On Sun, Nov 19, 2023 at 11:16 PM Jiri Slaby [off-list ref] wrote:

On 18. 11. 23, 4:38, Anup Patel wrote:

quoted

diff --git a/drivers/tty/hvc/hvc_riscv_sbi.c b/drivers/tty/hvc/hvc_riscv_sbi.c
index 31f53fa77e4a..697c981221b5 100644
--- a/drivers/tty/hvc/hvc_riscv_sbi.c
+++ b/drivers/tty/hvc/hvc_riscv_sbi.c

...

quoted

-static int __init hvc_sbi_console_init(void)
+static int hvc_sbi_dbcn_tty_get(uint32_t vtermno, char *buf, int count)
  {
-     hvc_instantiate(0, 0, &hvc_sbi_ops);
+     phys_addr_t pa;
+
+     if (is_vmalloc_addr(buf)) {

I wonder, where does this buf come from, so that you have to check for
vmalloc?

When VMAP_STCK is enabled, stack allocation depends on the vmalloc.
That's why we have to if the buf is allocated using vmalloc.

quoted

+             pa = page_to_phys(vmalloc_to_page(buf)) + offset_in_page(buf);
+             if (PAGE_SIZE < (offset_in_page(buf) + count))

Am I the only one who would prefer:
   if (count + offset_in_page(buf) > PAGE_SIZE)
?

quoted

+                     count = PAGE_SIZE - offset_in_page(buf);
+     } else {
+             pa = __pa(buf);
+     }
+
+     return sbi_debug_console_read(count, pa);
+}

thanks,
--
js
suse labs