Re: [RFC/PATCH][NOR] Fix cfi_cmdset_0001.c FL_SYNCING race
Alexander Belyakov
abelyako at mail.ru
Sun May 4 05:47:39 EDT 2008
>
> We found this issue was related to FL_SYNCING state race with other operations.
> The patch below tries to remove these races.
>
> Any comments on this patch are welcome.
>
No one is interested in fixing these issues?
I hope it is not because of some trailing whitespaces :)
I've fixed them and checkpatch.pl has passed.
---
Signed-off-by: Alexander Belyakov <abelyako at googlemail.com>
diff -uNrp a/drivers/mtd/chips/cfi_cmdset_0001.c b/drivers/mtd/chips/cfi_cmdset_0001.c
--- a/drivers/mtd/chips/cfi_cmdset_0001.c 2008-05-01 22:15:28.000000000 +0400
+++ b/drivers/mtd/chips/cfi_cmdset_0001.c 2008-05-04 13:51:20.000000000 +0400
@@ -701,6 +701,10 @@ static int chip_ready (struct map_info *
struct cfi_pri_intelext *cfip = cfi->cmdset_priv;
unsigned long timeo = jiffies + HZ;
+ /* Prevent setting state FL_SYNCING for chip in suspended state. */
+ if (FL_SYNCING == mode && FL_READY != chip->oldstate)
+ goto sleep;
+
switch (chip->state) {
case FL_STATUS:
@@ -806,8 +810,9 @@ static int get_chip(struct map_info *map
DECLARE_WAITQUEUE(wait, current);
retry:
- if (chip->priv && (mode == FL_WRITING || mode == FL_ERASING
- || mode == FL_OTP_WRITE || mode == FL_SHUTDOWN)) {
+ if (chip->priv &&
+ (mode == FL_WRITING || mode == FL_ERASING || mode == FL_OTP_WRITE
+ || mode == FL_SHUTDOWN) && chip->state != FL_SYNCING) {
/*
* OK. We have possibility for contention on the write/erase
* operations which are global to the real chip and not per
@@ -857,6 +862,14 @@ static int get_chip(struct map_info *map
return ret;
}
spin_lock(&shared->lock);
+
+ /* We should not own chip if it is already
+ * in FL_SYNCING state. Put contender and retry. */
+ if (chip->state == FL_SYNCING) {
+ put_chip(map, contender, contender->start);
+ spin_unlock(contender->mutex);
+ goto retry;
+ }
spin_unlock(contender->mutex);
}
More information about the linux-mtd
mailing list