sending DASD patches through linux-block
Subsystems (3)
Stefan Haberland
maintainercover sending DASD patches through linux-block
Hi Jens,
we would like to get our DASD related patches upstream through your linux-block git tree in the future.
We hope to get a better integration and that we are able to identify potential conflicts with generic blocklayer changes earlier.
I hope this is OK for you. I have attached a first patch that fixes a bug in the DASD driver. This patch is based on the master branch of your linux-block git tree.
Please let me know if I should change something in my workflow.
Kind regards,
Stefan
Stefan Haberland (1):
s390/dasd: fix endless loop after read unit address configuration
drivers/s390/block/dasd_alias.c | 22 ++++++++++++++++------ 1 file changed, 16 insertions(+), 6 deletions(-)
Signature · 1 line
2.17.1
Stefan Haberland
maintainerPATCH 1/1 s390/dasd: fix endless loop after read unit address configuration
After getting a storage server event that causes the DASD device driver to update its unit address configuration during a device shutdown there is the possibility of an endless loop in the device driver.
In the system log there will be ongoing DASD error messages with RC: -19.
The reason is that the loop starting the ruac request only terminates when the retry counter is decreased to 0. But in the sleep_on function there are early exit paths that do not decrease the retry counter.
Prevent an endless loop by handling those cases separately.
Remove the unnecessary do..while loop since the sleep_on function takes care of retries by itself.
--- drivers/s390/block/dasd_alias.c | 22 ++++++++++++++++------ 1 file changed, 16 insertions(+), 6 deletions(-)
drivers/s390/block/dasd_alias.c +16 -6
@@ -383,6 +383,20 @@ suborder_not_supported(struct dasd_ccw_req *cqr)char msg_format;char msg_no; +/* +* intrc values ENODEV, ENOLINK and EPERM +* will be optained from sleep_on to indicate that no +* IO operation can be started +*/ +if (cqr->intrc == -ENODEV ) +return 1 ; + +if (cqr->intrc == -ENOLINK ) +return 1 ; + +if (cqr->intrc == -EPERM ) +return 1 ; + sense =dasd_get_sense (&cqr->irb);if (!sense)return 0 ;
@@ -447,12 +461,8 @@ static int read_unit_address_configuration(struct dasd_device *device, lcu->flags &= ~NEED_UAC_UPDATE ;spin_unlock_irqrestore (&lcu->lock, flags); -do { - rc =dasd_sleep_on (cqr); -if (rc &&suborder_not_supported (cqr)) -return -EOPNOTSUPP ; - }while (rc && (cqr->retries >0 )); -if (rc) { + rc =dasd_sleep_on (cqr); +if (rc && !suborder_not_supported (cqr)) {spin_lock_irqsave (&lcu->lock, flags); lcu->flags |=NEED_UAC_UPDATE ;spin_unlock_irqrestore (&lcu->lock, flags);
Signature · 1 line
2.17.1
Jens Axboe
maintainerRe: [PATCH 1/1] s390/dasd: fix endless loop after read unit address configuration
Quoted 6 lines · Stefan Haberland
After getting a storage server event that causes the DASD device driver to update its unit address configuration during a device shutdown there is the possibility of an endless loop in the device driver.
In the system log there will be ongoing DASD error messages with RC: -19.
Quoted 8 lines
The reason is that the loop starting the ruac request only terminates when the retry counter is decreased to 0. But in the
sleep_onfunction there are early exit paths that do not decrease the retry counter.Prevent an endless loop by handling those cases separately.
Remove the unnecessary do..while loop since the
sleep_onfunction takes care of retries by itself.
Applied for 5.3, thanks.
Signature · 1 line
Jens Axboe
Jens Axboe
maintainerQuoted 10 lines · Stefan Haberland
Hi Jens,
we would like to get our DASD related patches upstream through your linux-block git tree in the future.
We hope to get a better integration and that we are able to identify potential conflicts with generic blocklayer changes earlier.I hope this is OK for you. I have attached a first patch that fixes a bug in the DASD driver. This patch is based on the master branch of your linux-block git tree.
Certainly. No extra hassle for me, and if it makes it easier for you, then by all means.
Quoted 1 line
Please let me know if I should change something in my workflow.
Looks fine to me, I generally prefer patches (individual or series) to git trees, unless it's a huge series.
Signature · 1 line
Jens Axboe