diff options
| author | Rafał Miłecki <rafal@milecki.pl> | 2023-02-07 18:09:40 +0100 |
|---|---|---|
| committer | Rafał Miłecki <rafal@milecki.pl> | 2023-07-10 13:31:08 +0200 |
| commit | e17f9fd0e8a9d7fbe159495b142b3a890f69db3a (patch) | |
| tree | 89172e0da1bd2562c4a841f1f7c208d6dea76122 /target/linux/bcm47xx | |
| parent | 80a6b0a9173a1182081e184fd0f22c0996bd16b2 (diff) | |
| download | upstream-e17f9fd0e8a9d7fbe159495b142b3a890f69db3a.tar.gz upstream-e17f9fd0e8a9d7fbe159495b142b3a890f69db3a.tar.bz2 upstream-e17f9fd0e8a9d7fbe159495b142b3a890f69db3a.zip | |
bcm47xx: revert bgmac back to the old limited max frame size
Bumping max frame size has significantly affected network performance
and memory usage. It was done by upstream commit that first appeared in
the 5.7 release.
Allocating 512 (BGMAC_RX_RING_SLOTS) buffers, 10 k each, is clearly a
bad idea on 32 MiB devices. This commit fixes support for Linksys E1000
V2.1 which gives up after allocating ~346 such buffers running 5.15
kernel.
Ref: 230c9da963aa ("bcm53xx: revert bgmac back to the old limited max frame size")
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
(cherry picked from commit 4970dd027bced78f8a72aa2baed22c962d3b587b)
Diffstat (limited to 'target/linux/bcm47xx')
| -rw-r--r-- | target/linux/bcm47xx/patches-5.10/701-net-bgmac-connect-to-PHY-even-if-it-is-BGMAC_PHY_NOR.patch | 33 |
1 files changed, 33 insertions, 0 deletions
diff --git a/target/linux/bcm47xx/patches-5.10/701-net-bgmac-connect-to-PHY-even-if-it-is-BGMAC_PHY_NOR.patch b/target/linux/bcm47xx/patches-5.10/701-net-bgmac-connect-to-PHY-even-if-it-is-BGMAC_PHY_NOR.patch new file mode 100644 index 0000000000..2c2eb07b82 --- /dev/null +++ b/target/linux/bcm47xx/patches-5.10/701-net-bgmac-connect-to-PHY-even-if-it-is-BGMAC_PHY_NOR.patch @@ -0,0 +1,33 @@ +From: =?UTF-8?q?Rafa=C5=82=20Mi=C5=82ecki?= <rafal@milecki.pl> +Date: Fri, 10 Jun 2022 13:10:47 +0200 +Subject: [PATCH] bgmac: reduce max frame size to support just MTU 1500 +MIME-Version: 1.0 +Content-Type: text/plain; charset=UTF-8 +Content-Transfer-Encoding: 8bit + +bgmac allocates new replacement buffer before handling each received +frame. Allocating & DMA-preparing 9724 B each time consumes a lot of CPU +time. Ideally bgmac should just respect currently set MTU but it isn't +the case right now. For now just revert back to the old limited frame +size. + +This change bumps NAT masquarade speed by ~95%. + +Ref: 8c7da63978f1 ("bgmac: configure MTU and add support for frames beyond 8192 byte size") +Signed-off-by: Rafał Miłecki <rafal@milecki.pl> +--- + drivers/net/ethernet/broadcom/bgmac.h | 3 +-- + 1 file changed, 1 insertion(+), 2 deletions(-) + +--- a/drivers/net/ethernet/broadcom/bgmac.h ++++ b/drivers/net/ethernet/broadcom/bgmac.h +@@ -366,8 +366,7 @@ + #define BGMAC_RX_FRAME_OFFSET 30 /* There are 2 unused bytes between header and real data */ + #define BGMAC_RX_BUF_OFFSET (NET_SKB_PAD + NET_IP_ALIGN - \ + BGMAC_RX_FRAME_OFFSET) +-/* Jumbo frame size with FCS */ +-#define BGMAC_RX_MAX_FRAME_SIZE 9724 ++#define BGMAC_RX_MAX_FRAME_SIZE 1536 + #define BGMAC_RX_BUF_SIZE (BGMAC_RX_FRAME_OFFSET + BGMAC_RX_MAX_FRAME_SIZE) + #define BGMAC_RX_ALLOC_SIZE (SKB_DATA_ALIGN(BGMAC_RX_BUF_SIZE + BGMAC_RX_BUF_OFFSET) + \ + SKB_DATA_ALIGN(sizeof(struct skb_shared_info))) |