| Message ID | 20250119020518.1962249-1-kuba@kernel.org (mailing list archive) |
|---|---|
| Headers | show |
| Series | net: ethtool: fixes for HDS threshold | expand |
On Sat, Jan 18, 2025 at 6:05 PM Jakub Kicinski <kuba@kernel.org> wrote: > > Quick follow up on the HDS threshold work, since the merge window > is upon us. > > Fix the bnxt implementation to apply the settings right away, > because we update the parameters _after_ configuring HW user > needed to reconfig the device twice to get the settings to stick. > > For this I took the liberty of moving the config to a separate > struct. This follows my original thinking for the queue API. > It should also fit more neatly into how many drivers which > support safe config update operate. Drivers can allocate > new objects using the "pending" struct. > > netdevsim: > > KTAP version 1 > 1..7 > ok 1 hds.get_hds > ok 2 hds.get_hds_thresh > ok 3 hds.set_hds_disable > ok 4 hds.set_hds_enable > ok 5 hds.set_hds_thresh_zero > ok 6 hds.set_hds_thresh_max > ok 7 hds.set_hds_thresh_gt > # Totals: pass:7 fail:0 xfail:0 xpass:0 skip:0 error:0 > > bnxt: > > KTAP version 1 > 1..7 > ok 1 hds.get_hds > ok 2 hds.get_hds_thresh > ok 3 hds.set_hds_disable # SKIP disabling of HDS not supported by the device > ok 4 hds.set_hds_enable > ok 5 hds.set_hds_thresh_zero > ok 6 hds.set_hds_thresh_max > ok 7 hds.set_hds_thresh_gt > # Totals: pass:6 fail:0 xfail:0 xpass:0 skip:1 error:0 > > v2: > - make sure we always manipulate cfg_pending under rtnl_lock > - split patch 2 into 2+3 for ease of review > v1: https://lore.kernel.org/20250117194815.1514410-1-kuba@kernel.org For the series, Reviewed-by: Michael Chan <michael.chan@broadcom.com>
Quick follow up on the HDS threshold work, since the merge window is upon us. Fix the bnxt implementation to apply the settings right away, because we update the parameters _after_ configuring HW user needed to reconfig the device twice to get the settings to stick. For this I took the liberty of moving the config to a separate struct. This follows my original thinking for the queue API. It should also fit more neatly into how many drivers which support safe config update operate. Drivers can allocate new objects using the "pending" struct. netdevsim: KTAP version 1 1..7 ok 1 hds.get_hds ok 2 hds.get_hds_thresh ok 3 hds.set_hds_disable ok 4 hds.set_hds_enable ok 5 hds.set_hds_thresh_zero ok 6 hds.set_hds_thresh_max ok 7 hds.set_hds_thresh_gt # Totals: pass:7 fail:0 xfail:0 xpass:0 skip:0 error:0 bnxt: KTAP version 1 1..7 ok 1 hds.get_hds ok 2 hds.get_hds_thresh ok 3 hds.set_hds_disable # SKIP disabling of HDS not supported by the device ok 4 hds.set_hds_enable ok 5 hds.set_hds_thresh_zero ok 6 hds.set_hds_thresh_max ok 7 hds.set_hds_thresh_gt # Totals: pass:6 fail:0 xfail:0 xpass:0 skip:1 error:0 v2: - make sure we always manipulate cfg_pending under rtnl_lock - split patch 2 into 2+3 for ease of review v1: https://lore.kernel.org/20250117194815.1514410-1-kuba@kernel.org Jakub Kicinski (7): net: move HDS config from ethtool state net: ethtool: store netdev in a temp variable in ethnl_default_set_doit() net: provide pending ring configuration in net_device eth: bnxt: apply hds_thrs settings correctly net: ethtool: populate the default HDS params in the core eth: bnxt: allocate enough buffer space to meet HDS threshold eth: bnxt: update header sizing defaults include/linux/ethtool.h | 4 --- include/linux/netdevice.h | 9 ++++++ include/net/netdev_queues.h | 10 +++++++ drivers/net/ethernet/broadcom/bnxt/bnxt.c | 16 ++++++---- .../net/ethernet/broadcom/bnxt/bnxt_ethtool.c | 4 +-- drivers/net/netdevsim/ethtool.c | 6 +--- drivers/net/netdevsim/netdev.c | 10 +++---- net/core/dev.c | 12 ++++++-- net/core/devmem.c | 4 +-- net/ethtool/netlink.c | 30 +++++++++++++++---- net/ethtool/rings.c | 16 ++++++---- 11 files changed, 84 insertions(+), 37 deletions(-)