* [PATCH] ata: pata_ep93xx: fix incorrect return value in data_xfer
@ 2026-06-01 19:21 Rosen Penev
0 siblings, 0 replies; 3+ messages in thread
From: Rosen Penev @ 2026-06-01 19:21 UTC (permalink / raw)
To: linux-ide; +Cc: Damien Le Moal, Niklas Cassel, open list
The while (words--) loop post-decrements words, so it underflows from
0 to UINT_MAX on exit. The function then returns "words << 1", which
is either 0 (for odd-length transfers after words++) or 0xFFFFFFFE
(for even-length transfers). Callers like __atapi_pio_bytes() depend
on the correct byte count to manage scatter-gather progress; a wrong
value causes ATAPI PIO transfer errors.
Fix by returning buflen instead of the corrupted words counter, matching
what ata_sff_data_xfer() does.
Assisted-by: opencode:big-pickle
Signed-off-by: Rosen Penev <rosenp@gmail.com>
---
drivers/ata/pata_ep93xx.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/ata/pata_ep93xx.c b/drivers/ata/pata_ep93xx.c
index 192f75636d18..5b79a7a019ce 100644
--- a/drivers/ata/pata_ep93xx.c
+++ b/drivers/ata/pata_ep93xx.c
@@ -506,7 +506,7 @@ static unsigned int ep93xx_pata_data_xfer(struct ata_queued_cmd *qc,
words++;
}
- return words << 1;
+ return buflen;
}
/* Note: original code is ata_devchk */
--
2.54.0
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [PATCH] ata: pata_ep93xx: fix incorrect return value in data_xfer
2026-07-20 2:09 Rosen Penev
@ 2026-07-23 7:41 ` Damien Le Moal
0 siblings, 0 replies; 3+ messages in thread
From: Damien Le Moal @ 2026-07-23 7:41 UTC (permalink / raw)
To: Rosen Penev, linux-ide; +Cc: Niklas Cassel, open list
On 7/20/26 11:09, Rosen Penev wrote:
> The while (words--) loop post-decrements words, so it underflows from
> 0 to UINT_MAX on exit. The function then returns "words << 1", which
> is either 0 (for odd-length transfers after words++) or 0xFFFFFFFE
> (for even-length transfers). Callers like __atapi_pio_bytes() depend
> on the correct byte count to manage scatter-gather progress; a wrong
> value causes ATAPI PIO transfer errors.
>
> Fix by returning ALIGN(buflen, 2) instead of the corrupted words
> counter, matching what ata_sff_data_xfer() effectively returns via
> its words counter.
>
> Assisted-by: opencode:big-pickle
> Signed-off-by: Rosen Penev <rosenp@gmail.com>
This does not apply cleanly. Please rebase on libata/for-7.3.
--
Damien Le Moal
Western Digital Research
^ permalink raw reply [flat|nested] 3+ messages in thread
* [PATCH] ata: pata_ep93xx: fix incorrect return value in data_xfer
@ 2026-07-20 2:09 Rosen Penev
2026-07-23 7:41 ` Damien Le Moal
0 siblings, 1 reply; 3+ messages in thread
From: Rosen Penev @ 2026-07-20 2:09 UTC (permalink / raw)
To: linux-ide; +Cc: Damien Le Moal, Niklas Cassel, open list
The while (words--) loop post-decrements words, so it underflows from
0 to UINT_MAX on exit. The function then returns "words << 1", which
is either 0 (for odd-length transfers after words++) or 0xFFFFFFFE
(for even-length transfers). Callers like __atapi_pio_bytes() depend
on the correct byte count to manage scatter-gather progress; a wrong
value causes ATAPI PIO transfer errors.
Fix by returning ALIGN(buflen, 2) instead of the corrupted words
counter, matching what ata_sff_data_xfer() effectively returns via
its words counter.
Assisted-by: opencode:big-pickle
Signed-off-by: Rosen Penev <rosenp@gmail.com>
---
drivers/ata/pata_ep93xx.c | 8 +++-----
1 file changed, 3 insertions(+), 5 deletions(-)
diff --git a/drivers/ata/pata_ep93xx.c b/drivers/ata/pata_ep93xx.c
index 21d7261f537e..97dda6213878 100644
--- a/drivers/ata/pata_ep93xx.c
+++ b/drivers/ata/pata_ep93xx.c
@@ -502,15 +502,13 @@ static unsigned int ep93xx_pata_data_xfer(struct ata_queued_cmd *qc,
if (unlikely(buflen & 0x01)) {
buf += buflen - 1;
- if (rw == READ) {
+ if (rw == READ)
*buf = ep93xx_pata_read_data(drv_data, IDECTRL_ADDR_DATA);
- } else {
+ else
ep93xx_pata_write_data(drv_data, *buf, IDECTRL_ADDR_DATA);
- }
- words++;
}
- return words << 1;
+ return ALIGN(buflen, 2);
}
/* Note: original code is ata_devchk */
--
2.55.0
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2026-07-23 7:41 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2026-06-01 19:21 [PATCH] ata: pata_ep93xx: fix incorrect return value in data_xfer Rosen Penev
2026-07-20 2:09 Rosen Penev
2026-07-23 7:41 ` Damien Le Moal
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®