This is the first of three patches which reduce the amount of
kmap/kunmap traffic on highmem machines.
The workload which was tested was RAM-only dbench. This is dominated
by copy_*_user() costs.
The three patches speed up my 4xPIII by 3%
The three patches speed up a 16P NUMA-Q by 100 to 150%
The first two patches (copy_strings and pagecache reads) speed up an
8-way by 15%. I expect that all three patches will speed up the 8-way
by 40%.
Some of the benefit is from reduced pressure on kmap_lock. Most of it
is from reducing the number of global TLB invalidations.
This patch fixes up copy_strings(). copy_strings does a huge amount of
kmapping. Martin Bligh has noted that across a kernel compile this
function is the second or third largest user of kmaps in the kernel.
The fix is pretty simple: just hang onto the previous kmap as we we go
around the loop. It reduces the number of kmappings from copy_strings
by a factor of 30.
This patch also applies to 2.4, and seems worthwhile there.
exec.c | 60 +++++++++++++++++++++++++++++++++++++++++++-----------------
1 files changed, 43 insertions, 17 deletions
--- 2.5.30/fs/exec.c~copy_strings-speedup Fri Aug 9 17:36:41 2002
+++ 2.5.30-akpm/fs/exec.c Fri Aug 9 17:36:41 2002
@@ -185,25 +185,39 @@ static int count(char ** argv, int max)
*/
int copy_strings(int argc,char ** argv, struct linux_binprm *bprm)
{
+ struct page *kmapped_page = NULL;
+ char *kaddr = NULL;
+ int ret;
+
while (argc-- > 0) {
char *str;
int len;
unsigned long pos;
- if (get_user(str, argv+argc) || !(len = strnlen_user(str, bprm->p)))
- return -EFAULT;
- if (bprm->p < len)
- return -E2BIG;
+ if (get_user(str, argv+argc) ||
+ !(len = strnlen_user(str, bprm->p))) {
+ ret = -EFAULT;
+ goto out;
+ }
+
+ if (bprm->p < len) {
+ ret = -E2BIG;
+ goto out;
+ }
bprm->p -= len;
/* XXX: add architecture specific overflow check here. */
-
pos = bprm->p;
+
+ /*
+ * The only sleeping function which we are allowed to call in
+ * this loop is copy_from_user(). Otherwise, copy_user_state
+ * could get trashed.
+ */
while (len > 0) {
- char *kaddr;
int i, new, err;
- struct page *page;
int offset, bytes_to_copy;
+ struct page *page;
offset = pos % PAGE_SIZE;
i = pos/PAGE_SIZE;
@@ -212,32 +226,44 @@ int copy_strings(int argc,char ** argv,
if (!page) {
page = alloc_page(GFP_HIGHUSER);
bprm->page[i] = page;
- if (!page)
- return -ENOMEM;
+ if (!page) {
+ ret = -ENOMEM;
+ goto out;
+ }
new = 1;
}
- kaddr = kmap(page);
+ if (page != kmapped_page) {
+ if (kmapped_page)
+ kunmap(kmapped_page);
+ kmapped_page = page;
+ kaddr = kmap(kmapped_page);
+ }
if (new && offset)
memset(kaddr, 0, offset);
bytes_to_copy = PAGE_SIZE - offset;
if (bytes_to_copy > len) {
bytes_to_copy = len;
if (new)
- memset(kaddr+offset+len, 0, PAGE_SIZE-offset-len);
+ memset(kaddr+offset+len, 0,
+ PAGE_SIZE-offset-len);
+ }
+ err = copy_from_user(kaddr+offset, str, bytes_to_copy);
+ if (err) {
+ ret = -EFAULT;
+ goto out;
}
- err = copy_from_user(kaddr + offset, str, bytes_to_copy);
- kunmap(page);
-
- if (err)
- return -EFAULT;
pos += bytes_to_copy;
str += bytes_to_copy;
len -= bytes_to_copy;
}
}
- return 0;
+ ret = 0;
+out:
+ if (kmapped_page)
+ kunmap(kmapped_page);
+ return ret;
}
/*
.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Thu Aug 15 2002 - 22:00:21 EST