Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:18 as permitted sender) client-ip=2620:137:e000::1:18;
From:   Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To:     linux-kernel@vger.kernel.org
Cc:     Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        stable@vger.kernel.org, Helge Deller <deller@gmx.de>
Subject: [PATCH 5.4 04/53] parisc/unaligned: Fix fldd and fstd unaligned handlers on 32-bit kernel
Date:   Mon, 28 Feb 2022 18:24:02 +0100
Message-Id: <20220228172248.619102510@linuxfoundation.org>
In-Reply-To: <20220228172248.232273337@linuxfoundation.org>
References: <20220228172248.232273337@linuxfoundation.org>
User-Agent: quilt/0.66
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Precedence: bulk

From: Helge Deller <deller@gmx.de>

commit dd2288f4a020d693360e3e8d72f8b9d9c25f5ef6 upstream.

Usually the kernel provides fixup routines to emulate the fldd and fstd
floating-point instructions if they load or store 8-byte from/to a not
natuarally aligned memory location.

On a 32-bit kernel I noticed that those unaligned handlers didn't worked and
instead the application got a SEGV.
While checking the code I found two problems:

First, the OPCODE_FLDD_L and OPCODE_FSTD_L cases were ifdef'ed out by the
CONFIG_PA20 option, and as such those weren't built on a pure 32-bit kernel.
This is now fixed by moving the CONFIG_PA20 #ifdef to prevent the compilation
of OPCODE_LDD_L and OPCODE_FSTD_L only, and handling the fldd and fstd
instructions.

The second problem are two bugs in the 32-bit inline assembly code, where the
wrong registers where used. The calculation of the natural alignment used %2
(vall) instead of %3 (ior), and the first word was stored back to address %1
(valh) instead of %3 (ior).

Signed-off-by: Helge Deller <deller@gmx.de>
Cc: stable@vger.kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
 arch/parisc/kernel/unaligned.c |    8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

--- a/arch/parisc/kernel/unaligned.c
+++ b/arch/parisc/kernel/unaligned.c
@@ -397,7 +397,7 @@ static int emulate_std(struct pt_regs *r
 	__asm__ __volatile__ (
 "	mtsp	%4, %%sr1\n"
 "	zdep	%2, 29, 2, %%r19\n"
-"	dep	%%r0, 31, 2, %2\n"
+"	dep	%%r0, 31, 2, %3\n"
 "	mtsar	%%r19\n"
 "	zvdepi	-2, 32, %%r19\n"
 "1:	ldw	0(%%sr1,%3),%%r20\n"
@@ -409,7 +409,7 @@ static int emulate_std(struct pt_regs *r
 "	andcm	%%r21, %%r19, %%r21\n"
 "	or	%1, %%r20, %1\n"
 "	or	%2, %%r21, %2\n"
-"3:	stw	%1,0(%%sr1,%1)\n"
+"3:	stw	%1,0(%%sr1,%3)\n"
 "4:	stw	%%r1,4(%%sr1,%3)\n"
 "5:	stw	%2,8(%%sr1,%3)\n"
 "	copy	%%r0, %0\n"
@@ -596,7 +596,6 @@ void handle_unaligned(struct pt_regs *re
 		ret = ERR_NOTHANDLED;	/* "undefined", but lets kill them. */
 		break;
 	}
-#ifdef CONFIG_PA20
 	switch (regs->iir & OPCODE2_MASK)
 	{
 	case OPCODE_FLDD_L:
@@ -607,14 +606,15 @@ void handle_unaligned(struct pt_regs *re
 		flop=1;
 		ret = emulate_std(regs, R2(regs->iir),1);
 		break;
+#ifdef CONFIG_PA20
 	case OPCODE_LDD_L:
 		ret = emulate_ldd(regs, R2(regs->iir),0);
 		break;
 	case OPCODE_STD_L:
 		ret = emulate_std(regs, R2(regs->iir),0);
 		break;
-	}
 #endif
+	}
 	switch (regs->iir & OPCODE3_MASK)
 	{
 	case OPCODE_FLDW_L: