets horribly expensive...
Oh, right. Reading the kcore is a high-performance path, my mistake.
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 61+ messages in thread
* Re: [PATCH v2 0/2] kdump: simplify code
@ 2021-12-13 14:43 ` Matthew Wilcox
0 siblings, 0 replies; 61+ messages in thread
From: Matthew Wilcox @ 2021-12-13 14:43 UTC (permalink / raw)
To: David Laight
Cc: 'Tiezhu Yang',
Dave Young, Baoquan He, Vivek Goyal, Andrew Morton,
linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li
On Mon, Dec 13, 2021 at 08:30:33AM +0000, David Laight wrote:
> From: Matthew Wilcox
> > Sent: 12 December 2021 11:48
> >
> > On Sat, Dec 11, 2021 at 05:53:46PM +0000, David Laight wrote:
> > > From: Tiezhu Yang
> > > > Sent: 11 December 2021 03:33
> > > >
> > > > v2:
> > > > -- add copy_to_user_or_kernel() in lib/usercopy.c
> > > > -- define userbuf as bool type
> > >
> > > Instead of having a flag to indicate whether the buffer is user or kernel,
> > > would it be better to have two separate buffer pointers.
> > > One for a user space buffer, the other for a kernel space buffer.
> > > Exactly one of the buffers should always be NULL.
> >
> > No. You should be using an iov_iter instead. See
> > https://lore.kernel.org/all/Ya4bdB0UBJCZhUSo@casper.infradead.org/
> > for a start on this.
>
> iov_iter gets horribly expensive...
Oh, right. Reading the kcore is a high-performance path, my mistake.
_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec
^ permalink raw reply [flat|nested] 61+ messages in thread
* Re: [PATCH v2 0/2] kdump: simplify code
2021-12-13 14:43 ` Matthew Wilcox
` (3 preceding siblings ...)
(?)
@ 2021-12-14 10:03 ` Tiezhu Yang
-1 siblings, 0 replies; 61+ messages in thread
From: Tiezhu Yang @ 2021-12-14 10:03 UTC (permalink / raw)
To: Matthew Wilcox, David Laight
Cc: Dave Young, Baoquan He, Vivek Goyal, Andrew Morton,
linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li, Christophe Leroy
On 12/13/2021 10:43 PM, Matthew Wilcox wrote:
> On Mon, Dec 13, 2021 at 08:30:33AM +0000, David Laight wrote:
>> From: Matthew Wilcox
>>> Sent: 12 December 2021 11:48
>>>
>>> On Sat, Dec 11, 2021 at 05:53:46PM +0000, David Laight wrote:
>>>> From: Tiezhu Yang
>>>>> Sent: 11 December 2021 03:33
>>>>>
>>>>> v2:
>>>>> -- add copy_to_user_or_kernel() in lib/usercopy.c
>>>>> -- define userbuf as bool type
>>>>
>>>> Instead of having a flag to indicate whether the buffer is user or kernel,
>>>> would it be better to have two separate buffer pointers.
>>>> One for a user space buffer, the other for a kernel space buffer.
>>>> Exactly one of the buffers should always be NULL.
>>>
>>> No. You should be using an iov_iter instead. See
>>> https://lore.kernel.org/all/Ya4bdB0UBJCZhUSo@casper.infradead.org/
>>> for a start on this.
>>
>> iov_iter gets horribly expensive...
>
> Oh, right. Reading the kcore is a high-performance path, my mistake.
>
Hi,
Thank you for your discussions.
The intention of this patchset is to simplify the related code with no
functional changes and no side effects.
At this moment, if you are OK, I will send v3 used with inline function
copy_to_user_or_kernel() to keep it simple, maybe other more changes can
be done in the future if no any side effect.
The v3 will contain the following three patches to make the changes
more clear:
kdump: vmcore: remove copy_to() and add copy_to_user_or_kernel()
kdump: crashdump: use copy_to_user_or_kernel() to simplify code
kdump: vmcore: crashdump: make variable type of userbuf as bool
^ permalink raw reply [flat|nested] 61+ messages in thread
* Re: [PATCH v2 0/2] kdump: simplify code
@ 2021-12-14 10:03 ` Tiezhu Yang
0 siblings, 0 replies; 61+ messages in thread
From: Tiezhu Yang @ 2021-12-14 10:03 UTC (permalink / raw)
To: Matthew Wilcox, David Laight
Cc: Dave Young, Baoquan He, Vivek Goyal, Andrew Morton,
linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li, Christophe Leroy
On 12/13/2021 10:43 PM, Matthew Wilcox wrote:
> On Mon, Dec 13, 2021 at 08:30:33AM +0000, David Laight wrote:
>> From: Matthew Wilcox
>>> Sent: 12 December 2021 11:48
>>>
>>> On Sat, Dec 11, 2021 at 05:53:46PM +0000, David Laight wrote:
>>>> From: Tiezhu Yang
>>>>> Sent: 11 December 2021 03:33
>>>>>
>>>>> v2:
>>>>> -- add copy_to_user_or_kernel() in lib/usercopy.c
>>>>> -- define userbuf as bool type
>>>>
>>>> Instead of having a flag to indicate whether the buffer is user or kernel,
>>>> would it be better to have two separate buffer pointers.
>>>> One for a user space buffer, the other for a kernel space buffer.
>>>> Exactly one of the buffers should always be NULL.
>>>
>>> No. You should be using an iov_iter instead. See
>>> https://lore.kernel.org/all/Ya4bdB0UBJCZhUSo@casper.infradead.org/
>>> for a start on this.
>>
>> iov_iter gets horribly expensive...
>
> Oh, right. Reading the kcore is a high-performance path, my mistake.
>
Hi,
Thank you for your discussions.
The intention of this patchset is to simplify the related code with no
functional changes and no side effects.
At this moment, if you are OK, I will send v3 used with inline function
copy_to_user_or_kernel() to keep it simple, maybe other more changes can
be done in the future if no any side effect.
The v3 will contain the following three patches to make the changes
more clear:
kdump: vmcore: remove copy_to() and add copy_to_user_or_kernel()
kdump: crashdump: use copy_to_user_or_kernel() to simplify code
kdump: vmcore: crashdump: make variable type of userbuf as bool
_______________________________________________
linux-riscv mailing list
linux-riscv@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-riscv
^ permalink raw reply [flat|nested] 61+ messages in thread
* Re: [PATCH v2 0/2] kdump: simplify code
@ 2021-12-14 10:03 ` Tiezhu Yang
0 siblings, 0 replies; 61+ messages in thread
From: Tiezhu Yang @ 2021-12-14 10:03 UTC (permalink / raw)
To: Matthew Wilcox, David Laight
Cc: linux-ia64, Baoquan He, linux-sh, linuxppc-dev, x86, kexec,
linux-mips, linux-kernel, Vivek Goyal, Xuefeng Li, linux-fsdevel,
Andrew Morton, Dave Young, linux-riscv, linux-arm-kernel
On 12/13/2021 10:43 PM, Matthew Wilcox wrote:
> On Mon, Dec 13, 2021 at 08:30:33AM +0000, David Laight wrote:
>> From: Matthew Wilcox
>>> Sent: 12 December 2021 11:48
>>>
>>> On Sat, Dec 11, 2021 at 05:53:46PM +0000, David Laight wrote:
>>>> From: Tiezhu Yang
>>>>> Sent: 11 December 2021 03:33
>>>>>
>>>>> v2:
>>>>> -- add copy_to_user_or_kernel() in lib/usercopy.c
>>>>> -- define userbuf as bool type
>>>>
>>>> Instead of having a flag to indicate whether the buffer is user or kernel,
>>>> would it be better to have two separate buffer pointers.
>>>> One for a user space buffer, the other for a kernel space buffer.
