mirror of
https://github.com/AuxXxilium/linux_dsm_epyc7002.git
synced 2024-12-21 12:26:47 +07:00
d46cda12e8
There is a miscompilation of csum_tcpudp_magic() due to the way we pass the asm() operands in. Fortunately, this doesn't affect the IP code, but can affect anyone who passes ntohs(udp->len) as the length argument, or protocols with more than 8 bits. The problem stems from passing 16-bit operands into an asm() - GCC makes no guarantees about what may be in the high 16-bits of such a register passed into assembly which is in the "HI" machine mode. Address this by changing the way we handle the 16-bit arguments - since accumulating the protocol and length can never overflow, we can delegate this to the compiler to perform, and then accumulate it into the checksum inside the asm(), taking account of the endian-ness via an appropriate 32-bit rotation. While we are here, also realise that there's a chance to optimise this a little: several callers from IP code pass a constant zero as the initial sum. This is wasteful - if we detect this condition, we can optimise away one instruction. Tested-by: Maxime Bizon <mbizon@freebox.fr> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
154 lines
3.7 KiB
C
154 lines
3.7 KiB
C
/*
|
|
* arch/arm/include/asm/checksum.h
|
|
*
|
|
* IP checksum routines
|
|
*
|
|
* Copyright (C) Original authors of ../asm-i386/checksum.h
|
|
* Copyright (C) 1996-1999 Russell King
|
|
*/
|
|
#ifndef __ASM_ARM_CHECKSUM_H
|
|
#define __ASM_ARM_CHECKSUM_H
|
|
|
|
#include <linux/in6.h>
|
|
|
|
/*
|
|
* computes the checksum of a memory block at buff, length len,
|
|
* and adds in "sum" (32-bit)
|
|
*
|
|
* returns a 32-bit number suitable for feeding into itself
|
|
* or csum_tcpudp_magic
|
|
*
|
|
* this function must be called with even lengths, except
|
|
* for the last fragment, which may be odd
|
|
*
|
|
* it's best to have buff aligned on a 32-bit boundary
|
|
*/
|
|
__wsum csum_partial(const void *buff, int len, __wsum sum);
|
|
|
|
/*
|
|
* the same as csum_partial, but copies from src while it
|
|
* checksums, and handles user-space pointer exceptions correctly, when needed.
|
|
*
|
|
* here even more important to align src and dst on a 32-bit (or even
|
|
* better 64-bit) boundary
|
|
*/
|
|
|
|
__wsum
|
|
csum_partial_copy_nocheck(const void *src, void *dst, int len, __wsum sum);
|
|
|
|
__wsum
|
|
csum_partial_copy_from_user(const void __user *src, void *dst, int len, __wsum sum, int *err_ptr);
|
|
|
|
/*
|
|
* Fold a partial checksum without adding pseudo headers
|
|
*/
|
|
static inline __sum16 csum_fold(__wsum sum)
|
|
{
|
|
__asm__(
|
|
"add %0, %1, %1, ror #16 @ csum_fold"
|
|
: "=r" (sum)
|
|
: "r" (sum)
|
|
: "cc");
|
|
return (__force __sum16)(~(__force u32)sum >> 16);
|
|
}
|
|
|
|
/*
|
|
* This is a version of ip_compute_csum() optimized for IP headers,
|
|
* which always checksum on 4 octet boundaries.
|
|
*/
|
|
static inline __sum16
|
|
ip_fast_csum(const void *iph, unsigned int ihl)
|
|
{
|
|
unsigned int tmp1;
|
|
__wsum sum;
|
|
|
|
__asm__ __volatile__(
|
|
"ldr %0, [%1], #4 @ ip_fast_csum \n\
|
|
ldr %3, [%1], #4 \n\
|
|
sub %2, %2, #5 \n\
|
|
adds %0, %0, %3 \n\
|
|
ldr %3, [%1], #4 \n\
|
|
adcs %0, %0, %3 \n\
|
|
ldr %3, [%1], #4 \n\
|
|
1: adcs %0, %0, %3 \n\
|
|
ldr %3, [%1], #4 \n\
|
|
tst %2, #15 @ do this carefully \n\
|
|
subne %2, %2, #1 @ without destroying \n\
|
|
bne 1b @ the carry flag \n\
|
|
adcs %0, %0, %3 \n\
|
|
adc %0, %0, #0"
|
|
: "=r" (sum), "=r" (iph), "=r" (ihl), "=r" (tmp1)
|
|
: "1" (iph), "2" (ihl)
|
|
: "cc", "memory");
|
|
return csum_fold(sum);
|
|
}
|
|
|
|
static inline __wsum
|
|
csum_tcpudp_nofold(__be32 saddr, __be32 daddr, unsigned short len,
|
|
unsigned short proto, __wsum sum)
|
|
{
|
|
u32 lenprot = len | proto << 16;
|
|
if (__builtin_constant_p(sum) && sum == 0) {
|
|
__asm__(
|
|
"adds %0, %1, %2 @ csum_tcpudp_nofold0 \n\t"
|
|
#ifdef __ARMEB__
|
|
"adcs %0, %0, %3 \n\t"
|
|
#else
|
|
"adcs %0, %0, %3, ror #8 \n\t"
|
|
#endif
|
|
"adc %0, %0, #0"
|
|
: "=&r" (sum)
|
|
: "r" (daddr), "r" (saddr), "r" (lenprot)
|
|
: "cc");
|
|
} else {
|
|
__asm__(
|
|
"adds %0, %1, %2 @ csum_tcpudp_nofold \n\t"
|
|
"adcs %0, %0, %3 \n\t"
|
|
#ifdef __ARMEB__
|
|
"adcs %0, %0, %4 \n\t"
|
|
#else
|
|
"adcs %0, %0, %4, ror #8 \n\t"
|
|
#endif
|
|
"adc %0, %0, #0"
|
|
: "=&r"(sum)
|
|
: "r" (sum), "r" (daddr), "r" (saddr), "r" (lenprot)
|
|
: "cc");
|
|
}
|
|
return sum;
|
|
}
|
|
/*
|
|
* computes the checksum of the TCP/UDP pseudo-header
|
|
* returns a 16-bit checksum, already complemented
|
|
*/
|
|
static inline __sum16
|
|
csum_tcpudp_magic(__be32 saddr, __be32 daddr, unsigned short len,
|
|
unsigned short proto, __wsum sum)
|
|
{
|
|
return csum_fold(csum_tcpudp_nofold(saddr, daddr, len, proto, sum));
|
|
}
|
|
|
|
|
|
/*
|
|
* this routine is used for miscellaneous IP-like checksums, mainly
|
|
* in icmp.c
|
|
*/
|
|
static inline __sum16
|
|
ip_compute_csum(const void *buff, int len)
|
|
{
|
|
return csum_fold(csum_partial(buff, len, 0));
|
|
}
|
|
|
|
#define _HAVE_ARCH_IPV6_CSUM
|
|
extern __wsum
|
|
__csum_ipv6_magic(const struct in6_addr *saddr, const struct in6_addr *daddr, __be32 len,
|
|
__be32 proto, __wsum sum);
|
|
|
|
static inline __sum16
|
|
csum_ipv6_magic(const struct in6_addr *saddr, const struct in6_addr *daddr, __u32 len,
|
|
unsigned short proto, __wsum sum)
|
|
{
|
|
return csum_fold(__csum_ipv6_magic(saddr, daddr, htonl(len),
|
|
htonl(proto), sum));
|
|
}
|
|
#endif
|