uprobe_write_opcode() must not cross page boundary; prepare_uprobe()
relies on arch_uprobe_analyze_insn() which should validate "vaddr" but
some architectures (csky, s390, and sparc) don't do this.
We can remove the BUG_ON() check in prepare_uprobe() and validate the
offset early in __uprobe_register(). The new IS_ALIGNED() check matches
the alignment check in arch_prepare_kprobe() on supported architectures,
so I think that all insns must be aligned to UPROBE_SWBP_INSN_SIZE.
Another problem is __update_ref_ctr() which was wrong from the very
beginning, it can read/write outside of kmap'ed page unless "vaddr" is
aligned to sizeof(short), __uprobe_register() should check this too.
Cc: [email protected]
Reported-by: Linus Torvalds <[email protected]>
Suggested-by: Linus Torvalds <[email protected]>
Signed-off-by: Oleg Nesterov <[email protected]>
---
kernel/events/uprobes.c | 16 ++++++++++++----
1 file changed, 12 insertions(+), 4 deletions(-)
diff --git a/kernel/events/uprobes.c b/kernel/events/uprobes.c
index ece7e13f6e4a..cc2095607c74 100644
--- a/kernel/events/uprobes.c
+++ b/kernel/events/uprobes.c
@@ -867,10 +867,6 @@ static int prepare_uprobe(struct uprobe *uprobe, struct file *file,
if (ret)
goto out;
- /* uprobe_write_opcode() assumes we don't cross page boundary */
- BUG_ON((uprobe->offset & ~PAGE_MASK) +
- UPROBE_SWBP_INSN_SIZE > PAGE_SIZE);
-
smp_wmb(); /* pairs with the smp_rmb() in handle_swbp() */
set_bit(UPROBE_COPY_INSN, &uprobe->flags);
@@ -1166,6 +1162,15 @@ static int __uprobe_register(struct inode *inode, loff_t offset,
if (offset > i_size_read(inode))
return -EINVAL;
+ /*
+ * This ensures that copy_from_page(), copy_to_page() and
+ * __update_ref_ctr() can't cross page boundary.
+ */
+ if (!IS_ALIGNED(offset, UPROBE_SWBP_INSN_SIZE))
+ return -EINVAL;
+ if (!IS_ALIGNED(ref_ctr_offset, sizeof(short)))
+ return -EINVAL;
+
retry:
uprobe = alloc_uprobe(inode, offset, ref_ctr_offset);
if (!uprobe)
@@ -2014,6 +2019,9 @@ static int is_trap_at_addr(struct mm_struct *mm, unsigned long vaddr)
uprobe_opcode_t opcode;
int result;
+ if (WARN_ON_ONCE(!IS_ALIGNED(vaddr, UPROBE_SWBP_INSN_SIZE)))
+ return -EINVAL;
+
pagefault_disable();
result = __get_user(opcode, (uprobe_opcode_t __user *)vaddr);
pagefault_enable();
--
2.25.1.362.g51ebf55
On 04.05.20 18:47, Oleg Nesterov wrote:
> uprobe_write_opcode() must not cross page boundary; prepare_uprobe()
> relies on arch_uprobe_analyze_insn() which should validate "vaddr" but
> some architectures (csky, s390, and sparc) don't do this.
I think the idea was that the uprobe instruction is 2 bytes and instructions
are always aligned to 2 bytes on s390. (we can have 2,4 or 6 bytes).
>
> We can remove the BUG_ON() check in prepare_uprobe() and validate the
> offset early in __uprobe_register(). The new IS_ALIGNED() check matches
> the alignment check in arch_prepare_kprobe() on supported architectures,
> so I think that all insns must be aligned to UPROBE_SWBP_INSN_SIZE.
Not sure if it would have been possible to try to create a uprobe on an
odd address. If yes, then the new IS_ALIGNED check certainly makes this
better for s390, so the patch looks sane. Adding Vasily and Sven to double
check.
