dpdk: rework hugepage hints to use per-numa information

Previous integration of hugepage analysis only fetched data
from /proc/meminfo. However this proved to be often
deceiving mainly for providing only global information and
not taking into account different hugepage sizes (e.g. 1GB
hugepages) and different NUMA nodes.

Ticket: #6419

doc/userguide/capture-hardware/dpdk.rst

@@ -15,6 +15,57 @@ learn more about the basic setup for DPDK.
 The following sections contain examples of how to set up DPDK and Suricata for
 more obscure use-cases.
 
+Hugepage analysis
+-----------------
+
+Suricata can analyse utilized hugepages on the system. This can be particularly 
+beneficial when there's a potential overallocation of hugepages. 
+The hugepage analysis is designed to examine the hugepages in use and 
+provide recommendations on an adequate number of hugepages. This then ensures 
+Suricata operates optimally while leaving sufficient memory for other 
+applications on the system. The analysis works by comparing snapshots of the
+hugepages before and after Suricata is initialized. After the initialization,
+no more hugepages are allocated by Suricata.
+The hugepage analysis can be seen in the Perf log level and is printed out 
+during the Suricata start. It is only printed when Suricata detects some 
+disrepancies in the system related to hugepage allocation.
+
+It's recommended to perform this analysis from a "clean" state - 
+that is a state when all your hugepages are free. It is especially recommended 
+when no other hugepage-dependent applications are running on your system.
+This can be checked in one of two ways:
+
+.. code-block:: 
+
+  # global check
+  cat /proc/meminfo
+
+  HugePages_Total:    1024
+  HugePages_Free:     1024
+
+  # per-numa check depends on NUMA node ID, hugepage size, 
+  # and nr_hugepages/free_hugepages - e.g.:
+  cat /sys/devices/system/node/node0/hugepages/hugepages-2048kB/free_hugepages
+
+After the termination of Suricata and other hugepage-related applications, 
+if the count of free hugepages is not equal with the total number of hugepages, 
+it indicates some hugepages were not freed completely.
+This can be fixed by removing DPDK-related files from the hugepage-mounted 
+directory (filesystem). 
+It's important to exercise caution while removing hugepages, especially when 
+other hugepage-dependent applications are in operation, as this action will 
+disrupt their memory functionality.
+Removing the DPDK files from the hugepage directory can often be done as:
+
+.. code-block:: bash
+
+  sudo rm -rf /dev/hugepages/rtemap_*
+
+  # To check where hugepages are mounted:
+  dpdk-hugepages.py -s
+  # or 
+  mount | grep huge
+
 Bond interface
 --------------
 

src/Makefile.am

@@ -552,6 +552,7 @@ noinst_HEADERS = \
 	util-hash-string.h \
 	util-host-info.h \
 	util-host-os-info.h \
+	util-hugepages.h \
 	util-hyperscan.h \
 	util-ioctl.h \
 	util-ip.h \
@@ -1152,6 +1153,7 @@ libsuricata_c_a_SOURCES = \
 	util-hash-string.c \
 	util-host-info.c \
 	util-host-os-info.c \
+	util-hugepages.c \
 	util-hyperscan.c \
 	util-ioctl.c \
 	util-ip.c \

src/suricata.c

@@ -127,6 +127,7 @@
 #include "util-ebpf.h"
 #include "util-exception-policy.h"
 #include "util-host-os-info.h"
+#include "util-hugepages.h"
 #include "util-ioctl.h"
 #include "util-landlock.h"
 #include "util-luajit.h"
@@ -2968,6 +2969,7 @@ int SuricataMain(int argc, char **argv)
         goto out;
     }
 
+    SystemHugepageSnapshot *prerun_snap = SystemHugepageSnapshotCreate();
     SCSetStartTime(&suricata);
     RunModeDispatch(suricata.run_mode, suricata.runmode_custom_mode,
             suricata.capture_plugin_name, suricata.capture_plugin_args);
@@ -3026,7 +3028,11 @@ int SuricataMain(int argc, char **argv)
 
     PostRunStartedDetectSetup(&suricata);
 
