Thread (3 messages) flat view 3 messages, 3 authors, 2021-05-13

Re: [PATCH 14/18] ath10k: drop MPDU which has discard flag set by firmware for SDIO

From: Jeff Johnson <hidden>
Date: 2021-05-13 17:19:12
Also in: linux-wireless

On 2021-05-12 11:35, Brian Norris wrote:
On Tue, May 11, 2021 at 11:03 AM Johannes Berg
[off-list ref] wrote:
quoted
--- a/drivers/net/wireless/ath/ath10k/htt_rx.c
+++ b/drivers/net/wireless/ath/ath10k/htt_rx.c
@@ -2312,6 +2312,11 @@ static bool ath10k_htt_rx_proc_rx_ind_hl(struct
ath10k_htt *htt,
        fw_desc = &rx->fw_desc;
        rx_desc_len = fw_desc->len;

+       if (fw_desc->u.bits.discard) {
+               ath10k_dbg(ar, ATH10K_DBG_HTT, "htt discard mpdu\n");
+               goto err;
+       }
+
        /* I have not yet seen any case where num_mpdu_ranges > 1.
         * qcacld does not seem handle that case either, so we 
introduce
the
         * same limitiation here as well.
diff --git a/drivers/net/wireless/ath/ath10k/rx_desc.h
b/drivers/net/wireless/ath/ath10k/rx_desc.h
index f2b6bf8f0d60..705b6295e466 100644
--- a/drivers/net/wireless/ath/ath10k/rx_desc.h
+++ b/drivers/net/wireless/ath/ath10k/rx_desc.h
@@ -1282,7 +1282,19 @@ struct fw_rx_desc_base {
 #define FW_RX_DESC_UDP              (1 << 6)

 struct fw_rx_desc_hl {
-       u8 info0;
+       union {
+               struct {
+               u8 discard:1,
+                  forward:1,
+                  any_err:1,
+                  dup_err:1,
+                  reserved:1,
+                  inspect:1,
+                  extension:2;
+               } bits;
+               u8 info0;
+       } u;
Am I misled here, or are you introducing endianness issues here? From 
C99:

"The order of allocation of bit-fields within a unit (high-order to
low-order or low-order to high-order) is implementation-defined."

Now, we're pretty well attuned to two implementations (big and little
endian), and this should work for the most common one (little endian),
but it's not wise to assume everyone is little endian.

Brian
This issue was identified in internal review, but due to the embargo 
expiring
we sent it out as-is since that is what had been tested. The author will 
have
a follow-up change to replace this.

-- 
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora 
Forum,
a Linux Foundation Collaborative Project
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help