>
> Another problem is __update_ref_ctr() which was wrong from the very
> beginning, it can read/write outside of kmap'ed page unless "vaddr" is
> aligned to sizeof(short), __uprobe_register() should check this too.
>
> Cc: [email protected]
> Reported-by: Linus Torvalds <[email protected]>
> Suggested-by: Linus Torvalds <[email protected]>
> Signed-off-by: Oleg Nesterov <[email protected]>
> ---
> kernel/events/uprobes.c | 16 ++++++++++++----
> 1 file changed, 12 insertions(+), 4 deletions(-)
>
> diff --git a/kernel/events/uprobes.c b/kernel/events/uprobes.c
> index ece7e13f6e4a..cc2095607c74 100644
> --- a/kernel/events/uprobes.c
> +++ b/kernel/events/uprobes.c
> @@ -867,10 +867,6 @@ static int prepare_uprobe(struct uprobe *uprobe, struct file *file,
> if (ret)
> goto out;
>
> - /* uprobe_write_opcode() assumes we don't cross page boundary */
> - BUG_ON((uprobe->offset & ~PAGE_MASK) +
> - UPROBE_SWBP_INSN_SIZE > PAGE_SIZE);
> -
> smp_wmb(); /* pairs with the smp_rmb() in handle_swbp() */
> set_bit(UPROBE_COPY_INSN, &uprobe->flags);
>
> @@ -1166,6 +1162,15 @@ static int __uprobe_register(struct inode *inode, loff_t offset,
> if (offset > i_size_read(inode))
> return -EINVAL;
>
> + /*
> + * This ensures that copy_from_page(), copy_to_page() and
> + * __update_ref_ctr() can't cross page boundary.
> + */
> + if (!IS_ALIGNED(offset, UPROBE_SWBP_INSN_SIZE))
> + return -EINVAL;
> + if (!IS_ALIGNED(ref_ctr_offset, sizeof(short)))
> + return -EINVAL;
> +
> retry:
> uprobe = alloc_uprobe(inode, offset, ref_ctr_offset);
> if (!uprobe)
> @@ -2014,6 +2019,9 @@ static int is_trap_at_addr(struct mm_struct *mm, unsigned long vaddr)
> uprobe_opcode_t opcode;
> int result;
>
> + if (WARN_ON_ONCE(!IS_ALIGNED(vaddr, UPROBE_SWBP_INSN_SIZE)))
> + return -EINVAL;
> +
> pagefault_disable();
> result = __get_user(opcode, (uprobe_opcode_t __user *)vaddr);
> pagefault_enable();
>
Hi,
On Mon, May 04, 2020 at 08:40:44PM +0200, Christian Borntraeger wrote:
>
>
> On 04.05.20 18:47, Oleg Nesterov wrote:
> > uprobe_write_opcode() must not cross page boundary; prepare_uprobe()
> > relies on arch_uprobe_analyze_insn() which should validate "vaddr" but
> > some architectures (csky, s390, and sparc) don't do this.
>
> I think the idea was that the uprobe instruction is 2 bytes and instructions
> are always aligned to 2 bytes on s390. (we can have 2,4 or 6 bytes).
>
> >
> > We can remove the BUG_ON() check in prepare_uprobe() and validate the
> > offset early in __uprobe_register(). The new IS_ALIGNED() check matches
> > the alignment check in arch_prepare_kprobe() on supported architectures,
> > so I think that all insns must be aligned to UPROBE_SWBP_INSN_SIZE.
>
> Not sure if it would have been possible to try to create a uprobe on an
> odd address. If yes, then the new IS_ALIGNED check certainly makes this
> better for s390, so the patch looks sane. Adding Vasily and Sven to double
> check.
I did a quick test, and without this patch it is possible to place a uprobe
at an odd address. With the patch it fails with EINVAL, which is more
reasonable.
