[PATCH 3/5] hwmon/w83781d: Drop REALLY_SLOW_IO setting

Juergen Gross posted 5 patches 5 days, 6 hours ago
[PATCH 3/5] hwmon/w83781d: Drop REALLY_SLOW_IO setting
Posted by Juergen Gross 5 days, 6 hours ago
In w83781d_isa_found() there is REALLY_SLOW_IO defined around some port
accesses, probably in order to wait between multiple accesses.

Unfortunately this isn't making any difference compared to not having
this define since more than a decade, as REALLY_SLOW_IO needs to be
defined while "#include <asm/io.h>" is called to have an effect.

As there seem not to be any outstanding issues in spite of this having
no effect, just drop the "#define" and add a remark to the related
comment.

Signed-off-by: Juergen Gross <jgross@suse.com>
---
 drivers/hwmon/w83781d.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/drivers/hwmon/w83781d.c b/drivers/hwmon/w83781d.c
index 076200ed2ec9..f664c2152a6d 100644
--- a/drivers/hwmon/w83781d.c
+++ b/drivers/hwmon/w83781d.c
@@ -1850,10 +1850,12 @@ w83781d_isa_found(unsigned short address)
 		}
 	}
 
-#define REALLY_SLOW_IO
 	/*
 	 * We need the timeouts for at least some W83781D-like
 	 * chips. But only if we read 'undefined' registers.
+	 * There used to be a "#define REALLY_SLOW_IO" to enforce that, but
+	 * this has been without any effect since more than a decade, so it
+	 * has been dropped.
 	 */
 	val = inb_p(address + 1);
 	if (inb_p(address + 2) != val
@@ -1862,7 +1864,6 @@ w83781d_isa_found(unsigned short address)
 		pr_debug("Detection failed at step %d\n", 1);
 		goto release;
 	}
-#undef REALLY_SLOW_IO
 
 	/*
 	 * We should be able to change the 7 LSB of the address port. The
-- 
2.51.0
Re: [PATCH 3/5] hwmon/w83781d: Drop REALLY_SLOW_IO setting
Posted by Guenter Roeck 5 days, 5 hours ago
On Wed, Nov 26, 2025 at 05:20:16PM +0100, Juergen Gross wrote:
> In w83781d_isa_found() there is REALLY_SLOW_IO defined around some port
> accesses, probably in order to wait between multiple accesses.
> 
> Unfortunately this isn't making any difference compared to not having
> this define since more than a decade, as REALLY_SLOW_IO needs to be
> defined while "#include <asm/io.h>" is called to have an effect.
> 
> As there seem not to be any outstanding issues in spite of this having
> no effect, just drop the "#define" and add a remark to the related
> comment.
> 
> Signed-off-by: Juergen Gross <jgross@suse.com>

Applied.

Thanks,
Guenter