>>>> Exactly one of the buffers should always be NULL.
>>>
>>> No. You should be using an iov_iter instead. See
>>> https://lore.kernel.org/all/Ya4bdB0UBJCZhUSo@casper.infradead.org/
>>> for a start on this.
>>
>> iov_iter gets horribly expensive...
>
> Oh, right. Reading the kcore is a high-performance path, my mistake.
>
Hi,
Thank you for your discussions.
The intention of this patchset is to simplify the related code with no
functional changes and no side effects.
At this moment, if you are OK, I will send v3 used with inline function
copy_to_user_or_kernel() to keep it simple, maybe other more changes can
be done in the future if no any side effect.
The v3 will contain the following three patches to make the changes
more clear:
kdump: vmcore: remove copy_to() and add copy_to_user_or_kernel()
kdump: crashdump: use copy_to_user_or_kernel() to simplify code
kdump: vmcore: crashdump: make variable type of userbuf as bool
^ permalink raw reply [flat|nested] 61+ messages in thread
* Re: [PATCH v2 0/2] kdump: simplify code
@ 2021-12-14 10:03 ` Tiezhu Yang
0 siblings, 0 replies; 61+ messages in thread
From: Tiezhu Yang @ 2021-12-14 10:03 UTC (permalink / raw)
To: Matthew Wilcox, David Laight
Cc: Dave Young, Baoquan He, Vivek Goyal, Andrew Morton,
linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li, Christophe Leroy
On 12/13/2021 10:43 PM, Matthew Wilcox wrote:
> On Mon, Dec 13, 2021 at 08:30:33AM +0000, David Laight wrote:
>> From: Matthew Wilcox
>>> Sent: 12 December 2021 11:48
>>>
>>> On Sat, Dec 11, 2021 at 05:53:46PM +0000, David Laight wrote:
>>>> From: Tiezhu Yang
>>>>> Sent: 11 December 2021 03:33
>>>>>
>>>>> v2:
>>>>> -- add copy_to_user_or_kernel() in lib/usercopy.c
>>>>> -- define userbuf as bool type
>>>>
>>>> Instead of having a flag to indicate whether the buffer is user or kernel,
>>>> would it be better to have two separate buffer pointers.
>>>> One for a user space buffer, the other for a kernel space buffer.
>>>> Exactly one of the buffers should always be NULL.
>>>
>>> No. You should be using an iov_iter instead. See
>>> https://lore.kernel.org/all/Ya4bdB0UBJCZhUSo@casper.infradead.org/
>>> for a start on this.
>>
>> iov_iter gets horribly expensive...
>
> Oh, right. Reading the kcore is a high-performance path, my mistake.
>
Hi,
Thank you for your discussions.
The intention of this patchset is to simplify the related code with no
functional changes and no side effects.
At this moment, if you are OK, I will send v3 used with inline function
copy_to_user_or_kernel() to keep it simple, maybe other more changes can
be done in the future if no any side effect.
The v3 will contain the following three patches to make the changes
more clear:
kdump: vmcore: remove copy_to() and add copy_to_user_or_kernel()
kdump: crashdump: use copy_to_user_or_kernel() to simplify code
kdump: vmcore: crashdump: make variable type of userbuf as bool
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 61+ messages in thread
* Re: [PATCH v2 0/2] kdump: simplify code
@ 2021-12-14 10:03 ` Tiezhu Yang
0 siblings, 0 replies; 61+ messages in thread
From: Tiezhu Yang @ 2021-12-14 10:03 UTC (permalink / raw)
To: Matthew Wilcox, David Laight
Cc: Dave Young, Baoquan He, Vivek Goyal, Andrew Morton,
linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li, Christophe Leroy
On 12/13/2021 10:43 PM, Matthew Wilcox wrote:
> On Mon, Dec 13, 2021 at 08:30:33AM +0000, David Laight wrote:
>> From: Matthew Wilcox
>>> Sent: 12 December 2021 11:48
>>>
>>> On Sat, Dec 11, 2021 at 05:53:46PM +0000, David Laight wrote:
>>>> From: Tiezhu Yang
>>>>> Sent: 11 December 2021 03:33
>>>>>
>>>>> v2:
>>>>> -- add copy_to_user_or_kernel() in lib/usercopy.c
>>>>> -- define userbuf as bool type
>>>>
>>>> Instead of having a flag to indicate whether the buffer is user or kernel,
>>>> would it be better to have two separate buffer pointers.