Regards
Sven
* Oleg Nesterov <[email protected]> [2020-05-04 18:47:25]:
> uprobe_write_opcode() must not cross page boundary; prepare_uprobe()
> relies on arch_uprobe_analyze_insn() which should validate "vaddr" but
> some architectures (csky, s390, and sparc) don't do this.
>
> We can remove the BUG_ON() check in prepare_uprobe() and validate the
> offset early in __uprobe_register(). The new IS_ALIGNED() check matches
> the alignment check in arch_prepare_kprobe() on supported architectures,
> so I think that all insns must be aligned to UPROBE_SWBP_INSN_SIZE.
>
> Another problem is __update_ref_ctr() which was wrong from the very
> beginning, it can read/write outside of kmap'ed page unless "vaddr" is
> aligned to sizeof(short), __uprobe_register() should check this too.
>
> Cc: [email protected]
> Reported-by: Linus Torvalds <[email protected]>
> Suggested-by: Linus Torvalds <[email protected]>
> Signed-off-by: Oleg Nesterov <[email protected]>
Thanks Oleg.
Looks good to me.
Reviewed-by: Srikar Dronamraju <[email protected]>
> ---
--
Thanks and Regards
Srikar Dronamraju
On Wed, 6 May 2020 10:59:55 +0530
Srikar Dronamraju <[email protected]> wrote:
> * Oleg Nesterov <[email protected]> [2020-05-04 18:47:25]:
>
> > uprobe_write_opcode() must not cross page boundary; prepare_uprobe()
> > relies on arch_uprobe_analyze_insn() which should validate "vaddr" but
> > some architectures (csky, s390, and sparc) don't do this.
> >
> > We can remove the BUG_ON() check in prepare_uprobe() and validate the
> > offset early in __uprobe_register(). The new IS_ALIGNED() check matches
> > the alignment check in arch_prepare_kprobe() on supported architectures,
> > so I think that all insns must be aligned to UPROBE_SWBP_INSN_SIZE.
> >
> > Another problem is __update_ref_ctr() which was wrong from the very
> > beginning, it can read/write outside of kmap'ed page unless "vaddr" is
> > aligned to sizeof(short), __uprobe_register() should check this too.
> >
> > Cc: [email protected]
> > Reported-by: Linus Torvalds <[email protected]>
> > Suggested-by: Linus Torvalds <[email protected]>
> > Signed-off-by: Oleg Nesterov <[email protected]>
>
> Thanks Oleg.
>
> Looks good to me.
>
> Reviewed-by: Srikar Dronamraju <[email protected]>
> > ---
>
Thanks Oleg, Srikar and Sven.
As this is in the kernel/events/ directory, I'm guessing it should be taken
through the tip tree?
-- Steve
On 05/06, Steven Rostedt wrote:
>
> As this is in the kernel/events/ directory, I'm guessing it should be taken
> through the tip tree?
this would be great, thanks Steven.
Oleg.
Looks like this patch was forgotten...
Should I resend it?
On 05/04, Oleg Nesterov wrote:
>
> uprobe_write_opcode() must not cross page boundary; prepare_uprobe()
> relies on arch_uprobe_analyze_insn() which should validate "vaddr" but
> some architectures (csky, s390, and sparc) don't do this.
>
> We can remove the BUG_ON() check in prepare_uprobe() and validate the
> offset early in __uprobe_register(). The new IS_ALIGNED() check matches
> the alignment check in arch_prepare_kprobe() on supported architectures,
> so I think that all insns must be aligned to UPROBE_SWBP_INSN_SIZE.
>
> Another problem is __update_ref_ctr() which was wrong from the very
> beginning, it can read/write outside of kmap'ed page unless "vaddr" is
> aligned to sizeof(short), __uprobe_register() should check this too.
