target/linux/generic/backport-6.1/795-v6.6-10-eth-r8152-try-to-use-a-normal-budget.patch

   1 From cf74eb5a5bc867258e7d0b0d1c3c4a60e1e3de2f Mon Sep 17 00:00:00 2001
   2 From: Jakub Kicinski <kuba@kernel.org>
   3 Date: Mon, 14 Aug 2023 08:35:21 -0700
   4 Subject: [PATCH] eth: r8152: try to use a normal budget
   5
   6 Mario reports that loading r8152 on his system leads to a:
   7
   8   netif_napi_add_weight() called with weight 256
   9
  10 warning getting printed. We don't have any solid data
  11 on why such high budget was chosen, and it may cause
  12 stalls in processing other softirqs and rt threads.
  13 So try to switch back to the default (64) weight.
  14
  15 If this slows down someone's system we should investigate
  16 which part of stopping starting the NAPI poll in this
  17 driver are expensive.
  18
  19 Reported-by: Mario Limonciello <mario.limonciello@amd.com>
  20 Link: https://lore.kernel.org/all/0bfd445a-81f7-f702-08b0-bd5a72095e49@amd.com/
  21 Acked-by: Hayes Wang <hayeswang@realtek.com>
  22 Link: https://lore.kernel.org/r/20230814153521.2697982-1-kuba@kernel.org
  23 Signed-off-by: Jakub Kicinski <kuba@kernel.org>
  24 ---
  25  drivers/net/usb/r8152.c | 3 +--
  26  1 file changed, 1 insertion(+), 2 deletions(-)
  27
  28 --- a/drivers/net/usb/r8152.c
  29 +++ b/drivers/net/usb/r8152.c
  30 @@ -9784,8 +9784,7 @@ static int rtl8152_probe(struct usb_inte
  31
  32         usb_set_intfdata(intf, tp);
  33
  34 -       netif_napi_add_weight(netdev, &tp->napi, r8152_poll,
  35 -                             tp->support_2500full ? 256 : 64);
  36 +       netif_napi_add(netdev, &tp->napi, r8152_poll);
  37
  38         ret = register_netdev(netdev);
  39         if (ret != 0) {