>>>> One for a user space buffer, the other for a kernel space buffer.
>>>> Exactly one of the buffers should always be NULL.
>>>
>>> No. You should be using an iov_iter instead. See
>>> https://lore.kernel.org/all/Ya4bdB0UBJCZhUSo@casper.infradead.org/
>>> for a start on this.
>>
>> iov_iter gets horribly expensive...
>
> Oh, right. Reading the kcore is a high-performance path, my mistake.
>
Hi,
Thank you for your discussions.
The intention of this patchset is to simplify the related code with no
functional changes and no side effects.
At this moment, if you are OK, I will send v3 used with inline function
copy_to_user_or_kernel() to keep it simple, maybe other more changes can
be done in the future if no any side effect.
The v3 will contain the following three patches to make the changes
more clear:
kdump: vmcore: remove copy_to() and add copy_to_user_or_kernel()
kdump: crashdump: use copy_to_user_or_kernel() to simplify code
kdump: vmcore: crashdump: make variable type of userbuf as bool
_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec
^ permalink raw reply [flat|nested] 61+ messages in thread
* Re: [PATCH v2 0/2] kdump: simplify code
@ 2021-12-14 10:03 ` Tiezhu Yang
0 siblings, 0 replies; 61+ messages in thread
From: Tiezhu Yang @ 2021-12-14 10:03 UTC (permalink / raw)
To: Matthew Wilcox, David Laight
Cc: Dave Young, Baoquan He, Vivek Goyal, Andrew Morton,
linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li, Christophe Leroy
On 12/13/2021 10:43 PM, Matthew Wilcox wrote:
> On Mon, Dec 13, 2021 at 08:30:33AM +0000, David Laight wrote:
>> From: Matthew Wilcox
>>> Sent: 12 December 2021 11:48
>>>
>>> On Sat, Dec 11, 2021 at 05:53:46PM +0000, David Laight wrote:
>>>> From: Tiezhu Yang
>>>>> Sent: 11 December 2021 03:33
>>>>>
>>>>> v2:
>>>>> -- add copy_to_user_or_kernel() in lib/usercopy.c
>>>>> -- define userbuf as bool type
>>>>
>>>> Instead of having a flag to indicate whether the buffer is user or kernel,
>>>> would it be better to have two separate buffer pointers.
>>>> One for a user space buffer, the other for a kernel space buffer.
>>>> Exactly one of the buffers should always be NULL.
>>>
>>> No. You should be using an iov_iter instead. See
>>> https://lore.kernel.org/all/Ya4bdB0UBJCZhUSo@casper.infradead.org/
>>> for a start on this.
>>
>> iov_iter gets horribly expensive...
>
> Oh, right. Reading the kcore is a high-performance path, my mistake.
>
Hi,
Thank you for your discussions.
The intention of this patchset is to simplify the related code with no
functional changes and no side effects.
At this moment, if you are OK, I will send v3 used with inline function
copy_to_user_or_kernel() to keep it simple, maybe other more changes can
be done in the future if no any side effect.
The v3 will contain the following three patches to make the changes
more clear:
kdump: vmcore: remove copy_to() and add copy_to_user_or_kernel()
kdump: crashdump: use copy_to_user_or_kernel() to simplify code
kdump: vmcore: crashdump: make variable type of userbuf as bool
^ permalink raw reply [flat|nested] 61+ messages in thread
* Re: [PATCH v2 0/2] kdump: simplify code
2021-12-14 10:03 ` Tiezhu Yang
` (2 preceding siblings ...)