>
> Cc: [email protected]
> Reported-by: Linus Torvalds <[email protected]>
> Suggested-by: Linus Torvalds <[email protected]>
> Signed-off-by: Oleg Nesterov <[email protected]>
> ---
> kernel/events/uprobes.c | 16 ++++++++++++----
> 1 file changed, 12 insertions(+), 4 deletions(-)
>
> diff --git a/kernel/events/uprobes.c b/kernel/events/uprobes.c
> index ece7e13f6e4a..cc2095607c74 100644
> --- a/kernel/events/uprobes.c
> +++ b/kernel/events/uprobes.c
> @@ -867,10 +867,6 @@ static int prepare_uprobe(struct uprobe *uprobe, struct file *file,
> if (ret)
> goto out;
>
> - /* uprobe_write_opcode() assumes we don't cross page boundary */
> - BUG_ON((uprobe->offset & ~PAGE_MASK) +
> - UPROBE_SWBP_INSN_SIZE > PAGE_SIZE);
> -
> smp_wmb(); /* pairs with the smp_rmb() in handle_swbp() */
> set_bit(UPROBE_COPY_INSN, &uprobe->flags);
>
> @@ -1166,6 +1162,15 @@ static int __uprobe_register(struct inode *inode, loff_t offset,
> if (offset > i_size_read(inode))
> return -EINVAL;
>
> + /*
> + * This ensures that copy_from_page(), copy_to_page() and
> + * __update_ref_ctr() can't cross page boundary.
> + */
> + if (!IS_ALIGNED(offset, UPROBE_SWBP_INSN_SIZE))
> + return -EINVAL;
> + if (!IS_ALIGNED(ref_ctr_offset, sizeof(short)))
> + return -EINVAL;
> +
> retry:
> uprobe = alloc_uprobe(inode, offset, ref_ctr_offset);
> if (!uprobe)
> @@ -2014,6 +2019,9 @@ static int is_trap_at_addr(struct mm_struct *mm, unsigned long vaddr)
> uprobe_opcode_t opcode;
> int result;
>
> + if (WARN_ON_ONCE(!IS_ALIGNED(vaddr, UPROBE_SWBP_INSN_SIZE)))
> + return -EINVAL;
> +
> pagefault_disable();
> result = __get_user(opcode, (uprobe_opcode_t __user *)vaddr);
> pagefault_enable();
> --
> 2.25.1.362.g51ebf55
>
On Tue, Jun 9, 2020 at 8:30 AM Oleg Nesterov <[email protected]> wrote:
>
> Looks like this patch was forgotten...
>
> Should I resend it?
I guess I'll just take it directly, since it was triggered by me
complaining anyway.
I had hoped it would go through the usual channels.
Linus
On Tue, 9 Jun 2020 09:48:45 -0700
Linus Torvalds <[email protected]> wrote:
> On Tue, Jun 9, 2020 at 8:30 AM Oleg Nesterov <[email protected]> wrote:
> >
> > Looks like this patch was forgotten...
> >
> > Should I resend it?
>
> I guess I'll just take it directly, since it was triggered by me
> complaining anyway.
>
> I had hoped it would go through the usual channels.
Perhaps there was confusion about which tree it was suppose to go
through :-/
-- Steve
Hi Oleg,
On Tue, May 5, 2020 at 12:47 AM Oleg Nesterov <[email protected]> wrote:
>
> uprobe_write_opcode() must not cross page boundary; prepare_uprobe()
> relies on arch_uprobe_analyze_insn() which should validate "vaddr" but
> some architectures (csky, s390, and sparc) don't do this.
>
> We can remove the BUG_ON() check in prepare_uprobe() and validate the
> offset early in __uprobe_register(). The new IS_ALIGNED() check matches
> the alignment check in arch_prepare_kprobe() on supported architectures,
> so I think that all insns must be aligned to UPROBE_SWBP_INSN_SIZE.
>
> Another problem is __update_ref_ctr() which was wrong from the very
> beginning, it can read/write outside of kmap'ed page unless "vaddr" is
> aligned to sizeof(short), __uprobe_register() should check this too.
