| Message ID | 20241212105555.10364-1-jinjian.song@fibocom.com (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | [net-next,v1] net: wwan: t7xx: Fix FSM command timeout issue | expand |
On 12/12/2024 11:55 AM, Jinjian Song wrote: > When driver processes the internal state change command, it use > asynchronous thread to process the command operation. If the main > thread detects that the task has timed out, the asynchronous thread > will panic when executing te completion notification because the > main thread completion object is released. > > BUG: unable to handle page fault for address: fffffffffffffff8 > PGD 1f283a067 P4D 1f283a067 PUD 1f283c067 PMD 0 > Oops: 0000 [#1] PREEMPT SMP NOPTI > RIP: 0010:complete_all+0x3e/0xa0 > [...] > Call Trace: > <TASK> > ? __die_body+0x68/0xb0 > ? page_fault_oops+0x379/0x3e0 > ? exc_page_fault+0x69/0xa0 > ? asm_exc_page_fault+0x22/0x30 > ? complete_all+0x3e/0xa0 > fsm_main_thread+0xa3/0x9c0 [mtk_t7xx (HASH:1400 5)] > ? __pfx_autoremove_wake_function+0x10/0x10 > kthread+0xd8/0x110 > ? __pfx_fsm_main_thread+0x10/0x10 [mtk_t7xx (HASH:1400 5)] > ? __pfx_kthread+0x10/0x10 > ret_from_fork+0x38/0x50 > ? __pfx_kthread+0x10/0x10 > ret_from_fork_asm+0x1b/0x30 > </TASK> > [...] > CR2: fffffffffffffff8 > ---[ end trace 0000000000000000 ]--- > > After the main thread determines that the task has timed out, mark > the completion invalid, and add judgment in the asynchronous task. > > Signed-off-by: Jinjian Song <jinjian.song@fibocom.com> > --- > drivers/net/wwan/t7xx/t7xx_state_monitor.c | 7 +++++-- > 1 file changed, 5 insertions(+), 2 deletions(-) > > diff --git a/drivers/net/wwan/t7xx/t7xx_state_monitor.c b/drivers/net/wwan/t7xx/t7xx_state_monitor.c > index 3931c7a13f5a..57f1a7730fff 100644 > --- a/drivers/net/wwan/t7xx/t7xx_state_monitor.c > +++ b/drivers/net/wwan/t7xx/t7xx_state_monitor.c > @@ -108,7 +108,8 @@ static void fsm_finish_command(struct t7xx_fsm_ctl *ctl, struct t7xx_fsm_command > { > if (cmd->flag & FSM_CMD_FLAG_WAIT_FOR_COMPLETION) { > *cmd->ret = result; > - complete_all(cmd->done); > + if (cmd->done) > + complete_all(cmd->done); > } > > kfree(cmd); > @@ -503,8 +504,10 @@ int t7xx_fsm_append_cmd(struct t7xx_fsm_ctl *ctl, enum t7xx_fsm_cmd_state cmd_id > > wait_ret = wait_for_completion_timeout(&done, > msecs_to_jiffies(FSM_CMD_TIMEOUT_MS)); > - if (!wait_ret) > + if (!wait_ret) { > + cmd->done = NULL; > return -ETIMEDOUT; > + } > > return ret; > } If this is a fix then should be targeted to net and not net-next and probably should have Fixes: tag.
On Thu, Dec 12, 2024 at 06:55:55PM +0800, Jinjian Song wrote: > When driver processes the internal state change command, it use > asynchronous thread to process the command operation. If the main > thread detects that the task has timed out, the asynchronous thread > will panic when executing te completion notification because the > main thread completion object is released. s/it use/it uses an/ s/te/the/ s/is released/has been released/
From: Mateusz Polchlopek <mateusz.polchlopek@intel.com> >On 12/12/2024 11:55 AM, Jinjian Song wrote: >> When driver processes the internal state change command, it use >> asynchronous thread to process the command operation. If the main >> thread detects that the task has timed out, the asynchronous thread >> will panic when executing te completion notification because the >> main thread completion object is released. >> >[...] >If this is a fix then should be targeted to net and not net-next >and probably should have Fixes: tag. Hi Mateusz, Please let me request to net and add the Fixes: tag. Thanks. Jinjian, Best Regards.
diff --git a/drivers/net/wwan/t7xx/t7xx_state_monitor.c b/drivers/net/wwan/t7xx/t7xx_state_monitor.c index 3931c7a13f5a..57f1a7730fff 100644 --- a/drivers/net/wwan/t7xx/t7xx_state_monitor.c +++ b/drivers/net/wwan/t7xx/t7xx_state_monitor.c @@ -108,7 +108,8 @@ static void fsm_finish_command(struct t7xx_fsm_ctl *ctl, struct t7xx_fsm_command { if (cmd->flag & FSM_CMD_FLAG_WAIT_FOR_COMPLETION) { *cmd->ret = result; - complete_all(cmd->done); + if (cmd->done) + complete_all(cmd->done); } kfree(cmd); @@ -503,8 +504,10 @@ int t7xx_fsm_append_cmd(struct t7xx_fsm_ctl *ctl, enum t7xx_fsm_cmd_state cmd_id wait_ret = wait_for_completion_timeout(&done, msecs_to_jiffies(FSM_CMD_TIMEOUT_MS)); - if (!wait_ret) + if (!wait_ret) { + cmd->done = NULL; return -ETIMEDOUT; + } return ret; }
When driver processes the internal state change command, it use asynchronous thread to process the command operation. If the main thread detects that the task has timed out, the asynchronous thread will panic when executing te completion notification because the main thread completion object is released. BUG: unable to handle page fault for address: fffffffffffffff8 PGD 1f283a067 P4D 1f283a067 PUD 1f283c067 PMD 0 Oops: 0000 [#1] PREEMPT SMP NOPTI RIP: 0010:complete_all+0x3e/0xa0 [...] Call Trace: <TASK> ? __die_body+0x68/0xb0 ? page_fault_oops+0x379/0x3e0 ? exc_page_fault+0x69/0xa0 ? asm_exc_page_fault+0x22/0x30 ? complete_all+0x3e/0xa0 fsm_main_thread+0xa3/0x9c0 [mtk_t7xx (HASH:1400 5)] ? __pfx_autoremove_wake_function+0x10/0x10 kthread+0xd8/0x110 ? __pfx_fsm_main_thread+0x10/0x10 [mtk_t7xx (HASH:1400 5)] ? __pfx_kthread+0x10/0x10 ret_from_fork+0x38/0x50 ? __pfx_kthread+0x10/0x10 ret_from_fork_asm+0x1b/0x30 </TASK> [...] CR2: fffffffffffffff8 ---[ end trace 0000000000000000 ]--- After the main thread determines that the task has timed out, mark the completion invalid, and add judgment in the asynchronous task. Signed-off-by: Jinjian Song <jinjian.song@fibocom.com> --- drivers/net/wwan/t7xx/t7xx_state_monitor.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-)