(?)
@ 2021-12-14 13:14 ` Matthew Wilcox
-1 siblings, 0 replies; 61+ messages in thread
From: Matthew Wilcox @ 2021-12-14 13:14 UTC (permalink / raw)
To: Tiezhu Yang
Cc: David Laight, Dave Young, Baoquan He, Vivek Goyal, Andrew Morton,
linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li, Christophe Leroy
On Tue, Dec 14, 2021 at 06:03:11PM +0800, Tiezhu Yang wrote:
> On 12/13/2021 10:43 PM, Matthew Wilcox wrote:
> > On Mon, Dec 13, 2021 at 08:30:33AM +0000, David Laight wrote:
> > > From: Matthew Wilcox
> > > > Sent: 12 December 2021 11:48
> > > >
> > > > On Sat, Dec 11, 2021 at 05:53:46PM +0000, David Laight wrote:
> > > > > From: Tiezhu Yang
> > > > > > Sent: 11 December 2021 03:33
> > > > > >
> > > > > > v2:
> > > > > > -- add copy_to_user_or_kernel() in lib/usercopy.c
> > > > > > -- define userbuf as bool type
> > > > >
> > > > > Instead of having a flag to indicate whether the buffer is user or kernel,
> > > > > would it be better to have two separate buffer pointers.
> > > > > One for a user space buffer, the other for a kernel space buffer.
> > > > > Exactly one of the buffers should always be NULL.
> > > >
> > > > No. You should be using an iov_iter instead. See
> > > > https://lore.kernel.org/all/Ya4bdB0UBJCZhUSo@casper.infradead.org/
> > > > for a start on this.
> > >
> > > iov_iter gets horribly expensive...
> >
> > Oh, right. Reading the kcore is a high-performance path, my mistake.
> >
>
> Hi,
>
> Thank you for your discussions.
>
> The intention of this patchset is to simplify the related code with no
> functional changes and no side effects.
>
> At this moment, if you are OK, I will send v3 used with inline function
> copy_to_user_or_kernel() to keep it simple, maybe other more changes can
> be done in the future if no any side effect.
That would be pointless. I already sent a series to remove this,
which you were cc'd on.
^ permalink raw reply [flat|nested] 61+ messages in thread
* Re: [PATCH v2 0/2] kdump: simplify code
@ 2021-12-14 13:14 ` Matthew Wilcox
0 siblings, 0 replies; 61+ messages in thread
From: Matthew Wilcox @ 2021-12-14 13:14 UTC (permalink / raw)
To: Tiezhu Yang
Cc: David Laight, Dave Young, Baoquan He, Vivek Goyal, Andrew Morton,
linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li, Christophe Leroy
On Tue, Dec 14, 2021 at 06:03:11PM +0800, Tiezhu Yang wrote:
> On 12/13/2021 10:43 PM, Matthew Wilcox wrote:
> > On Mon, Dec 13, 2021 at 08:30:33AM +0000, David Laight wrote:
> > > From: Matthew Wilcox
> > > > Sent: 12 December 2021 11:48
> > > >
> > > > On Sat, Dec 11, 2021 at 05:53:46PM +0000, David Laight wrote:
> > > > > From: Tiezhu Yang
> > > > > > Sent: 11 December 2021 03:33
> > > > > >
> > > > > > v2:
> > > > > > -- add copy_to_user_or_kernel() in lib/usercopy.c
> > > > > > -- define userbuf as bool type
> > > > >
> > > > > Instead of having a flag to indicate whether the buffer is user or kernel,
> > > > > would it be better to have two separate buffer pointers.
> > > > > One for a user space buffer, the other for a kernel space buffer.
> > > > > Exactly one of the buffers should always be NULL.
