From: Michael Schmitz <schmitzmic@gmail.com>
To: linux-m68k@vger.kernel.org, geert@linux-m68k.org
Cc: gerg@linux-m68k.org, Michael Schmitz <schmitzmic@gmail.com>,
Finn Thain <fthain@linux-m68k.org>,
linux-m68k@lists.linux-m68k.org
Subject: [PATCH v3 1/2] m68k: Handle __generic_copy_to_user faults more carefully
Date: Sat, 27 Apr 2024 20:48:14 +1200 [thread overview]
Message-ID: <20240427084815.1449-2-schmitzmic@gmail.com> (raw)
In-Reply-To: <20240427084815.1449-1-schmitzmic@gmail.com>
As mentioned by Finn Thain in his patch to improve put_user
exception handling on 040, a similar problem exists on 030
processors.
A moves instruction that crosses a page boundary from a
mapped page into an unmapped one will cause a mid-instruction
bus error exception (frame format b), with the PC pointing
(usually) two instructions past the faulting movesl instruction.
Our exception handling in __generic_copy_to_user only covers
the instruction immediately following the faulting one. As
a result, fixup_exception in send_fault_sig does not detect
this case, and cause send_fault_sig to oops.
Extend the exception table to cover one additional instruction
beyond the moves[lwb] instructions.
Tested on 68030 (Atari Falcon 030) with transfers beginning
at one to six bytes offset from the end of a mapped page,
followed by further bytes on an unmapped page (testcase
derived from stress-ng sysbadaddr stressor by Finn Thain).
Tested on 68040 (Mac Quadra) and 68030 (Mac IIci) by Finn Thain.
A similar problem is present in __clear_user(); modify the
exception table for that function in the same way (tested by
Finn Thain).
Cc: Finn Thain <fthain@linux-m68k.org>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: linux-m68k@lists.linux-m68k.org
Tested-by: Finn Thain <fthain@linux-m68k.org>
Link: https://lore.kernel.org/all/e0f23460779e6d16e2633486ac4841790ef2aca0.1713176294.git.fthain@linux-m68k.org
Signed-off-by: Michael Schmitz <schmitzmic@gmail.com>
---
Changes from RFC v2:
Finn Thain:
- add missing extension table entries and final NOP in
__generic_copy_to_user faults in 040 tests
Michael Schmitz:
- add yet another exception table entry in __clear_user
Changes from RFC v1:
Michael Schmitz:
- use extended exception table instead of additional NOPs
---
arch/m68k/lib/uaccess.c | 56 ++++++++++++++++++++++++-----------------
1 file changed, 33 insertions(+), 23 deletions(-)
diff --git a/arch/m68k/lib/uaccess.c b/arch/m68k/lib/uaccess.c
index 86b6fed5151c..e3ed893047f8 100644
--- a/arch/m68k/lib/uaccess.c
+++ b/arch/m68k/lib/uaccess.c
@@ -62,35 +62,42 @@ unsigned long __generic_copy_to_user(void __user *to, const void *from,
asm volatile ("\n"
" tst.l %0\n"
- " jeq 4f\n"
+ " jeq 5f\n"
"1: move.l (%1)+,%3\n"
"2: "MOVES".l %3,(%2)+\n"
"3: subq.l #1,%0\n"
