Re: [PATCH v3] tty/sysrq: replace smp_processor_id() with get_cpu()

[Greg Kroah-Hartman]

On Tue, Sep 19, 2023 at 07:52:42AM +0200, Jiri Slaby wrote:
> On 18. 09. 23, 10:10, Greg Kroah-Hartman wrote:
> > On Tue, Aug 22, 2023 at 03:26:06PM +0500, Muhammad Usama Anjum wrote:
> > > The smp_processor_id() shouldn't be called from preemptible code.
> > > Instead use get_cpu() and put_cpu() which disables preemption in
> > > addition to getting the processor id. This fixes the following bug:
> > > 
> > > [  119.143590] sysrq: Show backtrace of all active CPUs
> > > [  119.143902] BUG: using smp_processor_id() in preemptible [00000000] code: bash/873
> > > [  119.144586] caller is debug_smp_processor_id+0x20/0x30
> > > [  119.144827] CPU: 6 PID: 873 Comm: bash Not tainted 5.10.124-dirty #3
> > > [  119.144861] Hardware name: QEMU QEMU Virtual Machine, BIOS 2023.05-1 07/22/2023
> > > [  119.145053] Call trace:
> > > [  119.145093]  dump_backtrace+0x0/0x1a0
> > > [  119.145122]  show_stack+0x18/0x70
> > > [  119.145141]  dump_stack+0xc4/0x11c
> > > [  119.145159]  check_preemption_disabled+0x100/0x110
> > > [  119.145175]  debug_smp_processor_id+0x20/0x30
> > > [  119.145195]  sysrq_handle_showallcpus+0x20/0xc0
> > > [  119.145211]  __handle_sysrq+0x8c/0x1a0
> > > [  119.145227]  write_sysrq_trigger+0x94/0x12c
> > > [  119.145247]  proc_reg_write+0xa8/0xe4
> > > [  119.145266]  vfs_write+0xec/0x280
> > > [  119.145282]  ksys_write+0x6c/0x100
> > > [  119.145298]  __arm64_sys_write+0x20/0x30
> > > [  119.145315]  el0_svc_common.constprop.0+0x78/0x1e4
> > > [  119.145332]  do_el0_svc+0x24/0x8c
> > > [  119.145348]  el0_svc+0x10/0x20
> > > [  119.145364]  el0_sync_handler+0x134/0x140
> > > [  119.145381]  el0_sync+0x180/0x1c0
> > > 
> > > Cc: stable@xxxxxxxxxxxxxxx
> > > Fixes: 47cab6a722d4 ("debug lockups: Improve lockup detection, fix generic arch fallback")
> > > Signed-off-by: Muhammad Usama Anjum <usama.anjum@xxxxxxxxxxxxx>
> > > ---
> > > Changes since v2:
> > > - Add changelog and resend
> > > 
> > > Changes since v1:
> > > - Add "Cc: stable@xxxxxxxxxxxxxxx" tag
> > > ---
> > >   drivers/tty/sysrq.c | 3 ++-
> > >   1 file changed, 2 insertions(+), 1 deletion(-)
> > > 
> > > diff --git a/drivers/tty/sysrq.c b/drivers/tty/sysrq.c
> > > index 23198e3f1461a..6b4a28bcf2f5f 100644
> > > --- a/drivers/tty/sysrq.c
> > > +++ b/drivers/tty/sysrq.c
> > > @@ -262,13 +262,14 @@ static void sysrq_handle_showallcpus(u8 key)
> > >   		if (in_hardirq())
> > >   			regs = get_irq_regs();
> > > -		pr_info("CPU%d:\n", smp_processor_id());
> > > +		pr_info("CPU%d:\n", get_cpu());
> > 
> > Why not call put_cpu() right here?
> > 
> > >   		if (regs)
> > >   			show_regs(regs);
> > >   		else
> > >   			show_stack(NULL, NULL, KERN_INFO);
> > >   		schedule_work(&sysrq_showallcpus);
> > > +		put_cpu();
> > 
> > Why wait so long here after you have scheduled work?  Please drop the
> > cpu reference right away, you don't need to hold it for this length of
> > time, right?
> 
> As I understand it, this way, schedule_work() will queue the work on the
> "gotten" (current) CPU. So sysrq_showregs_othercpus() will really dump other
> than the "gotten" cpu.

Ok, that makes a bit more sense, but that's not what the code does
today, have people seen the regs dumped from the wrong cpu in the past?

> If that is the case, it indeed should have been described in the commit log.

Agreed.

thanks for the review,

greg k-h