> > > >
> > > > No. You should be using an iov_iter instead. See
> > > > https://lore.kernel.org/all/Ya4bdB0UBJCZhUSo@casper.infradead.org/
> > > > for a start on this.
> > >
> > > iov_iter gets horribly expensive...
> >
> > Oh, right. Reading the kcore is a high-performance path, my mistake.
> >
>
> Hi,
>
> Thank you for your discussions.
>
> The intention of this patchset is to simplify the related code with no
> functional changes and no side effects.
>
> At this moment, if you are OK, I will send v3 used with inline function
> copy_to_user_or_kernel() to keep it simple, maybe other more changes can
> be done in the future if no any side effect.
That would be pointless. I already sent a series to remove this,
which you were cc'd on.
_______________________________________________
linux-riscv mailing list
linux-riscv@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-riscv
^ permalink raw reply [flat|nested] 61+ messages in thread
* Re: [PATCH v2 0/2] kdump: simplify code
@ 2021-12-14 13:14 ` Matthew Wilcox
0 siblings, 0 replies; 61+ messages in thread
From: Matthew Wilcox @ 2021-12-14 13:14 UTC (permalink / raw)
To: Tiezhu Yang
Cc: linux-ia64, Baoquan He, linux-sh, linuxppc-dev, x86, kexec,
linux-mips, linux-kernel, David Laight, Vivek Goyal, Xuefeng Li,
linux-fsdevel, Andrew Morton, Dave Young, linux-riscv,
linux-arm-kernel
On Tue, Dec 14, 2021 at 06:03:11PM +0800, Tiezhu Yang wrote:
> On 12/13/2021 10:43 PM, Matthew Wilcox wrote:
> > On Mon, Dec 13, 2021 at 08:30:33AM +0000, David Laight wrote:
> > > From: Matthew Wilcox
> > > > Sent: 12 December 2021 11:48
> > > >
> > > > On Sat, Dec 11, 2021 at 05:53:46PM +0000, David Laight wrote:
> > > > > From: Tiezhu Yang
> > > > > > Sent: 11 December 2021 03:33
> > > > > >
> > > > > > v2:
> > > > > > -- add copy_to_user_or_kernel() in lib/usercopy.c
> > > > > > -- define userbuf as bool type
> > > > >
> > > > > Instead of having a flag to indicate whether the buffer is user or kernel,
> > > > > would it be better to have two separate buffer pointers.
> > > > > One for a user space buffer, the other for a kernel space buffer.
> > > > > Exactly one of the buffers should always be NULL.
> > > >
> > > > No. You should be using an iov_iter instead. See
> > > > https://lore.kernel.org/all/Ya4bdB0UBJCZhUSo@casper.infradead.org/
> > > > for a start on this.
> > >
> > > iov_iter gets horribly expensive...
> >
> > Oh, right. Reading the kcore is a high-performance path, my mistake.
> >
>
> Hi,
>
> Thank you for your discussions.
>
> The intention of this patchset is to simplify the related code with no
> functional changes and no side effects.
>
> At this moment, if you are OK, I will send v3 used with inline function
> copy_to_user_or_kernel() to keep it simple, maybe other more changes can
> be done in the future if no any side effect.
That would be pointless. I already sent a series to remove this,
which you were cc'd on.
^ permalink raw reply [flat|nested] 61+ messages in thread
* Re: [PATCH v2 0/2] kdump: simplify code
@ 2021-12-14 13:14 ` Matthew Wilcox
0 siblings, 0 replies; 61+ messages in thread
From: Matthew Wilcox @ 2021-12-14 13:14 UTC (permalink / raw)
To: Tiezhu Yang
Cc: David Laight, Dave Young, Baoquan He, Vivek Goyal, Andrew Morton,
linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li, Christophe Leroy
On Tue, Dec 14, 2021 at 06:03:11PM +0800, Tiezhu Yang wrote:
> On 12/13/2021 10:43 PM, Matthew Wilcox wrote:
> > On Mon, Dec 13, 2021 at 08:30:33AM +0000, David Laight wrote:
> > > From: Matthew Wilcox
> > > > Sent: 12 December 2021 11:48
> > > >
> > > > On Sat, Dec 11, 2021 at 05:53:46PM +0000, David Laight wrote:
> > > > > From: Tiezhu Yang
> > > > > > Sent: 11 December 2021 03:33
> > > > > >
> > > > > > v2:
> > > > > > -- add copy_to_user_or_kernel() in lib/usercopy.c
> > > > > > -- define userbuf as bool type
> > > > >
> > > > > Instead of having a flag to indicate whether the buffer is user or kernel,
> > > > > would it be better to have two separate buffer pointers.
