Skip to contentSkip to conversation

powerpc/fault: kernel can extend a user process's stack

PATCH In review Daniel Axtens 4 participants 7 messages Dec 11, 2019 – Jul 21, 2020 PowerPC

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 powerpc/fault: kernel can extend a user process's stack

If a process page-faults trying to write beyond the end of its stack, we attempt to grow the stack.

However, if the kernel attempts to write beyond the end of a process's stack, it takes a bad fault. This can occur when the kernel is trying to set up a signal frame.

Permit the kernel to grow a process's stack. The same general limits as to how and when the stack can be grown apply: the kernel code still passes through expand_stack(), so anything that goes beyond e.g. the rlimit should still be blocked.

---
 arch/powerpc/mm/fault.c | 10 ++++++++++
 1 file changed, 10 insertions(+)
arch/powerpc/mm/fault.c +10 -0
index b5047f9b5dec..00183731ea22 100644
@@ -287,7 +287,17 @@ static bool bad_stack_expansion(struct pt_regs *regs, unsigned long address,
 			if (!res)
 				return !store_updates_sp(inst);
 			*must_retry = true;
+		} else if ((flags & FAULT_FLAG_WRITE) &&
+			   !(flags & FAULT_FLAG_USER)) {
+			/*
+			 * the kernel can also attempt to write beyond the end
+			 * of a process's stack - for example setting up a
+			 * signal frame. We assume this is valid, subject to
+			 * the checks in expand_stack() later.
+			 */
+			return false;
 		}
+
 		return true;
 	}
 	return false;
Signature · 1 line
2.20.1
#1 lore raw
Quoted 7 lines · Daniel Axtens

If a process page-faults trying to write beyond the end of its stack, we attempt to grow the stack.

However, if the kernel attempts to write beyond the end of a process's stack, it takes a bad fault. This can occur when the kernel is trying to set up a signal frame.

Quoted 6 lines

Permit the kernel to grow a process's stack. The same general limits as to how and when the stack can be grown apply: the kernel code still passes through expand_stack(), so anything that goes beyond e.g. the rlimit should still be blocked.

Fixes 14cf11af6cf6 ("powerpc: Merge enough to start building in

arch/powerpc.")

Quoted 32 lines
---
 arch/powerpc/mm/fault.c | 10 ++++++++++
 1 file changed, 10 insertions(+)
arch/powerpc/mm/fault.c +10 -0
index b5047f9b5dec..00183731ea22 100644
@@ -287,7 +287,17 @@ static bool bad_stack_expansion(struct pt_regs *regs, unsigned long address,
 			if (!res)
 				return !store_updates_sp(inst);
 			*must_retry = true;
+		} else if ((flags & FAULT_FLAG_WRITE) &&
+			   !(flags & FAULT_FLAG_USER)) {
+			/*
+			 * the kernel can also attempt to write beyond the end
+			 * of a process's stack - for example setting up a
+			 * signal frame. We assume this is valid, subject to
+			 * the checks in expand_stack() later.
+			 */
+			return false;
 		}
+
 		return true;
 	}
 	return false;
Signature · 1 line
2.20.1
#2 lore raw
Quoted 2 lines
Fixes 14cf11af6cf6 ("powerpc: Merge enough to start building in

arch/powerpc.")

Wow, that's pretty ancient! I'm also not sure it's right - in that same patch, arch/ppc64/mm/fault.c contains:

^1da177e4c3f4 (Linus Torvalds         2005-04-16 15:20:36 -0700 213)            if (address + 2048 < uregs->gpr[1]
^1da177e4c3f4 (Linus Torvalds         2005-04-16 15:20:36 -0700 214)                && (!user_mode(regs) || !store_updates_sp(regs)))
^1da177e4c3f4 (Linus Torvalds         2005-04-16 15:20:36 -0700 215)                    goto bad_area;

Which is the same as the new arch/powerpc/mm/fault.c code:

14cf11af6cf60 (Paul Mackerras 2005-09-26 16:04:21 +1000 234) if (address + 2048 < uregs->gpr[1]
14cf11af6cf60 (Paul Mackerras 2005-09-26 16:04:21 +1000 235) && (!user_mode(regs) || !store_updates_sp(regs)))
14cf11af6cf60 (Paul Mackerras 2005-09-26 16:04:21 +1000 236) goto bad_area;

So either they're both right or they're both wrong, either way I'm not sure how this patch is to blame.

I guess we should also cc stable@...

