DPDK patches and discussions
 help / color / mirror / Atom feed
From: Ralf Hoffmann <ralf.hoffmann@allegro-packets.com>
To: sergio.gonzalez.monroy@intel.com
Cc: dev@dpdk.org
Subject: [dpdk-dev] [PATCH v5 1/1] eal/linux: change hugepage sorting to avoid overlapping memcpy
Date: Thu,  7 Jan 2016 15:54:02 +0100	[thread overview]
Message-ID: <1452178442-22154-1-git-send-email-ralf.hoffmann@allegro-packets.com> (raw)
In-Reply-To: <568E77F3.8070408@intel.com>

with only one hugepage or already sorted hugepage addresses, the sort
function called memcpy with same src and dst pointer. Debugging with
valgrind will issue a warning about overlapping area. This patch changes
the sort method to qsort to avoid this behavior. The separate sort
function is no longer necessary.

Signed-off-by: Ralf Hoffmann <ralf.hoffmann@allegro-packets.com>
Acked-by: Sergio Gonzalez Monroy <sergio.gonzalez.monroy@intel.com>
---
v5:
* set commit title to eal/linux

v4:
* add linebreak to qsort statement

v3:
* set commit title to eal/linux

v2:

* incorporate patch from http://dpdk.org/dev/patchwork/patch/2061/
  to use qsort instead of bubble sort,
  original patch by Jay Rolette <rolette@infiniteio.com>

 lib/librte_eal/linuxapp/eal/eal_memory.c | 61 ++++++++------------------------
 1 file changed, 15 insertions(+), 46 deletions(-)

diff --git a/lib/librte_eal/linuxapp/eal/eal_memory.c b/lib/librte_eal/linuxapp/eal/eal_memory.c
index 846fd31..c0e510b 100644
--- a/lib/librte_eal/linuxapp/eal/eal_memory.c
+++ b/lib/librte_eal/linuxapp/eal/eal_memory.c
@@ -701,54 +701,23 @@ error:
 	return -1;
 }
 
-/*
- * Sort the hugepg_tbl by physical address (lower addresses first on x86,
- * higher address first on powerpc). We use a slow algorithm, but we won't
- * have millions of pages, and this is only done at init time.
- */
 static int
-sort_by_physaddr(struct hugepage_file *hugepg_tbl, struct hugepage_info *hpi)
+cmp_physaddr(const void *a, const void *b)
 {
-	unsigned i, j;
-	int compare_idx;
-	uint64_t compare_addr;
-	struct hugepage_file tmp;
-
-	for (i = 0; i < hpi->num_pages[0]; i++) {
-		compare_addr = 0;
-		compare_idx = -1;
-
-		/*
-		 * browse all entries starting at 'i', and find the
-		 * entry with the smallest addr
-		 */
-		for (j=i; j< hpi->num_pages[0]; j++) {
-
-			if (compare_addr == 0 ||
-#ifdef RTE_ARCH_PPC_64
-				hugepg_tbl[j].physaddr > compare_addr) {
+#ifndef RTE_ARCH_PPC_64
+	const struct hugepage_file *p1 = (const struct hugepage_file *)a;
+	const struct hugepage_file *p2 = (const struct hugepage_file *)b;
 #else
-				hugepg_tbl[j].physaddr < compare_addr) {
+	/* PowerPC needs memory sorted in reverse order from x86 */
+	const struct hugepage_file *p1 = (const struct hugepage_file *)b;
+	const struct hugepage_file *p2 = (const struct hugepage_file *)a;
 #endif
-				compare_addr = hugepg_tbl[j].physaddr;
-				compare_idx = j;
-			}
-		}
-
-		/* should not happen */
-		if (compare_idx == -1) {
-			RTE_LOG(ERR, EAL, "%s(): error in physaddr sorting\n", __func__);
-			return -1;
-		}
-
-		/* swap the 2 entries in the table */
-		memcpy(&tmp, &hugepg_tbl[compare_idx],
-			sizeof(struct hugepage_file));
-		memcpy(&hugepg_tbl[compare_idx], &hugepg_tbl[i],
-			sizeof(struct hugepage_file));
-		memcpy(&hugepg_tbl[i], &tmp, sizeof(struct hugepage_file));
-	}
-	return 0;
+	if (p1->physaddr < p2->physaddr)
+		return -1;
+	else if (p1->physaddr > p2->physaddr)
+		return 1;
+	else
+		return 0;
 }
 
 /*
@@ -1195,8 +1164,8 @@ rte_eal_hugepage_init(void)
 			goto fail;
 		}
 
-		if (sort_by_physaddr(&tmp_hp[hp_offset], hpi) < 0)
-			goto fail;
+		qsort(&tmp_hp[hp_offset], hpi->num_pages[0],
+		      sizeof(struct hugepage_file), cmp_physaddr);
 
 #ifdef RTE_EAL_SINGLE_FILE_SEGMENTS
 		/* remap all hugepages into single file segments */
-- 
2.5.0

  reply	other threads:[~2016-01-07 14:54 UTC|newest]

Thread overview: 17+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-09-04 10:14 [dpdk-dev] [PATCH v1] " Ralf Hoffmann
2015-09-08 12:45 ` Gonzalez Monroy, Sergio
2015-09-08 13:29   ` Jay Rolette
2015-09-08 14:24     ` Gonzalez Monroy, Sergio
2016-01-05  9:25       ` [dpdk-dev] [PATCH v2 0/1] " Ralf Hoffmann
2016-01-05  9:25         ` [dpdk-dev] [PATCH v2 1/1] " Ralf Hoffmann
2016-01-05  9:37       ` [dpdk-dev] [PATCH v3 0/1] eal/linux: " Ralf Hoffmann
2016-01-05  9:37         ` [dpdk-dev] [PATCH v3 1/1] " Ralf Hoffmann
2016-01-07  9:33           ` Sergio Gonzalez Monroy
2016-01-07  9:51             ` Sergio Gonzalez Monroy
2016-01-07 10:38               ` Sergio Gonzalez Monroy
2016-01-07 13:58                 ` Ralf Hoffmann
2016-01-07 13:59               ` [dpdk-dev] [PATCH v4 1/1] " Ralf Hoffmann
2016-01-07 14:36                 ` Sergio Gonzalez Monroy
2016-01-07 14:54                   ` Ralf Hoffmann [this message]
2016-03-02 16:13                     ` [dpdk-dev] [PATCH v5 1/1] eal/linux: " Thomas Monjalon
2015-09-09  8:41   ` [dpdk-dev] [PATCH v1] " Ralf Hoffmann

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1452178442-22154-1-git-send-email-ralf.hoffmann@allegro-packets.com \
    --to=ralf.hoffmann@allegro-packets.com \
    --cc=dev@dpdk.org \
    --cc=sergio.gonzalez.monroy@intel.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).