I'll revert that commit and re-release the 6.6.14 I just pushed
On Wed, Jan 24, 2024 at 9:48 AM Sebastian Andrzej Siewior
<bigeasy@xxxxxxxxxxxxx> wrote:
>
> On 2024-01-18 20:59:45 [-0000], Clark Williams wrote:
> > Hello RT-list!
> Hi Clark,
>
> > I'm pleased to announce the 6.6.12-rt20 stable release.
>
> I've been looking over it and it looks okay. Then I compared how it
> would do it vs your outcome and noticed this:
>
> --- a/arch/riscv/kernel/cpufeature.c
> +++ b/arch/riscv/kernel/cpufeature.c
> @@ -575,12 +575,6 @@ static int check_unaligned_access(void *param)
> if (per_cpu(misaligned_access_speed, cpu) != RISCV_HWPROBE_MISALIGNED_UNKNOWN)
> return;
>
> - page = alloc_pages(GFP_NOWAIT, get_order(MISALIGNED_BUFFER_SIZE));
> - if (!page) {
> - pr_warn("Can't alloc pages to measure memcpy performance");
> - return;
> - }
> -
> /* Make an unaligned destination buffer. */
> dst = (void *)((unsigned long)page_address(page) | 0x1);
> /* Unalign src as well, but differently (off by 1 + 2 = 3). */
>
> You shouldn't allocate that page. Nobody will free it, that page is
> passed via an argument now. Please drop hunk.
>
> While at it, do you think you can drop patch
> preempt-Put-preempt_enable-within-an-instrumentation.patch
>
> or revert commit
> c15abad8f7159 ("preempt: Put preempt_enable() within an instrumentation*() section.")
>
> I've been looking over it and it is fixed as of v6.6 so this patch is no
> longer needed.
>
> > Enjoy!
> > Clark
>
> Sebastian
>
