Small optimization

git-svn-id: https://scst.svn.sourceforge.net/svnroot/scst/trunk@559 d57e44dd-8a1f-0410-8b47-8ef2f437770f

diff --git a/scst/src/dev_handlers/scst_user.c b/scst/src/dev_handlers/scst_user.c
index a2b2d67..5a4f6ba 100644 (file)
--- a/scst/src/dev_handlers/scst_user.c
+++ b/scst/src/dev_handlers/scst_user.c
@@ -806,14 +806,24 @@ static void dev_user_flush_dcache(struct scst_user_cmd *ucmd)
 {
        struct scst_user_cmd *buf_ucmd = ucmd->buf_ucmd;
        unsigned long start = buf_ucmd->ubuff;
-       int i;
+       int i, bufflen = ucmd->cmd->bufflen;
 
        TRACE_ENTRY();
 
        if (start == 0)
                goto out;
 
-       for (i = 0; i < buf_ucmd->num_data_pages; i++) {
+       /*
+        * Possibly, flushing of all the pages from ucmd->cmd->sg can be
+        * faster, since it should be cache hot, while ucmd->buf_ucmd and
+        * buf_ucmd->data_pages are cache cold. But, from other side,
+        * sizeof(buf_ucmd->data_pages[0]) is considerably smaller, than
+        * sizeof(ucmd->cmd->sg[0]), so on big buffers going over 
+        * data_pages array can lead to less cache misses. So, real numbers are
+        * needed. ToDo.
+        */
+
+       for (i = 0; (bufflen > 0) && (i < buf_ucmd->num_data_pages); i++) {
                struct page *page;
                page = buf_ucmd->data_pages[i];
 #ifdef ARCH_HAS_FLUSH_ANON_PAGE
@@ -823,6 +833,7 @@ static void dev_user_flush_dcache(struct scst_user_cmd *ucmd)
 #endif
                flush_dcache_page(page);
                start += PAGE_SIZE;
+               bufflen -= PAGE_SIZE;
        }
 
 out: