Return-path: Received: from mail-bk0-f49.google.com ([209.85.214.49]:41216 "EHLO mail-bk0-f49.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1757250AbaCDOGA convert rfc822-to-8bit (ORCPT ); Tue, 4 Mar 2014 09:06:00 -0500 Received: by mail-bk0-f49.google.com with SMTP id my13so223138bkb.22 for ; Tue, 04 Mar 2014 06:05:59 -0800 (PST) MIME-Version: 1.0 In-Reply-To: <1393937548-7482-14-git-send-email-janusz.dziedzic@tieto.com> References: <1393937548-7482-1-git-send-email-janusz.dziedzic@tieto.com> <1393937548-7482-14-git-send-email-janusz.dziedzic@tieto.com> Date: Tue, 4 Mar 2014 15:05:58 +0100 Message-ID: (sfid-20140304_150606_467838_124DE691) Subject: Re: [RFC 13/14] ath10k: return error when ath10k_htt_rx_amsdu_pop() fail From: Michal Kazior To: Janusz Dziedzic Cc: "ath10k@lists.infradead.org" , linux-wireless Content-Type: text/plain; charset=UTF-8 Sender: linux-wireless-owner@vger.kernel.org List-ID: On 4 March 2014 13:52, Janusz Dziedzic wrote: > Signed-off-by: Janusz Dziedzic > --- > drivers/net/wireless/ath/ath10k/htt_rx.c | 22 +++++++++++++--------- > 1 file changed, 13 insertions(+), 9 deletions(-) > > diff --git a/drivers/net/wireless/ath/ath10k/htt_rx.c b/drivers/net/wireless/ath/ath10k/htt_rx.c > index 555aecf..e300a74 100644 > --- a/drivers/net/wireless/ath/ath10k/htt_rx.c > +++ b/drivers/net/wireless/ath/ath10k/htt_rx.c > @@ -310,7 +310,7 @@ static int ath10k_htt_rx_amsdu_pop(struct ath10k_htt *htt, > > if (htt->rx_confused) { > ath10k_warn("htt is confused. refusing rx\n"); > - return 0; > + return -1; > } Add a comment before function definition, please: /* return: <0 fatal error, >=0 success, >0 number of chained msdus */ (this includes return-value logic Ben's unchaining patch added..) > > msdu = *head_msdu = ath10k_htt_rx_netbuf_pop(htt); > @@ -442,6 +442,9 @@ static int ath10k_htt_rx_amsdu_pop(struct ath10k_htt *htt, > } > *tail_msdu = msdu; > > + if (*head_msdu == NULL) > + msdu_chaining = -1; > + > /* > * Don't refill the ring yet. > * > @@ -1089,11 +1092,6 @@ static bool ath10k_htt_rx_amsdu_allowed(struct ath10k_htt *htt, > enum htt_rx_mpdu_status status, > bool channel_set) > { > - if (!head) { > - ath10k_warn("htt rx no data!\n"); > - return false; > - } > - > if (head->len == 0) { > ath10k_dbg(ATH10K_DBG_HTT, > "htt rx dropping due to zero-len\n"); > @@ -1209,8 +1207,14 @@ static void ath10k_htt_rx_handler(struct ath10k_htt *htt, > &msdu_head, > &msdu_tail); > > + if (msdu_chaining == -1) { if (msdu_chaining < 0) I think having `ret` makes more sense here now. > + ath10k_warn("htt rx no data!\n"); "failed to pop amsdu from htt rx ring: %d", ret > + ath10k_htt_rx_free_msdu_chain(msdu_head); > + continue; > + } > + > if (!ath10k_htt_rx_amsdu_allowed(htt, msdu_head, > - !!msdu_chaining, > + msdu_chaining > 0, > status, > channel_set)) { > ath10k_htt_rx_free_msdu_chain(msdu_head); > @@ -1280,12 +1284,12 @@ static void ath10k_htt_rx_frag_handler(struct ath10k_htt *htt, > > ath10k_dbg(ATH10K_DBG_HTT_DUMP, "htt rx frag ahead\n"); > > - if (!msdu_head) { > + if (msdu_chaining == -1) { ret < 0 > ath10k_warn("htt rx frag no data\n"); "failed to .. : %d", ret > return; > } > > - if (msdu_chaining || msdu_head != msdu_tail) { > + if (msdu_chaining > 0 || msdu_head != msdu_tail) { > ath10k_warn("aggregation with fragmentation?!\n"); > ath10k_htt_rx_free_msdu_chain(msdu_head); > return; Actually the 2 checks above can be squashed: if (ret) { ath10k_warn("failed to pop amsdu from htt rx for fragmented rx: %d", ret) Michał