On Wed, Sep 22, 2021 at 10:36:31AM +0300, Shai Malin wrote:
> If the HW device is during recovery, the HW resources will never return,
> hence we shouldn't wait for the CID (HW context ID) bitmaps to clear.
> This fix speeds up the error recovery flow.
>
> Changes since v1:
> - Fix race condition (thanks to Leon Romanovsky).
Please put changelog under "---", there is a little value for them in the
commit message.
>
> Fixes: 64515dc899df ("qed: Add infrastructure for error detection and recovery")
> Signed-off-by: Michal Kalderon <mkalderon@xxxxxxxxxxx>
> Signed-off-by: Ariel Elior <aelior@xxxxxxxxxxx>
> Signed-off-by: Shai Malin <smalin@xxxxxxxxxxx>
> ---
> drivers/net/ethernet/qlogic/qed/qed_iwarp.c | 8 ++++++++
> drivers/net/ethernet/qlogic/qed/qed_roce.c | 8 ++++++++
> 2 files changed, 16 insertions(+)
>
> diff --git a/drivers/net/ethernet/qlogic/qed/qed_iwarp.c b/drivers/net/ethernet/qlogic/qed/qed_iwarp.c
> index fc8b3e64f153..186d0048a9d1 100644
> --- a/drivers/net/ethernet/qlogic/qed/qed_iwarp.c
> +++ b/drivers/net/ethernet/qlogic/qed/qed_iwarp.c
> @@ -1297,6 +1297,14 @@ qed_iwarp_wait_cid_map_cleared(struct qed_hwfn *p_hwfn, struct qed_bmap *bmap)
> prev_weight = weight;
>
> while (weight) {
> + /* If the HW device is during recovery, all resources are
> + * immediately reset without receiving a per-cid indication
> + * from HW. In this case we don't expect the cid_map to be
> + * cleared.
> + */
> + if (p_hwfn->cdev->recov_in_prog)
> + return 0;
> +
> msleep(QED_IWARP_MAX_CID_CLEAN_TIME);
>
> weight = bitmap_weight(bmap->bitmap, bmap->max_count);
> diff --git a/drivers/net/ethernet/qlogic/qed/qed_roce.c b/drivers/net/ethernet/qlogic/qed/qed_roce.c
> index f16a157bb95a..cf5baa5e59bc 100644
> --- a/drivers/net/ethernet/qlogic/qed/qed_roce.c
> +++ b/drivers/net/ethernet/qlogic/qed/qed_roce.c
> @@ -77,6 +77,14 @@ void qed_roce_stop(struct qed_hwfn *p_hwfn)
> * Beyond the added delay we clear the bitmap anyway.
> */
> while (bitmap_weight(rcid_map->bitmap, rcid_map->max_count)) {
> + /* If the HW device is during recovery, all resources are
> + * immediately reset without receiving a per-cid indication
> + * from HW. In this case we don't expect the cid bitmap to be
> + * cleared.
> + */
> + if (p_hwfn->cdev->recov_in_prog)
> + return;
> +
> msleep(100);
> if (wait_count++ > 20) {
> DP_NOTICE(p_hwfn, "cid bitmap wait timed out\n");
> --
> 2.27.0
>