Regards,
Daniel

Quoted 32 lines
Quoted 32 lines
---
 arch/powerpc/mm/fault.c | 10 ++++++++++
 1 file changed, 10 insertions(+)
arch/powerpc/mm/fault.c +10 -0
index b5047f9b5dec..00183731ea22 100644
@@ -287,7 +287,17 @@ static bool bad_stack_expansion(struct pt_regs *regs, unsigned long address,
 			if (!res)
 				return !store_updates_sp(inst);
 			*must_retry = true;
+		} else if ((flags & FAULT_FLAG_WRITE) &&
+			   !(flags & FAULT_FLAG_USER)) {
+			/*
+			 * the kernel can also attempt to write beyond the end
+			 * of a process's stack - for example setting up a
+			 * signal frame. We assume this is valid, subject to
+			 * the checks in expand_stack() later.
+			 */
+			return false;
 		}
+
 		return true;
 	}
 	return false;
Signature · 1 line
2.20.1
#3 lore raw

Hello,

Quoted 10 lines · Daniel Axtens
Quoted 2 lines
Fixes 14cf11af6cf6 ("powerpc: Merge enough to start building in

arch/powerpc.")

Wow, that's pretty ancient! I'm also not sure it's right - in that same patch, arch/ppc64/mm/fault.c contains:

^1da177e4c3f4 (Linus Torvalds         2005-04-16 15:20:36 -0700 213)            if (address + 2048 < uregs->gpr[1]
^1da177e4c3f4 (Linus Torvalds         2005-04-16 15:20:36 -0700 214)                && (!user_mode(regs) || !store_updates_sp(regs)))
^1da177e4c3f4 (Linus Torvalds         2005-04-16 15:20:36 -0700 215)                    goto bad_area;
Quoted 8 lines

Which is the same as the new arch/powerpc/mm/fault.c code:

14cf11af6cf60 (Paul Mackerras 2005-09-26 16:04:21 +1000 234) if (address + 2048 < uregs->gpr[1]
14cf11af6cf60 (Paul Mackerras 2005-09-26 16:04:21 +1000 235) && (!user_mode(regs) || !store_updates_sp(regs)))
14cf11af6cf60 (Paul Mackerras 2005-09-26 16:04:21 +1000 236) goto bad_area;

So either they're both right or they're both wrong, either way I'm not sure how this patch is to blame.

Is there any progress on resolving this?

I did not notice any followup patch nor this one being merged/refuted.

Thanks

Michal

Quoted 38 lines

I guess we should also cc stable@...

Regards,
Daniel

Quoted 32 lines
Quoted 32 lines
---
 arch/powerpc/mm/fault.c | 10 ++++++++++
 1 file changed, 10 insertions(+)
arch/powerpc/mm/fault.c +10 -0
index b5047f9b5dec..00183731ea22 100644
@@ -287,7 +287,17 @@ static bool bad_stack_expansion(struct pt_regs *regs, unsigned long address,
 			if (!res)
 				return !store_updates_sp(inst);
 			*must_retry = true;
+		} else if ((flags & FAULT_FLAG_WRITE) &&
+			   !(flags & FAULT_FLAG_USER)) {
+			/*
+			 * the kernel can also attempt to write beyond the end
+			 * of a process's stack - for example setting up a
+			 * signal frame. We assume this is valid, subject to
+			 * the checks in expand_stack() later.
+			 */
+			return false;
 		}
+
 		return true;
 	}
 	return false;
Signature · 1 line
2.20.1
#4 lore raw
Quoted 19 lines · Michal Suchánek

Hello,

Quoted 16 lines · Daniel Axtens
Quoted 2 lines
Fixes 14cf11af6cf6 ("powerpc: Merge enough to start building in

arch/powerpc.")

Wow, that's pretty ancient! I'm also not sure it's right - in that same patch, arch/ppc64/mm/fault.c contains:

^1da177e4c3f4 (Linus Torvalds         2005-04-16 15:20:36 -0700 213)            if (address + 2048 < uregs->gpr[1]
^1da177e4c3f4 (Linus Torvalds         2005-04-16 15:20:36 -0700 214)                && (!user_mode(regs) || !store_updates_sp(regs)))
^1da177e4c3f4 (Linus Torvalds         2005-04-16 15:20:36 -0700 215)                    goto bad_area;

Which is the same as the new arch/powerpc/mm/fault.c code:

14cf11af6cf60 (Paul Mackerras 2005-09-26 16:04:21 +1000 234) if (address + 2048 < uregs->gpr[1]
14cf11af6cf60 (Paul Mackerras 2005-09-26 16:04:21 +1000 235) && (!user_mode(regs) || !store_updates_sp(regs)))
14cf11af6cf60 (Paul Mackerras 2005-09-26 16:04:21 +1000 236) goto bad_area;

Quoted 6 lines
Quoted 2 lines

So either they're both right or they're both wrong, either way I'm not sure how this patch is to blame.

Is there any progress on resolving this?

I did not notice any followup patch nor this one being merged/refuted.