- " jne 1b\n"
- "4: btst #1,%5\n"
- " jeq 6f\n"
- " move.w (%1)+,%3\n"
- "5: "MOVES".w %3,(%2)+\n"
- "6: btst #0,%5\n"
+ "4: jne 1b\n"
+ "5: btst #1,%5\n"
" jeq 8f\n"
- " move.b (%1)+,%3\n"
- "7: "MOVES".b %3,(%2)+\n"
- "8:\n"
+ "6: move.w (%1)+,%3\n"
+ "7: "MOVES".w %3,(%2)+\n"
+ "8: btst #0,%5\n"
+ "9: jeq 13f\n"
+ "10: move.b (%1)+,%3\n"
+ "11: "MOVES".b %3,(%2)+\n"
+ "12: nop\n"
+ "13:\n"
" .section .fixup,\"ax\"\n"
" .even\n"
"20: lsl.l #2,%0\n"
"50: add.l %5,%0\n"
- " jra 8b\n"
+ " jra 13b\n"
" .previous\n"
"\n"
" .section __ex_table,\"a\"\n"
" .align 4\n"
+ " .long 1b,20b\n"
" .long 2b,20b\n"
" .long 3b,20b\n"
- " .long 5b,50b\n"
+ " .long 4b,20b\n"
+ " .long 5b,20b\n"
" .long 6b,50b\n"
" .long 7b,50b\n"
" .long 8b,50b\n"
+ " .long 9b,50b\n"
+ " .long 10b,50b\n"
+ " .long 11b,50b\n"
+ " .long 12b,50b\n"
" .previous"
: "=d" (res), "+a" (from), "+a" (to), "=&d" (tmp)
: "0" (n / 4), "d" (n & 3));
@@ -109,32 +116,35 @@ unsigned long __clear_user(void __user *to, unsigned long n)
asm volatile ("\n"
" tst.l %0\n"
- " jeq 3f\n"
+ " jeq 4f\n"
"1: "MOVES".l %2,(%1)+\n"
"2: subq.l #1,%0\n"
- " jne 1b\n"
- "3: btst #1,%4\n"
- " jeq 5f\n"
- "4: "MOVES".w %2,(%1)+\n"
- "5: btst #0,%4\n"
- " jeq 7f\n"
- "6: "MOVES".b %2,(%1)\n"
- "7:\n"
+ "3: jne 1b\n"
+ "4: btst #1,%4\n"
+ " jeq 6f\n"
+ "5: "MOVES".w %2,(%1)+\n"
+ "6: btst #0,%4\n"
+ "7: jeq 9f\n"
+ "8: "MOVES".b %2,(%1)\n"
+ "9:\n"
" .section .fixup,\"ax\"\n"
" .even\n"
"10: lsl.l #2,%0\n"
"40: add.l %4,%0\n"
- " jra 7b\n"
+ " jra 9b\n"
" .previous\n"
"\n"
" .section __ex_table,\"a\"\n"
" .align 4\n"
" .long 1b,10b\n"
" .long 2b,10b\n"
- " .long 4b,40b\n"
+ " .long 3b,10b\n"
+ " .long 4b,10b\n"
" .long 5b,40b\n"
" .long 6b,40b\n"
" .long 7b,40b\n"
+ " .long 8b,40b\n"
+ " .long 9b,40b\n"
" .previous"
: "=d" (res), "+a" (to)
: "d" (0), "0" (n / 4), "d" (n & 3));
--
2.17.1
next prev parent reply other threads:[~2024-04-27 8:48 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-04-27 8:48 [PATCH v3 0/2] m68k uaccess fault handling fixes Michael Schmitz
2024-04-27 8:48 ` Michael Schmitz [this message]
2024-04-27 9:34 ` [PATCH v3 1/2] m68k: Handle __generic_copy_to_user faults more carefully Finn Thain
2024-04-27 23:16 ` Michael Schmitz
2024-04-28 3:28 ` Finn Thain
2024-04-28 3:51 ` Michael Schmitz
2024-04-28 8:03 ` Finn Thain
2024-04-27 8:48 ` [PATCH v3 2/2] m68k: improve __constant_copy_to_user_asm() fault handling Michael Schmitz
2024-04-27 9:24 ` Finn Thain
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20240427084815.1449-2-schmitzmic@gmail.com \
--to=schmitzmic@gmail.com \
--cc=fthain@linux-m68k.org \
--cc=geert@linux-m68k.org \
--cc=gerg@linux-m68k.org \
--cc=linux-m68k@lists.linux-m68k.org \
--cc=linux-m68k@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox