netdev - [PATCH net-next 04/11] net: ipa: separate memory validation from initialization

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <20210609223503.2649114-5-elder@linaro.org>
Date:   Wed,  9 Jun 2021 17:34:56 -0500
From:   Alex Elder <elder@...aro.org>
To:     davem@...emloft.net, kuba@...nel.org
Cc:     bjorn.andersson@...aro.org, evgreen@...omium.org,
        cpratapa@...eaurora.org, subashab@...eaurora.org, elder@...nel.org,
        netdev@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: [PATCH net-next 04/11] net: ipa: separate memory validation from initialization

Currently, memory regions are validated in the loop that initializes
them.  Instead, validate them separately.

Rename ipa_mem_valid() to be ipa_mem_valid_one().  Define a *new*
function named ipa_mem_valid() that performs validation of the array
of memory regions provided.  This function calls ipa_mem_valid_one()
for each region in turn.

Skip validation for any "empty" region descriptors, which have zero
size and are not preceded by any canary values.  Issue a warning for
such descriptors if the offset is non-zero.

Signed-off-by: Alex Elder <elder@...aro.org>
---
 drivers/net/ipa/ipa_mem.c | 40 +++++++++++++++++++++++++++++++--------
 1 file changed, 32 insertions(+), 8 deletions(-)

diff --git a/drivers/net/ipa/ipa_mem.c b/drivers/net/ipa/ipa_mem.c
index effaa745a4061..62e1b8280d982 100644
--- a/drivers/net/ipa/ipa_mem.c
+++ b/drivers/net/ipa/ipa_mem.c
@@ -99,7 +99,7 @@ int ipa_mem_setup(struct ipa *ipa)
 	return 0;
 }
 
-static bool ipa_mem_valid(struct ipa *ipa, const struct ipa_mem *mem)
+static bool ipa_mem_valid_one(struct ipa *ipa, const struct ipa_mem *mem)
 {
 	struct device *dev = &ipa->pdev->dev;
 	enum ipa_mem_id mem_id = mem->id;
@@ -126,6 +126,31 @@ static bool ipa_mem_valid(struct ipa *ipa, const struct ipa_mem *mem)
 	return false;
 }
 
+/* Verify each defined memory region is valid. */
+static bool ipa_mem_valid(struct ipa *ipa)
+{
+	struct device *dev = &ipa->pdev->dev;
+	enum ipa_mem_id mem_id;
+
+	for (mem_id = 0; mem_id < ipa->mem_count; mem_id++) {
+		const struct ipa_mem *mem = &ipa->mem[mem_id];
+
+		/* Defined regions have non-zero size and/or canary count */
+		if (mem->size || mem->canary_count) {
+			if (ipa_mem_valid_one(ipa, mem))
+				continue;
+			return false;
+		}
+
+		/* It's harmless, but warn if an offset is provided */
+		if (mem->offset)
+			dev_warn(dev, "empty region %u has non-zero offset\n",
+				 mem_id);
+	}
+
+	return true;
+}
+
 /**
  * ipa_mem_config() - Configure IPA shared memory
  * @ipa:	IPA pointer
@@ -167,19 +192,18 @@ int ipa_mem_config(struct ipa *ipa)
 	ipa->zero_virt = virt;
 	ipa->zero_size = IPA_MEM_MAX;
 
-	/* Verify each defined memory region is valid, and if indicated
-	 * for the region, write "canary" values in the space prior to
-	 * the region's base address.
+	/* Make sure all defined memory regions are valid */
+	if (!ipa_mem_valid(ipa))
+		goto err_dma_free;
+
+	/* For each region, write "canary" values in the space prior to
+	 * the region's base address if indicated.
 	 */
 	for (mem_id = 0; mem_id < ipa->mem_count; mem_id++) {
 		const struct ipa_mem *mem = &ipa->mem[mem_id];
 		u16 canary_count;
 		__le32 *canary;
 
-		/* Validate all regions (even undefined ones) */
-		if (!ipa_mem_valid(ipa, mem))
-			goto err_dma_free;
-
 		/* Skip over undefined regions */
 		if (!mem->offset && !mem->size)
 			continue;
-- 
2.27.0