Skip to contentSkip to conversation

rhashtable: make test actually random

PATCH RESEND Applied Rolf Eike Beer 4 participants 5 messages Oct 21, 2022 – Oct 26, 2022

Subsystems (4)
lore
Share & tools
lore
#0 lore raw

PATCH RESEND 1/1 rhashtable: make test actually random

The "random rhlist add/delete operations" actually wasn't very random, as all cases tested the same bit. Since the later parts of this loop depend on the first case execute this unconditionally, and then test on different bits for the remaining tests. While at it only request as much random bits as are actually used.

---
 lib/test_rhashtable.c | 58 ++++++++++++++++---------------------------
 1 file changed, 22 insertions(+), 36 deletions(-)

[Resend with correct from address]

lib/test_rhashtable.c +22 -36
index b358a74ed7ed..f2ba5787055a 100644
@@ -369,18 +369,10 @@ static int __init test_rhltable(unsigned int entries)
 	pr_info("test %d random rhlist add/delete operations\n", entries);
 	for (j = 0; j < entries; j++) {
 		u32 i = prandom_u32_max(entries);
-		u32 prand = get_random_u32();
+		u32 prand = prandom_u32_max(4);
 
 		cond_resched();
 
-		if (prand == 0)
-			prand = get_random_u32();
-
-		if (prand & 1) {
-			prand >>= 1;
-			continue;
-		}
-
 		err = rhltable_remove(&rhlt, &rhl_test_objects[i].list_node, test_rht_params);
 		if (test_bit(i, obj_in_table)) {
 			clear_bit(i, obj_in_table);
@@ -393,35 +385,29 @@ static int __init test_rhltable(unsigned int entries)
 		}
 
 		if (prand & 1) {
-			prand >>= 1;
-			continue;
-		}
-
-		err = rhltable_insert(&rhlt, &rhl_test_objects[i].list_node, test_rht_params);
-		if (err == 0) {
-			if (WARN(test_and_set_bit(i, obj_in_table), "succeeded to insert same object %d", i))
-				continue;
-		} else {
-			if (WARN(!test_bit(i, obj_in_table), "failed to insert object %d", i))
-				continue;
-		}
-
-		if (prand & 1) {
-			prand >>= 1;
-			continue;
+			err = rhltable_insert(&rhlt, &rhl_test_objects[i].list_node, test_rht_params);
+			if (err == 0) {
+				if (WARN(test_and_set_bit(i, obj_in_table), "succeeded to insert same object %d", i))
+					continue;
+			} else {
+				if (WARN(!test_bit(i, obj_in_table), "failed to insert object %d", i))
+					continue;
+			}
 		}
 
-		i = prandom_u32_max(entries);
-		if (test_bit(i, obj_in_table)) {
-			err = rhltable_remove(&rhlt, &rhl_test_objects[i].list_node, test_rht_params);
-			WARN(err, "cannot remove element at slot %d", i);
-			if (err == 0)
-				clear_bit(i, obj_in_table);
-		} else {
-			err = rhltable_insert(&rhlt, &rhl_test_objects[i].list_node, test_rht_params);
-			WARN(err, "failed to insert object %d", i);
-			if (err == 0)
-				set_bit(i, obj_in_table);
+		if (prand & 2) {
+			i = prandom_u32_max(entries);
+			if (test_bit(i, obj_in_table)) {
+				err = rhltable_remove(&rhlt, &rhl_test_objects[i].list_node, test_rht_params);
+				WARN(err, "cannot remove element at slot %d", i);
+				if (err == 0)
+					clear_bit(i, obj_in_table);
+			} else {
+				err = rhltable_insert(&rhlt, &rhl_test_objects[i].list_node, test_rht_params);
+				WARN(err, "failed to insert object %d", i);
+				if (err == 0)
+					set_bit(i, obj_in_table);
+			}
 		}
 	}
 
Signature · 1 line
2.35.3
#1 lore raw

Hi,

Quoted 6 lines · Rolf Eike Beer

The "random rhlist add/delete operations" actually wasn't very random, as all cases tested the same bit. Since the later parts of this loop depend on the first case execute this unconditionally, and then test on different bits for the remaining tests. While at it only request as much random bits as are actually used.

Seems reasonable to me. If it's okay with Thomas, who you CC'd, I'd like to take this through my random tree, as that'll prevent it from conflicting with a series I have out currently:
https://lore.kernel.org/lkml/[email protected]/

Jason

maintainer Acked-by on v1 1/1
#3 lore raw
In reply to Rolf Eike Beer
Quoted 10 lines · Rolf Eike Beer

The "random rhlist add/delete operations" actually wasn't very random, as all cases tested the same bit. Since the later parts of this loop depend on the first case execute this unconditionally, and then test on different bits for the remaining tests. While at it only request as much random bits as are actually used.

---
 lib/test_rhashtable.c | 58 ++++++++++++++++---------------------------
 1 file changed, 22 insertions(+), 36 deletions(-)
Signature · 3 lines
Email: Herbert Xu <[email protected]>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
maintainer
#2 lore raw
In reply to Jason A. Donenfeld
Quoted 5 lines · Jason A. Donenfeld

Seems reasonable to me. If it's okay with Thomas, who you CC'd, I'd like to take this through my random tree, as that'll prevent it from conflicting with a series I have out currently:
https://lore.kernel.org/lkml/[email protected]/

Sure, this code rarely changes so it should be fine to go through your tree.

Thanks,

Signature · 3 lines
Email: Herbert Xu <[email protected]>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
#4 lore raw
In reply to Rolf Eike Beer
This patch was applied to netdev/net.git (master) by David S. Miller <[email protected]>:

Hello:

This patch was applied to netdev/net.git (master)
by David S. Miller <[email protected]>:

Quoted 9 lines · On Fri, 21 Oct 2022 15:47:03 +0200 you wrote:

The "random rhlist add/delete operations" actually wasn't very random, as all cases tested the same bit. Since the later parts of this loop depend on the first case execute this unconditionally, and then test on different bits for the remaining tests. While at it only request as much random bits as are actually used.

[...]

Here is the summary with links:
- [Resend] rhashtable: make test actually random

    https://git.kernel.org/netdev/net/c/c5f0a1728874

You are awesome, thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html