mirror of
https://github.com/AuxXxilium/linux_dsm_epyc7002.git
synced 2024-11-26 04:40:55 +07:00
d02d6be5d5
As noted by Matthew Wilcox:
Kyle McMartin just tracked down a bug on parisc to a missing
"memory" clobber in the inline assembly implementation of
ip_fast_csum. The FRV, SH and Xtensa ports are also missing a
memory clobber, so I thought it would be polite to let you know.
The bug manifests as dropped network packets (obviously they have
the wrong checksum). It started appearing for parisc with GCC 4.3.
The GCC manual says:
If your assembler instructions access memory in an unpredictable
fashion, add `memory' to the list of clobbered registers. This
will cause GCC to not keep memory values cached in registers
across the assembler instruction and not optimize stores or loads
to that memory.
I see that FRV has a 400 byte memory output which may prevent this
problem from appearing, but SH and Xtensa have nothing to prevent
this bug. Hope this saves you a few days of debugging.
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
216 lines
5.3 KiB
C
216 lines
5.3 KiB
C
#ifndef __ASM_SH_CHECKSUM_H
|
|
#define __ASM_SH_CHECKSUM_H
|
|
|
|
/*
|
|
* This file is subject to the terms and conditions of the GNU General Public
|
|
* License. See the file "COPYING" in the main directory of this archive
|
|
* for more details.
|
|
*
|
|
* Copyright (C) 1999 by Kaz Kojima & Niibe Yutaka
|
|
*/
|
|
|
|
#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
|
|
*/
|
|
asmlinkage __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
|
|
*/
|
|
|
|
asmlinkage __wsum csum_partial_copy_generic(const void *src, void *dst,
|
|
int len, __wsum sum,
|
|
int *src_err_ptr, int *dst_err_ptr);
|
|
|
|
/*
|
|
* Note: when you get a NULL pointer exception here this means someone
|
|
* passed in an incorrect kernel address to one of these functions.
|
|
*
|
|
* If you use these functions directly please don't forget the
|
|
* access_ok().
|
|
*/
|
|
static inline
|
|
__wsum csum_partial_copy_nocheck(const void *src, void *dst,
|
|
int len, __wsum sum)
|
|
{
|
|
return csum_partial_copy_generic(src, dst, len, sum, NULL, NULL);
|
|
}
|
|
|
|
static inline
|
|
__wsum csum_partial_copy_from_user(const void __user *src, void *dst,
|
|
int len, __wsum sum, int *err_ptr)
|
|
{
|
|
return csum_partial_copy_generic((__force const void *)src, dst,
|
|
len, sum, err_ptr, NULL);
|
|
}
|
|
|
|
/*
|
|
* Fold a partial checksum
|
|
*/
|
|
|
|
static inline __sum16 csum_fold(__wsum sum)
|
|
{
|
|
unsigned int __dummy;
|
|
__asm__("swap.w %0, %1\n\t"
|
|
"extu.w %0, %0\n\t"
|
|
"extu.w %1, %1\n\t"
|
|
"add %1, %0\n\t"
|
|
"swap.w %0, %1\n\t"
|
|
"add %1, %0\n\t"
|
|
"not %0, %0\n\t"
|
|
: "=r" (sum), "=&r" (__dummy)
|
|
: "0" (sum)
|
|
: "t");
|
|
return (__force __sum16)sum;
|
|
}
|
|
|
|
/*
|
|
* This is a version of ip_compute_csum() optimized for IP headers,
|
|
* which always checksum on 4 octet boundaries.
|
|
*
|
|
* i386 version by Jorge Cwik <jorge@laser.satlink.net>, adapted
|
|
* for linux by * Arnt Gulbrandsen.
