um: Use the x86 checksum implementation on 32-bit

When UML is compiled under 32-bit x86, it uses its own copy of
checksum_32.S, which is terribly out-of-date and doesn't support
checksumming unaligned data.

This causes the new "checksum" KUnit test to fail:
./tools/testing/kunit/kunit.py run --kconfig_add CONFIG_64BIT=n --cross_compile i686-linux-gnu- checksum
KTAP version 1
# Subtest: checksum
1..3
# test_csum_fixed_random_inputs: ASSERTION FAILED at lib/checksum_kunit.c:243
Expected result == expec, but
result == 33316 (0x8224)
expec == 33488 (0x82d0)
not ok 1 test_csum_fixed_random_inputs
# test_csum_all_carry_inputs: ASSERTION FAILED at lib/checksum_kunit.c:267
Expected result == expec, but
result == 65280 (0xff00)
expec == 0 (0x0)
not ok 2 test_csum_all_carry_inputs
# test_csum_no_carry_inputs: ASSERTION FAILED at lib/checksum_kunit.c:306
Expected result == expec, but
result == 65531 (0xfffb)
expec == 0 (0x0)
not ok 3 test_csum_no_carry_inputs

Sharing the normal implementation in arch/x86/lib both fixes all of
these issues and means any further fixes only need to be done once.

x86_64 already seems to share the same implementation between UML and
"normal" x86.

Signed-off-by: David Gow <[e-mail]>
Signed-off-by: Richard Weinberger <[e-mail]>

arch/x86/um/Makefile

···
 
 ifeq ($(CONFIG_X86_32),y)
 
-obj-y += checksum_32.o syscalls_32.o
+obj-y += syscalls_32.o
 obj-$(CONFIG_ELF_CORE) += elfcore.o
 
 subarch-y = ../lib/string_32.o ../lib/atomic64_32.o ../lib/atomic64_cx8_32.o
 subarch-y += ../lib/cmpxchg8b_emu.o ../lib/atomic64_386_32.o
+subarch-y += ../lib/checksum_32.o
 subarch-y += ../kernel/sys_ia32.o
 
 else

arch/x86/um/checksum_32.S

···
-/* SPDX-License-Identifier: GPL-2.0-or-later */
-/*
- * INET		An implementation of the TCP/IP protocol suite for the LINUX
- *		operating system.  INET is implemented using the  BSD Socket
- *		interface as the means of communication with the user level.
- *
- *		IP/TCP/UDP checksumming routines
- *
- * Authors:	Jorge Cwik, <[e-mail]>
- *		Arnt Gulbrandsen, <[e-mail]>
- *		Tom May, <[e-mail]>
- *              Pentium Pro/II routines:
- *              Alexander Kjeldaas <[e-mail]>
- *              Finn Arne Gangstad <[e-mail]>
- *		Lots of code moved from tcp.c and ip.c; see those files
- *		for more names.
- *
- * Changes:     Ingo Molnar, converted csum_partial_copy() to 2.1 exception
- *			     handling.
- *		Andi Kleen,  add zeroing on error
- *                   converted to pure assembler
- */
-
-#include <asm/errno.h>
-#include <asm/asm.h>
-#include <asm/export.h>
-				
-/*
- * computes a partial checksum, e.g. for TCP/UDP fragments
- */
-
-/*	
-unsigned int csum_partial(const unsigned char * buff, int len, unsigned int sum)
- */
-		
-.text
-.align 4
-.globl csum_partial
-		
-#ifndef CONFIG_X86_USE_PPRO_CHECKSUM
-
-	  /*		
-	   * Experiments with Ethernet and SLIP connections show that buff
-	   * is aligned on either a 2-byte or 4-byte boundary.  We get at
-	   * least a twofold speedup on 486 and Pentium if it is 4-byte aligned.
-	   * Fortunately, it is easy to convert 2-byte alignment to 4-byte
-	   * alignment for the unrolled loop.
-	   */		
-csum_partial:
-	pushl %esi
-	pushl %ebx
-	movl 20(%esp),%eax	# Function arg: unsigned int sum
-	movl 16(%esp),%ecx	# Function arg: int len
-	movl 12(%esp),%esi	# Function arg: unsigned char *buff
-	testl $2, %esi		# Check alignment.
-	jz 2f			# Jump if alignment is ok.
-	subl $2, %ecx		# Alignment uses up two bytes.
-	jae 1f			# Jump if we had at least two bytes.
-	addl $2, %ecx		# ecx was < 2.  Deal with it.
-	jmp 4f
-1:	movw (%esi), %bx
-	addl $2, %esi
-	addw %bx, %ax
-	adcl $0, %eax
-2:
-	movl %ecx, %edx
-	shrl $5, %ecx
-	jz 2f
-	testl %esi, %esi
-1:	movl (%esi), %ebx
-	adcl %ebx, %eax
-	movl 4(%esi), %ebx
-	adcl %ebx, %eax
-	movl 8(%esi), %ebx
-	adcl %ebx, %eax
-	movl 12(%esi), %ebx
-	adcl %ebx, %eax
-	movl 16(%esi), %ebx
-	adcl %ebx, %eax
-	movl 20(%esi), %ebx
-	adcl %ebx, %eax
-	movl 24(%esi), %ebx
-	adcl %ebx, %eax
-	movl 28(%esi), %ebx
-	adcl %ebx, %eax
-	lea 32(%esi), %esi
-	dec %ecx
-	jne 1b
-	adcl $0, %eax
-2:	movl %edx, %ecx
-	andl $0x1c, %edx
-	je 4f
-	shrl $2, %edx		# This clears CF
-3:	adcl (%esi), %eax
-	lea 4(%esi), %esi
-	dec %edx
-	jne 3b
-	adcl $0, %eax
-4:	andl $3, %ecx
-	jz 7f
-	cmpl $2, %ecx
-	jb 5f
-	movw (%esi),%cx
-	leal 2(%esi),%esi
-	je 6f
-	shll $16,%ecx
-5:	movb (%esi),%cl
-6:	addl %ecx,%eax
-	adcl $0, %eax 
-7:	
-	popl %ebx
-	popl %esi
-	RET
-
-#else
-
-/* Version for PentiumII/PPro */
-
-csum_partial:
-	pushl %esi
-	pushl %ebx
-	movl 20(%esp),%eax	# Function arg: unsigned int sum
-	movl 16(%esp),%ecx	# Function arg: int len
-	movl 12(%esp),%esi	# Function arg:	const unsigned char *buf
-
-	testl $2, %esi         
-	jnz 30f                 
-10:
-	movl %ecx, %edx
-	movl %ecx, %ebx
-	andl $0x7c, %ebx
-	shrl $7, %ecx
-	addl %ebx,%esi
-	shrl $2, %ebx  
-	negl %ebx
-	lea 45f(%ebx,%ebx,2), %ebx
-	testl %esi, %esi
-	jmp *%ebx
-
-	# Handle 2-byte-aligned regions
-20:	addw (%esi), %ax
-	lea 2(%esi), %esi
-	adcl $0, %eax
-	jmp 10b
-
-30:	subl $2, %ecx          
-	ja 20b                 
-	je 32f
-	movzbl (%esi),%ebx	# csumming 1 byte, 2-aligned
-	addl %ebx, %eax
-	adcl $0, %eax
-	jmp 80f
-32:
-	addw (%esi), %ax	# csumming 2 bytes, 2-aligned
-	adcl $0, %eax
-	jmp 80f
-
-40: 
-	addl -128(%esi), %eax
-	adcl -124(%esi), %eax
-	adcl -120(%esi), %eax
-	adcl -116(%esi), %eax   
-	adcl -112(%esi), %eax   
-	adcl -108(%esi), %eax
-	adcl -104(%esi), %eax
-	adcl -100(%esi), %eax
-	adcl -96(%esi), %eax
-	adcl -92(%esi), %eax
-	adcl -88(%esi), %eax
-	adcl -84(%esi), %eax
-	adcl -80(%esi), %eax
-	adcl -76(%esi), %eax
-	adcl -72(%esi), %eax
-	adcl -68(%esi), %eax
-	adcl -64(%esi), %eax     
-	adcl -60(%esi), %eax     
-	adcl -56(%esi), %eax     
-	adcl -52(%esi), %eax   
-	adcl -48(%esi), %eax   
-	adcl -44(%esi), %eax
-	adcl -40(%esi), %eax
-	adcl -36(%esi), %eax
-	adcl -32(%esi), %eax
-	adcl -28(%esi), %eax
-	adcl -24(%esi), %eax
-	adcl -20(%esi), %eax
-	adcl -16(%esi), %eax
-	adcl -12(%esi), %eax
-	adcl -8(%esi), %eax
-	adcl -4(%esi), %eax
-45:
-	lea 128(%esi), %esi
-	adcl $0, %eax
-	dec %ecx
-	jge 40b
-	movl %edx, %ecx
-50:	andl $3, %ecx
-	jz 80f
-
-	# Handle the last 1-3 bytes without jumping
-	notl %ecx		# 1->2, 2->1, 3->0, higher bits are masked
-	movl $0xffffff,%ebx	# by the shll and shrl instructions
-	shll $3,%ecx
-	shrl %cl,%ebx
-	andl -128(%esi),%ebx	# esi is 4-aligned so should be ok
-	addl %ebx,%eax
-	adcl $0,%eax
-80: 
-	popl %ebx
-	popl %esi
-	RET
-				
-#endif
-	EXPORT_SYMBOL(csum_partial)