> > > > > One for a user space buffer, the other for a kernel space buffer.
> > > > > Exactly one of the buffers should always be NULL.
> > > >
> > > > No. You should be using an iov_iter instead. See
> > > > https://lore.kernel.org/all/Ya4bdB0UBJCZhUSo@casper.infradead.org/
> > > > for a start on this.
> > >
> > > iov_iter gets horribly expensive...
> >
> > Oh, right. Reading the kcore is a high-performance path, my mistake.
> >
>
> Hi,
>
> Thank you for your discussions.
>
> The intention of this patchset is to simplify the related code with no
> functional changes and no side effects.
>
> At this moment, if you are OK, I will send v3 used with inline function
> copy_to_user_or_kernel() to keep it simple, maybe other more changes can
> be done in the future if no any side effect.
That would be pointless. I already sent a series to remove this,
which you were cc'd on.
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 61+ messages in thread
* Re: [PATCH v2 0/2] kdump: simplify code
@ 2021-12-14 13:14 ` Matthew Wilcox
0 siblings, 0 replies; 61+ messages in thread
From: Matthew Wilcox @ 2021-12-14 13:14 UTC (permalink / raw)
To: Tiezhu Yang
Cc: David Laight, Dave Young, Baoquan He, Vivek Goyal, Andrew Morton,
linux-arm-kernel, linux-ia64, linux-mips, linuxppc-dev,
linux-riscv, linux-sh, x86, linux-fsdevel, kexec, linux-kernel,
Xuefeng Li, Christophe Leroy
On Tue, Dec 14, 2021 at 06:03:11PM +0800, Tiezhu Yang wrote:
> On 12/13/2021 10:43 PM, Matthew Wilcox wrote:
> > On Mon, Dec 13, 2021 at 08:30:33AM +0000, David Laight wrote:
> > > From: Matthew Wilcox
> > > > Sent: 12 December 2021 11:48
> > > >
> > > > On Sat, Dec 11, 2021 at 05:53:46PM +0000, David Laight wrote:
> > > > > From: Tiezhu Yang
> > > > > > Sent: 11 December 2021 03:33
> > > > > >
> > > > > > v2:
> > > > > > -- add copy_to_user_or_kernel() in lib/usercopy.c
> > > > > > -- define userbuf as bool type
> > > > >
> > > > > Instead of having a flag to indicate whether the buffer is user or kernel,
> > > > > would it be better to have two separate buffer pointers.
> > > > > One for a user space buffer, the other for a kernel space buffer.
> > > > > Exactly one of the buffers should always be NULL.
> > > >
> > > > No. You should be using an iov_iter instead. See
> > > > https://lore.kernel.org/all/Ya4bdB0UBJCZhUSo@casper.infradead.org/
> > > > for a start on this.
> > >
> > > iov_iter gets horribly expensive...
> >
> > Oh, right. Reading the kcore is a high-performance path, my mistake.
> >
>
> Hi,
>
> Thank you for your discussions.
>
> The intention of this patchset is to simplify the related code with no
> functional changes and no side effects.
>
> At this moment, if you are OK, I will send v3 used with inline function
> copy_to_user_or_kernel() to keep it simple, maybe other more changes can
> be done in the future if no any side effect.
That would be pointless. I already sent a series to remove this,
which you were cc'd on.
_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec
^ permalink raw reply [flat|nested] 61+ messages in thread
end of thread, other threads:[~2021-12-14 13:16 UTC | newest]
Thread overview: 61+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-12-11 3:33 [PATCH v2 0/2] kdump: simplify code Tiezhu Yang
2021-12-11 3:33 ` Tiezhu Yang
2021-12-11 3:33 ` Tiezhu Yang
2021-12-11 3:33 ` Tiezhu Yang
2021-12-11 3:33 ` Tiezhu Yang
2021-12-11 3:33 ` Tiezhu Yang
2021-12-11 3:33 ` [PATCH v2 1/2] kdump: vmcore: remove copy_to() and add copy_to_user_or_kernel() Tiezhu Yang
2021-12-11 3:33 ` Tiezhu Yang
2021-12-11 3:33 ` Tiezhu Yang
2021-12-11 3:33 ` Tiezhu Yang
2021-12-11 3:33 ` Tiezhu Yang
2021-12-11 3:33 ` Tiezhu Yang
2021-12-11 10:32 ` Christophe Leroy
2021-12-11 10:32 ` Christophe Leroy
2021-12-11 10:32 ` Christophe Leroy
2021-12-11 10:32 ` Christophe Leroy
2021-12-11 10:32 ` Christophe Leroy
2021-12-11 10:32 ` Christophe Leroy
2021-12-11 3:33 ` [PATCH v2 2/2] kdump: crashdump: use copy_to_user_or_kernel() to simplify code Tiezhu Yang
2021-12-11 3:33 ` Tiezhu Yang
2021-12-11 3:33 ` Tiezhu Yang
2021-12-11 3:33 ` Tiezhu Yang
2021-12-11 3:33 ` Tiezhu Yang
2021-12-11 3:33 ` Tiezhu Yang
2021-12-11 17:53 ` [PATCH v2 0/2] kdump: " David Laight
2021-12-11 17:53 ` David Laight
2021-12-11 17:53 ` David Laight
2021-12-11 17:53 ` David Laight
2021-12-11 17:53 ` David Laight
2021-12-11 19:32 ` Christophe Leroy
2021-12-11 19:32 ` Christophe Leroy
2021-12-11 19:32 ` Christophe Leroy
2021-12-11 19:32 ` Christophe Leroy
2021-12-11 19:32 ` Christophe Leroy
2021-12-11 19:32 ` Christophe Leroy
2021-12-12 11:47 ` Matthew Wilcox
2021-12-12 11:47 ` Matthew Wilcox
2021-12-12 11:47 ` Matthew Wilcox
2021-12-12 11:47 ` Matthew Wilcox
2021-12-12 11:47 ` Matthew Wilcox
2021-12-13 8:30 ` David Laight
2021-12-13 8:30 ` David Laight
2021-12-13 8:30 ` David Laight
2021-12-13 8:30 ` David Laight
2021-12-13 8:30 ` David Laight
2021-12-13 14:43 ` Matthew Wilcox
2021-12-13 14:43 ` Matthew Wilcox
2021-12-13 14:43 ` Matthew Wilcox
2021-12-13 14:43 ` Matthew Wilcox
2021-12-13 14:43 ` Matthew Wilcox
2021-12-14 10:03 ` Tiezhu Yang
2021-12-14 10:03 ` Tiezhu Yang
2021-12-14 10:03 ` Tiezhu Yang
2021-12-14 10:03 ` Tiezhu Yang
2021-12-14 10:03 ` Tiezhu Yang
2021-12-14 10:03 ` Tiezhu Yang
2021-12-14 13:14 ` Matthew Wilcox
2021-12-14 13:14 ` Matthew Wilcox
2021-12-14 13:14 ` Matthew Wilcox
2021-12-14 13:14 ` Matthew Wilcox
2021-12-14 13:14 ` Matthew Wilcox
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.