>
> Cc: [email protected]
> Reported-by: Linus Torvalds <[email protected]>
> Suggested-by: Linus Torvalds <[email protected]>
> Signed-off-by: Oleg Nesterov <[email protected]>
> ---
> kernel/events/uprobes.c | 16 ++++++++++++----
> 1 file changed, 12 insertions(+), 4 deletions(-)
>
> diff --git a/kernel/events/uprobes.c b/kernel/events/uprobes.c
> index ece7e13f6e4a..cc2095607c74 100644
> --- a/kernel/events/uprobes.c
> +++ b/kernel/events/uprobes.c
> @@ -867,10 +867,6 @@ static int prepare_uprobe(struct uprobe *uprobe, struct file *file,
> if (ret)
> goto out;
>
> - /* uprobe_write_opcode() assumes we don't cross page boundary */
> - BUG_ON((uprobe->offset & ~PAGE_MASK) +
> - UPROBE_SWBP_INSN_SIZE > PAGE_SIZE);
> -
> smp_wmb(); /* pairs with the smp_rmb() in handle_swbp() */
> set_bit(UPROBE_COPY_INSN, &uprobe->flags);
>
> @@ -1166,6 +1162,15 @@ static int __uprobe_register(struct inode *inode, loff_t offset,
> if (offset > i_size_read(inode))
> return -EINVAL;
>
> + /*
> + * This ensures that copy_from_page(), copy_to_page() and
> + * __update_ref_ctr() can't cross page boundary.
> + */
> + if (!IS_ALIGNED(offset, UPROBE_SWBP_INSN_SIZE))
> + return -EINVAL;
> + if (!IS_ALIGNED(ref_ctr_offset, sizeof(short)))
> + return -EINVAL;
> +
> retry:
> uprobe = alloc_uprobe(inode, offset, ref_ctr_offset);
> if (!uprobe)
> @@ -2014,6 +2019,9 @@ static int is_trap_at_addr(struct mm_struct *mm, unsigned long vaddr)
> uprobe_opcode_t opcode;
> int result;
>
> + if (WARN_ON_ONCE(!IS_ALIGNED(vaddr, UPROBE_SWBP_INSN_SIZE)))
> + return -EINVAL;
> +
> pagefault_disable();
> result = __get_user(opcode, (uprobe_opcode_t __user *)vaddr);
> pagefault_enable();
> --
> 2.25.1.362.g51ebf55
>
>
Looks good to me, thx.
Reviewed-by: Guo Ren <[email protected]>
--
Best Regards
Guo Ren
ML: https://lore.kernel.org/linux-csky/
On Tue, May 5, 2020 at 2:41 AM Christian Borntraeger
<[email protected]> wrote:
>
>
>
> On 04.05.20 18:47, Oleg Nesterov wrote:
> > uprobe_write_opcode() must not cross page boundary; prepare_uprobe()
> > relies on arch_uprobe_analyze_insn() which should validate "vaddr" but
> > some architectures (csky, s390, and sparc) don't do this.
>
> I think the idea was that the uprobe instruction is 2 bytes and instructions
> are always aligned to 2 bytes on s390. (we can have 2,4 or 6 bytes).
Agree, csky has two length-types of instructions (2,4 bytes).
>
> >
> > We can remove the BUG_ON() check in prepare_uprobe() and validate the
> > offset early in __uprobe_register(). The new IS_ALIGNED() check matches
> > the alignment check in arch_prepare_kprobe() on supported architectures,
> > so I think that all insns must be aligned to UPROBE_SWBP_INSN_SIZE.
>
> Not sure if it would have been possible to try to create a uprobe on an
> odd address. If yes, then the new IS_ALIGNED check certainly makes this
> better for s390, so the patch looks sane. Adding Vasily and Sven to double
> check.
Also good to csky.
--
Best Regards
Guo Ren
ML: https://lore.kernel.org/linux-csky/