It ended up with this:

https://lore.kernel.org/linuxppc-dev/[email protected]/

Which I was hoping would get some reviews :)

I'll probably merge the whole series into next this week.

cheers

#5 lore raw
Quoted 25 lines · Michael Ellerman
Quoted 24 lines · Michal Suchánek

Hello,

Quoted 10 lines · Daniel Axtens
Quoted 2 lines
Fixes 14cf11af6cf6 ("powerpc: Merge enough to start building in

arch/powerpc.")

Wow, that's pretty ancient! I'm also not sure it's right - in that same patch, arch/ppc64/mm/fault.c contains:

^1da177e4c3f4 (Linus Torvalds         2005-04-16 15:20:36 -0700 213)            if (address + 2048 < uregs->gpr[1]
^1da177e4c3f4 (Linus Torvalds         2005-04-16 15:20:36 -0700 214)                && (!user_mode(regs) || !store_updates_sp(regs)))
^1da177e4c3f4 (Linus Torvalds         2005-04-16 15:20:36 -0700 215)                    goto bad_area;
Quoted 8 lines

Which is the same as the new arch/powerpc/mm/fault.c code:

14cf11af6cf60 (Paul Mackerras 2005-09-26 16:04:21 +1000 234) if (address + 2048 < uregs->gpr[1]
14cf11af6cf60 (Paul Mackerras 2005-09-26 16:04:21 +1000 235) && (!user_mode(regs) || !store_updates_sp(regs)))
14cf11af6cf60 (Paul Mackerras 2005-09-26 16:04:21 +1000 236) goto bad_area;

So either they're both right or they're both wrong, either way I'm not sure how this patch is to blame.

Is there any progress on resolving this?

Quoted 8 lines
Quoted 1 line

I did not notice any followup patch nor this one being merged/refuted.

It ended up with this:

https://lore.kernel.org/linuxppc-dev/[email protected]/

Which I was hoping would get some reviews :)

Ah, I missed this. I'll give it a look as soon as I can.

Kind regards,
Daniel

Quoted 4 lines

I'll probably merge the whole series into next this week.

cheers

Tested-by on v1 1/1
#6 lore raw
In reply to Daniel Axtens
Quoted 11 lines · On Wed, 11 Dec 2019 12:43:37 +1100 Daniel Axtens <[email protected]> wrote:

If a process page-faults trying to write beyond the end of its stack, we attempt to grow the stack.

However, if the kernel attempts to write beyond the end of a process's stack, it takes a bad fault. This can occur when the kernel is trying to set up a signal frame.

Permit the kernel to grow a process's stack. The same general limits as to how and when the stack can be grown apply: the kernel code still passes through expand_stack(), so anything that goes beyond e.g. the rlimit should still be blocked.

Thanks Daniel.

Looks good from a function perspective.

danielgb@ozrom2:~$ gcc -g -Wall -O stactest.c 
danielgb@ozrom2:~$ ./a.out 1240000 &
[1] 4223
danielgb@ozrom2:~$ cat /proc/$(pidof a.out)/maps | grep stack
^[[3~7ffff14d0000-7ffff1600000 rw-p 00000000 00:00 0                          [stack]
danielgb@ozrom2:~$  kill -USR1 %1
danielgb@ozrom2:~$ signal delivered, stack base 0x7ffff1600000 top 0x7ffff14d1427 (1240025 used)

[1]+ Done ./a.out 1240000

danielgb@ozrom2:~$  ./a.out 1241000 &
[1] 4227
danielgb@ozrom2:~$ kill -USR1 %1
danielgb@ozrom2:~$ signal delivered, stack base 0x7fffff630000 top 0x7fffff501057 (1241001 used)

[1]+ Done ./a.out 1241000

danielgb@ozrom2:~$ uname -a
Linux ozrom2 5.5.0-rc1-00001-g83ab444248c1 #1 SMP Wed Dec 11 17:01:50 AEDT 2019 ppc64le ppc64le ppc64le GNU/Linux
Quoted 31 lines
---
 arch/powerpc/mm/fault.c | 10 ++++++++++
 1 file changed, 10 insertions(+)
arch/powerpc/mm/fault.c +10 -0
index b5047f9b5dec..00183731ea22 100644
@@ -287,7 +287,17 @@ static bool bad_stack_expansion(struct pt_regs *regs, unsigned long address,
 			if (!res)
 				return !store_updates_sp(inst);
 			*must_retry = true;
+		} else if ((flags & FAULT_FLAG_WRITE) &&
+			   !(flags & FAULT_FLAG_USER)) {
+			/*
+			 * the kernel can also attempt to write beyond the end
+			 * of a process's stack - for example setting up a
+			 * signal frame. We assume this is valid, subject to
+			 * the checks in expand_stack() later.
+			 */
+			return false;
 		}
+
 		return true;
 	}
 	return false;