On Tue, Jun 04, 2019 at 11:18:44PM +0800, Ming Lei wrote:
> Hi Guenter,
>
> Thanks for your test and report!
>
> On Tue, Jun 04, 2019 at 06:38:49AM -0700, Guenter Roeck wrote:
> > On Tue, Jun 04, 2019 at 04:23:06PM +0800, Ming Lei wrote:
> > > Hi,
> > >
> > > Guenter reported scsi boot issue caused by commit c3288dd8c232
> > > ("scsi: core: avoid pre-allocating big SGL for data").
> > >
> > > Turns out there are at least two issues.
> > >
> > > The 1st patch fixes issue in case that NO_SG_CHAIN on some ARCHs,
> > > such as alpha, arm and parisc.
> > >
> > > The 2nd patch makes esp scsi working with SG_CHAIN.
> > >
> >
> > Both patches applied on top of next-20190604.
> >
> > Results on alpha:
> >
> > Waiting for root device /dev/sda...
> > ------------[ cut here ]------------
> > WARNING: CPU: 0 PID: 7 at mm/slab.h:359 kmem_cache_free+0x120/0x2a0
> > virt_to_cache: Object is not a Slab page!
>
> Please apply the following patch against the posted two:
>
> diff --git a/lib/sg_pool.c b/lib/sg_pool.c
> index 47eecbe094d8..e042a1722615 100644
> --- a/lib/sg_pool.c
> +++ b/lib/sg_pool.c
> @@ -122,7 +122,7 @@ int sg_alloc_table_chained(struct sg_table *table, int
> nents,
> }
>
> /* User supposes that the 1st SGL includes real entry */
> - if (nents_first_chunk == 1) {
> + if (nents_first_chunk <= 1) {
> first_chunk = NULL;
> nents_first_chunk = 0;
> }
>
That fixes the problem.
Thanks,
Guenter
