netdev - Re: [PATCH v4 06/17] octeontx2-pf: Receive packet handling support

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: <20200121083322.1d0b6e86@cakuba>
Date:   Tue, 21 Jan 2020 08:33:22 -0800
From:   Jakub Kicinski <kuba@...nel.org>
To:     sunil.kovvuri@...il.com
Cc:     netdev@...r.kernel.org, davem@...emloft.net, mkubecek@...e.cz,
        Sunil Goutham <sgoutham@...vell.com>
Subject: Re: [PATCH v4 06/17] octeontx2-pf: Receive packet handling support

On Tue, 21 Jan 2020 18:51:40 +0530, sunil.kovvuri@...il.com wrote:
>  static int otx2_rx_napi_handler(struct otx2_nic *pfvf,
>  				struct napi_struct *napi,
>  				struct otx2_cq_queue *cq, int budget)

> +	int processed_cqe = 0;
> +	s64 bufptr;
> +
> +	/* Make sure HW writes to CQ are done */
> +	dma_rmb();

What is this memory barrier between?

Usually dma_rmb() barrier is placed between accesses to the part of the
descriptor which tells us device is done and the rest of descriptor
accesses.

> +	while (likely(processed_cqe < budget)) {
> +		cqe = (struct nix_cqe_rx_s *)CQE_ADDR(cq, cq->cq_head);
> +		if (cqe->hdr.cqe_type == NIX_XQE_TYPE_INVALID ||
> +		    !cqe->sg.subdc) {
> +			if (!processed_cqe)
> +				return 0;
> +			break;
> +		}
> +		cq->cq_head++;
> +		cq->cq_head &= (cq->cqe_cnt - 1);
> +
> +		otx2_rcv_pkt_handler(pfvf, napi, cq, cqe);
> +
> +		cqe->hdr.cqe_type = NIX_XQE_TYPE_INVALID;
> +		cqe->sg.subdc = NIX_SUBDC_NOP;
> +		processed_cqe++;
> +	}

> +void otx2_cleanup_rx_cqes(struct otx2_nic *pfvf, struct otx2_cq_queue *cq)
> +{
> +	struct nix_cqe_rx_s *cqe;
> +	int processed_cqe = 0;
> +	u64 iova, pa;
> +
> +	/* Make sure HW writes to CQ are done */
> +	dma_rmb();

ditto

> +	while ((cqe = (struct nix_cqe_rx_s *)otx2_get_next_cqe(cq))) {
> +		if (!cqe->sg.subdc)
> +			continue;
> +		iova = cqe->sg.seg_addr - OTX2_HEAD_ROOM;
> +		pa = otx2_iova_to_phys(pfvf->iommu_domain, iova);
> +		otx2_dma_unmap_page(pfvf, iova, pfvf->rbsize, DMA_FROM_DEVICE);
> +		put_page(virt_to_page(phys_to_virt(pa)));
> +		processed_cqe++;
> +	}
> +
> +	/* Free CQEs to HW */
> +	otx2_write64(pfvf, NIX_LF_CQ_OP_DOOR,
> +		     ((u64)cq->cq_idx << 32) | processed_cqe);
> +}