-    DPDKEvaluateHugepages();
+    SystemHugepageSnapshot *postrun_snap = SystemHugepageSnapshotCreate();
+    if (run_mode == RUNMODE_DPDK) // only DPDK uses hpages at the moment
+        SystemHugepageEvaluateHugepages(prerun_snap, postrun_snap);
+    SystemHugepageSnapshotDestroy(prerun_snap);
+    SystemHugepageSnapshotDestroy(postrun_snap);
 
     SCPledge();
     SuricataMainLoop(&suricata);

src/util-dpdk.c

@@ -66,106 +66,7 @@ void DPDKFreeDevice(LiveDevice *ldev)
 #endif
 }
 
-static FILE *HugepagesMeminfoOpen(void)
-{
-    FILE *fp = fopen("/proc/meminfo", "r");
-    if (fp == NULL) {
-        SCLogInfo("Can't analyze hugepage usage: failed to open /proc/meminfo");
-    }
-    return fp;
-}
-
-static void HugepagesMeminfoClose(FILE *fp)
-{
-    if (fp) {
-        fclose(fp);
-    }
-}
-
-/**
- * Parsing values of meminfo
- *
- * \param fp Opened file pointer for reading of file /proc/meminfo at beginning
- * \param keyword Entry to look for e.g. "HugePages_Free:"
- * \return n Value of the entry
- * \return -1 On error
- *
- */
-static int32_t MemInfoParseValue(FILE *fp, const char *keyword)
-{
-    char path[256], value_str[64];
-    int32_t value = -1;
-
-    while (fscanf(fp, "%255s", path) != EOF) {
-        if (strcmp(path, keyword) == 0) {
-            if (fscanf(fp, "%63s", value_str) == EOF) {
-                SCLogDebug("%s: not followed by any number", keyword);
-                break;
-            }
-
-            if (StringParseInt32(&value, 10, 23, value_str) < 0) {
-                SCLogDebug("Failed to convert %s from /proc/meminfo", keyword);
-                value = -1;
-            }
-            break;
-        }
-    }
-    return value;
-}
-
-static void MemInfoEvaluateHugepages(FILE *fp)
-{
-    int32_t free_hugepages = MemInfoParseValue(fp, "HugePages_Free:");
-    if (free_hugepages < 0) {
-        SCLogInfo("HugePages_Free information not found in /proc/meminfo");
-        return;
-    }
-
-    rewind(fp);
-
-    int32_t total_hugepages = MemInfoParseValue(fp, "HugePages_Total:");
-    if (total_hugepages < 0) {
-        SCLogInfo("HugePages_Total information not found in /proc/meminfo");
-        return;
-    } else if (total_hugepages == 0) {
-        SCLogInfo("HugePages_Total equals to zero");
-        return;
-    }
-
-    float free_hugepages_ratio = (float)free_hugepages / (float)total_hugepages;
-    if (free_hugepages_ratio > 0.5) {
-        SCLogInfo("%" PRIu32 " of %" PRIu32
-                  " of hugepages are free - number of hugepages can be lowered to e.g. %.0lf",
-                free_hugepages, total_hugepages, ceil((total_hugepages - free_hugepages) * 1.15));
-    }
-}
-
-static void MemInfoWith(void (*callback)(FILE *))
-{
-    FILE *fp = HugepagesMeminfoOpen();
-    if (fp) {
-        callback(fp);
-        HugepagesMeminfoClose(fp);
-    }
-}
-
-void DPDKEvaluateHugepages(void)
-{
-    if (run_mode != RUNMODE_DPDK)
-        return;
-
-#ifdef HAVE_DPDK
-    if (rte_eal_has_hugepages() == 0) { // hugepages disabled
-        SCLogPerf("Hugepages not enabled - enabling hugepages can improve performance");
-        return;
-    }
-#endif
-
-    MemInfoWith(MemInfoEvaluateHugepages);
-}
-
 #ifdef HAVE_DPDK
-
 /**
  * Retrieves name of the port from port id
  * Not thread-safe

src/util-dpdk.h

@@ -121,7 +121,6 @@ void DPDKCleanupEAL(void);
 
 void DPDKCloseDevice(LiveDevice *ldev);
 void DPDKFreeDevice(LiveDevice *ldev);
-void DPDKEvaluateHugepages(void);
 
 #ifdef HAVE_DPDK
 const char *DPDKGetPortNameByPortID(uint16_t pid);