|
|
*/
|
|
static inline __sum16 ip_fast_csum(const void *iph, unsigned int ihl)
|
|
{
|
|
unsigned int sum, __dummy0, __dummy1;
|
|
|
|
__asm__ __volatile__(
|
|
"mov.l @%1+, %0\n\t"
|
|
"mov.l @%1+, %3\n\t"
|
|
"add #-2, %2\n\t"
|
|
"clrt\n\t"
|
|
"1:\t"
|
|
"addc %3, %0\n\t"
|
|
"movt %4\n\t"
|
|
"mov.l @%1+, %3\n\t"
|
|
"dt %2\n\t"
|
|
"bf/s 1b\n\t"
|
|
" cmp/eq #1, %4\n\t"
|
|
"addc %3, %0\n\t"
|
|
"addc %2, %0" /* Here %2 is 0, add carry-bit */
|
|
/* Since the input registers which are loaded with iph and ihl
|
|
are modified, we must also specify them as outputs, or gcc
|
|
will assume they contain their original values. */
|
|
: "=r" (sum), "=r" (iph), "=r" (ihl), "=&r" (__dummy0), "=&z" (__dummy1)
|
|
: "1" (iph), "2" (ihl)
|
|
: "t", "memory");
|
|
|
|
return csum_fold(sum);
|
|
}
|
|
|
|
static inline __wsum csum_tcpudp_nofold(__be32 saddr, __be32 daddr,
|
|
unsigned short len,
|
|
unsigned short proto,
|
|
__wsum sum)
|
|
{
|
|
#ifdef __LITTLE_ENDIAN__
|
|
unsigned long len_proto = (proto + len) << 8;
|
|
#else
|
|
unsigned long len_proto = proto + len;
|
|
#endif
|
|
__asm__("clrt\n\t"
|
|
"addc %0, %1\n\t"
|
|
"addc %2, %1\n\t"
|
|
"addc %3, %1\n\t"
|
|
"movt %0\n\t"
|
|
"add %1, %0"
|
|
: "=r" (sum), "=r" (len_proto)
|
|
: "r" (daddr), "r" (saddr), "1" (len_proto), "0" (sum)
|
|
: "t");
|
|
|
|
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
|
|
static inline __sum16 csum_ipv6_magic(const struct in6_addr *saddr,
|
|
const struct in6_addr *daddr,
|
|
__u32 len, unsigned short proto,
|
|
__wsum sum)
|
|
{
|
|
unsigned int __dummy;
|
|
__asm__("clrt\n\t"
|
|
"mov.l @(0,%2), %1\n\t"
|
|
"addc %1, %0\n\t"
|
|
"mov.l @(4,%2), %1\n\t"
|
|
"addc %1, %0\n\t"
|
|
"mov.l @(8,%2), %1\n\t"
|
|
"addc %1, %0\n\t"
|
|
"mov.l @(12,%2), %1\n\t"
|
|
"addc %1, %0\n\t"
|
|
"mov.l @(0,%3), %1\n\t"
|
|
"addc %1, %0\n\t"
|
|
"mov.l @(4,%3), %1\n\t"
|
|
"addc %1, %0\n\t"
|
|
"mov.l @(8,%3), %1\n\t"
|
|
"addc %1, %0\n\t"
|
|
"mov.l @(12,%3), %1\n\t"
|
|
"addc %1, %0\n\t"
|
|
"addc %4, %0\n\t"
|
|
"addc %5, %0\n\t"
|
|
"movt %1\n\t"
|
|
"add %1, %0\n"
|
|
: "=r" (sum), "=&r" (__dummy)
|
|
: "r" (saddr), "r" (daddr),
|
|
"r" (htonl(len)), "r" (htonl(proto)), "0" (sum)
|
|
: "t");
|
|
|
|
return csum_fold(sum);
|
|
}
|
|
|
|
/*
|
|
* Copy and checksum to user
|
|
*/
|
|
#define HAVE_CSUM_COPY_USER
|
|
static inline __wsum csum_and_copy_to_user(const void *src,
|
|
void __user *dst,
|
|
int len, __wsum sum,
|
|
int *err_ptr)
|
|
{
|
|
if (access_ok(VERIFY_WRITE, dst, len))
|
|
return csum_partial_copy_generic((__force const void *)src,
|
|
dst, len, sum, NULL, err_ptr);
|
|
|
|
if (len)
|
|
*err_ptr = -EFAULT;
|
|
|
|
return (__force __wsum)-1; /* invalid checksum */
|
|
}
|
|
#endif /* __ASM_SH_CHECKSUM_H */
|