diff mbox series

[net] NFC: nci: fix sleep in atomic context bugs caused by nci_skb_alloc

Message ID 20220513133355.113222-1-duoming@zju.edu.cn (mailing list archive)
State Superseded
Delegated to: Netdev Maintainers
Headers show
Series [net] NFC: nci: fix sleep in atomic context bugs caused by nci_skb_alloc | expand

Checks

Context Check Description
netdev/tree_selection success Clearly marked for net
netdev/fixes_present success Fixes tag present in non-next series
netdev/subject_prefix success Link
netdev/cover_letter success Single patches do not need cover letters
netdev/patch_count success Link
netdev/header_inline success No static functions without inline keyword in header files
netdev/build_32bit success Errors and warnings before: 0 this patch: 0
netdev/cc_maintainers fail 4 blamed authors not CCed: ilane@ti.com linville@tuxdriver.com christophe.ricard@gmail.com sameo@linux.intel.com; 4 maintainers not CCed: sameo@linux.intel.com linville@tuxdriver.com christophe.ricard@gmail.com ilane@ti.com
netdev/build_clang success Errors and warnings before: 0 this patch: 0
netdev/module_param success Was 0 now: 0
netdev/verify_signedoff success Signed-off-by tag matches author and committer
netdev/verify_fixes success Fixes tag looks correct
netdev/build_allmodconfig_warn success Errors and warnings before: 0 this patch: 0
netdev/checkpatch success total: 0 errors, 0 warnings, 0 checks, 24 lines checked
netdev/kdoc success Errors and warnings before: 0 this patch: 0
netdev/source_inline success Was 0 now: 0

Commit Message

Duoming Zhou May 13, 2022, 1:33 p.m. UTC
There are sleep in atomic context bugs when the request to secure
element of st-nci is timeout. The root cause is that nci_skb_alloc
with GFP_KERNEL parameter is called in st_nci_se_wt_timeout which is
a timer handler. The call paths that could trigger bugs are shown below:

    (interrupt context 1)
st_nci_se_wt_timeout
  nci_hci_send_event
    nci_hci_send_data
      nci_skb_alloc(..., GFP_KERNEL) //may sleep

   (interrupt context 2)
st_nci_se_wt_timeout
  nci_hci_send_event
    nci_hci_send_data
      nci_send_data
        nci_queue_tx_data_frags
          nci_skb_alloc(..., GFP_KERNEL) //may sleep

This patch changes allocation mode of nci_skb_alloc from GFP_KERNEL to
GFP_ATOMIC in order to prevent atomic context sleeping. The GFP_ATOMIC
flag makes memory allocation operation could be used in atomic context.

Fixes: 6a2968aaf50c ("NFC: basic NCI protocol implementation ")
Fixes: 11f54f228643 ("NFC: nci: Add HCI over NCI protocol support")
Signed-off-by: Duoming Zhou <duoming@zju.edu.cn>
---
 net/nfc/nci/data.c | 2 +-
 net/nfc/nci/hci.c  | 4 ++--
 2 files changed, 3 insertions(+), 3 deletions(-)

Comments

Krzysztof Kozlowski May 16, 2022, 6:15 a.m. UTC | #1
On 13/05/2022 15:33, Duoming Zhou wrote:
> There are sleep in atomic context bugs when the request to secure
> element of st-nci is timeout. The root cause is that nci_skb_alloc
> with GFP_KERNEL parameter is called in st_nci_se_wt_timeout which is
> a timer handler. The call paths that could trigger bugs are shown below:
> 
>     (interrupt context 1)
> st_nci_se_wt_timeout
>   nci_hci_send_event
>     nci_hci_send_data
>       nci_skb_alloc(..., GFP_KERNEL) //may sleep
> 
>    (interrupt context 2)
> st_nci_se_wt_timeout
>   nci_hci_send_event
>     nci_hci_send_data
>       nci_send_data
>         nci_queue_tx_data_frags
>           nci_skb_alloc(..., GFP_KERNEL) //may sleep
> 
> This patch changes allocation mode of nci_skb_alloc from GFP_KERNEL to
> GFP_ATOMIC in order to prevent atomic context sleeping. The GFP_ATOMIC
> flag makes memory allocation operation could be used in atomic context.
> 
> Fixes: 6a2968aaf50c ("NFC: basic NCI protocol implementation ")
> Fixes: 11f54f228643 ("NFC: nci: Add HCI over NCI protocol support")
> Signed-off-by: Duoming Zhou <duoming@zju.edu.cn>


Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>


Best regards,
Krzysztof
Jakub Kicinski May 16, 2022, 8:07 p.m. UTC | #2
On Fri, 13 May 2022 21:33:55 +0800 Duoming Zhou wrote:
> Fixes: 6a2968aaf50c ("NFC: basic NCI protocol implementation ")
> Fixes: 11f54f228643 ("NFC: nci: Add HCI over NCI protocol support")

Are there more bad callers? If st_nci_se_wt_timeout is the only source
of trouble then the fixes tag should point to when it was added, rather
than when the callee was added.
Duoming Zhou May 17, 2022, 1:10 a.m. UTC | #3
Hello,

On Mon, 16 May 2022 13:07:05 -0700 Jakub Kicinski wrote:

> On Fri, 13 May 2022 21:33:55 +0800 Duoming Zhou wrote:
> > Fixes: 6a2968aaf50c ("NFC: basic NCI protocol implementation ")
> > Fixes: 11f54f228643 ("NFC: nci: Add HCI over NCI protocol support")
> 
> Are there more bad callers? If st_nci_se_wt_timeout is the only source
> of trouble then the fixes tag should point to when it was added, rather
> than when the callee was added.

The st_nci_se_wt_timeout is the only source of trouble, it was added in
ed06aeefdac3 ("nfc: st-nci: Rename st21nfcb to st-nci"). I will send
patch v2.

Best regards,
Duoming Zhou
diff mbox series

Patch

diff --git a/net/nfc/nci/data.c b/net/nfc/nci/data.c
index 6055dc9a82a..aa5e712adf0 100644
--- a/net/nfc/nci/data.c
+++ b/net/nfc/nci/data.c
@@ -118,7 +118,7 @@  static int nci_queue_tx_data_frags(struct nci_dev *ndev,
 
 		skb_frag = nci_skb_alloc(ndev,
 					 (NCI_DATA_HDR_SIZE + frag_len),
-					 GFP_KERNEL);
+					 GFP_ATOMIC);
 		if (skb_frag == NULL) {
 			rc = -ENOMEM;
 			goto free_exit;
diff --git a/net/nfc/nci/hci.c b/net/nfc/nci/hci.c
index 19703a649b5..78c4b6addf1 100644
--- a/net/nfc/nci/hci.c
+++ b/net/nfc/nci/hci.c
@@ -153,7 +153,7 @@  static int nci_hci_send_data(struct nci_dev *ndev, u8 pipe,
 
 	i = 0;
 	skb = nci_skb_alloc(ndev, conn_info->max_pkt_payload_len +
-			    NCI_DATA_HDR_SIZE, GFP_KERNEL);
+			    NCI_DATA_HDR_SIZE, GFP_ATOMIC);
 	if (!skb)
 		return -ENOMEM;
 
@@ -184,7 +184,7 @@  static int nci_hci_send_data(struct nci_dev *ndev, u8 pipe,
 		if (i < data_len) {
 			skb = nci_skb_alloc(ndev,
 					    conn_info->max_pkt_payload_len +
-					    NCI_DATA_HDR_SIZE, GFP_KERNEL);
+					    NCI_DATA_HDR_SIZE, GFP_ATOMIC);
 			if (!skb)
 				return -ENOMEM;