Re: [PATCH] Swiotlb: Add CONFIG_HAS_IOMEM check around memremap() in the swiotlb_mem_remap()

[Wei Liu]

On Sun, Jan 02, 2022 at 11:54:46PM -0800, Christoph Hellwig wrote:
> On Fri, Dec 31, 2021 at 11:56:40AM -0500, Tianyu Lan wrote:
> > From: Tianyu Lan <Tianyu.Lan@xxxxxxxxxxxxx>
> > 
> > HAS_IOMEM option may not be selected on some platforms(e.g, s390) and this
> > will cause compile error due to miss memremap() implementation. Fix it via
> > adding HAS_IOMEM check around memremap() in the swiotlb.c.
> > 
> > Reported-by: kernel test robot <lkp@xxxxxxxxx>
> > Signed-off-by: Tianyu Lan <Tianyu.Lan@xxxxxxxxxxxxx>
> > ---
> >  kernel/dma/swiotlb.c | 2 ++
> >  1 file changed, 2 insertions(+)
> > 
> > diff --git a/kernel/dma/swiotlb.c b/kernel/dma/swiotlb.c
> > index b36c1cdd0c4f..3de651ba38cc 100644
> > --- a/kernel/dma/swiotlb.c
> > +++ b/kernel/dma/swiotlb.c
> > @@ -167,6 +167,7 @@ static void *swiotlb_mem_remap(struct io_tlb_mem *mem, unsigned long bytes)
> >  {
> >  	void *vaddr = NULL;
> >  
> > +#ifdef CONFIG_HAS_IOMEM
> 
> Please stub out all of swiotlb_mem_remap instead of the ifdef inside the
> function.

Does this look okay to you?

---8<---
diff --git a/kernel/dma/swiotlb.c b/kernel/dma/swiotlb.c
index b36c1cdd0c4f..f1e7ea160b43 100644
--- a/kernel/dma/swiotlb.c
+++ b/kernel/dma/swiotlb.c
@@ -163,6 +163,7 @@ static inline unsigned long nr_slots(u64 val)
  * when swiotlb_unencrypted_base is set. (e.g. for Hyper-V AMD SEV-SNP
  * Isolation VMs).
  */
+#ifdef CONFIG_HAS_IOMEM
 static void *swiotlb_mem_remap(struct io_tlb_mem *mem, unsigned long bytes)
 {
 	void *vaddr = NULL;
@@ -178,6 +179,12 @@ static void *swiotlb_mem_remap(struct io_tlb_mem *mem, unsigned long bytes)
 
 	return vaddr;
 }
+#else
+static void *swiotlb_mem_remap(struct io_tlb_mem *mem, unsigned long bytes)
+{
+	return NULL;
+}
+#endif
 
 /*
  * Early SWIOTLB allocation may be too early to allow an architecture to