Re: [PATCH] workqueue: fix duplicate wq_update_pod_attrs_buf allocation

[Kees Cook]

On Fri, Sep 15, 2023 at 01:24:25PM +0200, Geert Uytterhoeven wrote:
> On Wed, Sep 13, 2023 at 8:05 PM John Ogness <john.ogness@xxxxxxxxxxxxx> wrote:
> > One of the hunks for the patch resulting in
> > commit 84193c07105c ("workqueue: Generalize unbound CPU pods") was
> > applied incorrectly. This resulted in @wq_update_pod_attrs_buf
> > being allocated twice.
> >
> > From the kmemleak detector:
> >
> > unreferenced object 0xc0000000040074c0 (size 64):
> >   comm "swapper/0", pid 0, jiffies 4294937296 (age 1936.580s)
> >   hex dump (first 32 bytes):
> >     00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 03  ................
> >     00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
> >   backtrace:
> >     [<0000000076b83a6e>] .kmalloc_trace+0x54/0x190
> >     [<00000000898f2356>] .alloc_workqueue_attrs+0x2c/0x60
> >     [<0000000063365e1f>] .workqueue_init_early+0xe4/0x4b8
> >     [<00000000ca97ff39>] .start_kernel+0x8d0/0xba4
> >     [<000000002ee12080>] start_here_common+0x1c/0x20
> >
> > Remove the redundant allocation.
> >
> > Signed-off-by: John Ogness <john.ogness@xxxxxxxxxxxxx>
> 
> Stephen posted a similar patch before:
> https://lore.kernel.org/all/20230905174935.2d75feab@xxxxxxxxxxxxxxxxxx

I tripped over this kmemleak report too. It'd be nice to get this into
-rc2. Tejun, are able to get this or the Sep 5th patch to Linus soon?

-Kees

-- 
Kees Cook