Re: [Syzkaller & bisect] There is "io_ring_exit_work" related Call Trace in v5.2-rc5 kernel

[Jens Axboe]

On 2/2/23 12:44?PM, Jens Axboe wrote:
> On 2/2/23 1:56?AM, Pengfei Xu wrote:
>> Hi Pavel Begunkov,
>>
>> On 2023-02-01 at 15:52:47 +0000, Pavel Begunkov wrote:
>>> On 1/28/23 14:49, Pavel Begunkov wrote:
>>>> On 1/28/23 08:44, Pengfei Xu wrote:
>>>>> Hi Pavel Begunkov and kernel expert,
>>>>>
>>>>> Greeting!
>>>>>
>>>>> There is "io_ring_exit_work" related Call Trace in v5.2-rc5 kernel in guest
>>>>> on Sapphire Rapids server.
>>>>
>>>> Thanks for the report, we'll take a look
>>>
>>> Not reproducible for me. Apparently, the repro creates a normal ring
>>> and immediately closes it, then io_ring_exit_work() hangs waiting to
>>> the task to execute task_work.
>>   Ah, I put the reproduced virtual machine in gitlab.
>>   git clone https://gitlab.com/xupengfe/repro_vm_env.git
>>   cd repro_vm_env
>>   tar -xvf repro_vm_env.tar.gz
>>   cd repro_vm_env; ./start3.sh  // it needs qemu-system-x86_64 and I used v7.1.0
>>      //start3.sh will load bzImage_2241ab53cbb5cdb08a6b2d4688feb13971058f65 v6.2-rc5 kernel
>>   In vm and login with root,  there is no password for root.
>>   After login vm successfully, ./repro   to reproduce this issue.
>>   It could be reproduced in about 6min.
>>
>>   If you already have qemu-system-x86_64, please ignore below info.
>>   If you want to install qemu v7.1.0 version,
>>   git clone https://github.com/qemu/qemu.git
>>   cd qemu
>>   git checkout -f v7.1.0
>>   mkdir build
>>   cd build
>>   yum install -y ninja-build.x86_64
>>   ../configure --target-list=x86_64-softmmu --enable-kvm --enable-vnc --enable-gtk --enable-sdl
>>   make
>>   make install
>>   Done.
> 
> Are you sure you attached the right kconfig in the initial report? It
> is from 5.18.x.

Looks like a bug in 9p, can you try this?


diff --git a/net/9p/client.c b/net/9p/client.c
index 622ec6a586ee..5756d812d3d4 100644
--- a/net/9p/client.c
+++ b/net/9p/client.c
@@ -18,6 +18,7 @@
 #include <linux/sched/signal.h>
 #include <linux/uaccess.h>
 #include <linux/uio.h>
+#include <linux/task_work.h>
 #include <net/9p/9p.h>
 #include <linux/parser.h>
 #include <linux/seq_file.h>
@@ -652,6 +653,24 @@ static struct p9_req_t *p9_client_prepare_req(struct p9_client *c,
 	return ERR_PTR(err);
 }
 
+static int p9_sigpending(void)
+{
+	int sigpending = 0;
+
+	if (signal_pending(current)) {
+		if (test_thread_flag(TIF_NOTIFY_SIGNAL))
+			clear_notify_signal();
+		if (task_work_pending(current))
+			task_work_run();
+		if (signal_pending(current)) {
+			sigpending = 1;
+			clear_thread_flag(TIF_SIGPENDING);
+		}
+	}
+
+	return sigpending;
+}
+
 /**
  * p9_client_rpc - issue a request and wait for a response
  * @c: client session
@@ -687,12 +706,7 @@ p9_client_rpc(struct p9_client *c, int8_t type, const char *fmt, ...)
 	req->tc.zc = false;
 	req->rc.zc = false;
 
-	if (signal_pending(current)) {
-		sigpending = 1;
-		clear_thread_flag(TIF_SIGPENDING);
-	} else {
-		sigpending = 0;
-	}
+	sigpending = p9_sigpending();
 
 	err = c->trans_mod->request(c, req);
 	if (err < 0) {
@@ -789,12 +803,7 @@ static struct p9_req_t *p9_client_zc_rpc(struct p9_client *c, int8_t type,
 	req->tc.zc = true;
 	req->rc.zc = true;
 
-	if (signal_pending(current)) {
-		sigpending = 1;
-		clear_thread_flag(TIF_SIGPENDING);
-	} else {
-		sigpending = 0;
-	}
+	sigpending = p9_sigpending();
 
 	err = c->trans_mod->zc_request(c, req, uidata, uodata,
 				       inlen, olen, in_hdrlen);

-- 
Jens Axboe