[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20191112084259.GI100264@gmail.com>
Date: Tue, 12 Nov 2019 09:42:59 +0100
From: Ingo Molnar <mingo@...nel.org>
To: Thomas Gleixner <tglx@...utronix.de>
Cc: LKML <linux-kernel@...r.kernel.org>, x86@...nel.org,
Linus Torvalds <torvalds@...uxfoundation.org>,
Andy Lutomirski <luto@...nel.org>,
Stephen Hemminger <stephen@...workplumber.org>,
Willy Tarreau <w@....eu>, Juergen Gross <jgross@...e.com>,
Sean Christopherson <sean.j.christopherson@...el.com>,
"H. Peter Anvin" <hpa@...or.com>
Subject: Re: [patch V2 14/16] x86/iopl: Restrict iopl() permission scope
* Thomas Gleixner <tglx@...utronix.de> wrote:
> +static void task_update_io_bitmap(void)
> +{
> + struct thread_struct *t = ¤t->thread;
> +
> + preempt_disable();
> + if (t->iopl_emul == 3 || t->io_bitmap) {
> + /* TSS update is handled on exit to user space */
> + set_thread_flag(TIF_IO_BITMAP);
> + } else {
> + clear_thread_flag(TIF_IO_BITMAP);
> + /* Invalidate TSS */
> + tss_update_io_bitmap();
> + }
> + preempt_enable();
> +}
> +
> void io_bitmap_exit(void)
> {
> struct io_bitmap *iobm = current->thread.io_bitmap;
>
> - preempt_disable();
> current->thread.io_bitmap = NULL;
> - clear_thread_flag(TIF_IO_BITMAP);
> - tss_update_io_bitmap();
> - preempt_enable();
> + task_update_io_bitmap();
BTW., isn't the preempt_disable()/enable() sequence only needed around
the tss_update_io_bitmap() call?
->iopl_emul, ->io_bitmap and TIF_IO_BITMAP can only be set by the current
task AFAICS.
I.e. critical section could be narrowed a bit.
Thanks,
Ingo
Powered by blists - more mailing lists