Re: [PATCH] zcrypt: handle AP Info notification from CHSC SEI command

From: Cornelia Huck
Date: Thu Jan 31 2019 - 04:56:03 EST


On Wed, 30 Jan 2019 12:48:46 -0500
Tony Krowiak <akrowiak@xxxxxxxxxxxxx> wrote:

> The current AP bus implementation periodically polls the AP configuration
> to detect changes. When the AP configuration is dynamically changed via the
> SE or an SCLP instruction, the changes will not be reflected to sysfs until
> the next time the AP configuration is polled. The CHSC architecture
> provides a Store Event Information (SEI)command to make notification of an

missing blank ---------------------------^

> AP configuration change. This patch introduces a handler to process
> notification from the CHSC SEI command BY immediately kickING off an AP bus

s/BY/by/

s/kickING/kicking/

> scan.

Scan-after-event sounds useful, yeah.

Two questions:
- Does the event cover _any_ change to the AP configuration, or can the
periodic scan detect changes that are not signaled?
- Do we want to generate such an event in QEMU on plugging/unplugging
the vfio-ap device?

>
> Signed-off-by: Tony Krowiak <akrowiak@xxxxxxxxxxxxx>
> Reviewed-by: Halil Pasic <pasic@xxxxxxxxxxxxx>
> ---
> arch/s390/include/asm/ap.h | 12 ++++++++++++
> drivers/s390/cio/chsc.c | 12 ++++++++++++
> drivers/s390/cio/chsc.h | 1 +
> drivers/s390/crypto/ap_bus.c | 12 ++++++++++++
> 4 files changed, 37 insertions(+)
>
> diff --git a/arch/s390/include/asm/ap.h b/arch/s390/include/asm/ap.h
> index 1a6a7092d942..c778593d509f 100644
> --- a/arch/s390/include/asm/ap.h
> +++ b/arch/s390/include/asm/ap.h
> @@ -360,4 +360,16 @@ static inline struct ap_queue_status ap_dqap(ap_qid_t qid,
> return reg1;
> }
>
> +/*
> + * Interface to tell the AP bus code that a configuration
> + * change has happened. The bus code should at least do
> + * an ap bus resource rescan.
> + */
> +#if IS_ENABLED(CONFIG_ZCRYPT)
> +void ap_bus_cfg_chg(void);
> +#else
> +#error "no CONFIG_ZCRYPT"

That #error looks like a development debugging leftover.

> +static void ap_bus_cfg_chg(void){};
> +#endif
> +
> #endif /* _ASM_S390_AP_H_ */
> diff --git a/drivers/s390/cio/chsc.c b/drivers/s390/cio/chsc.c
> index a0baee25134c..dccccc337078 100644
> --- a/drivers/s390/cio/chsc.c
> +++ b/drivers/s390/cio/chsc.c
> @@ -586,6 +586,15 @@ static void chsc_process_sei_scm_avail(struct chsc_sei_nt0_area *sei_area)
> " failed (rc=%d).\n", ret);
> }
>
> +static void chsc_process_sei_ap_cfg_chg(struct chsc_sei_nt0_area *sei_area)
> +{
> + CIO_CRW_EVENT(3, "chsc: ap config changed\n");
> + if (sei_area->rs != 5)
> + return;

I'm guessing that a reporting source of 5 means ap, right? (The code is
silent on all those magic rs values :/)

If so, should the debug logging be moved after the check?

> +
> + ap_bus_cfg_chg();
> +}
> +
> static void chsc_process_sei_nt2(struct chsc_sei_nt2_area *sei_area)
> {
> switch (sei_area->cc) {
> @@ -612,6 +621,9 @@ static void chsc_process_sei_nt0(struct chsc_sei_nt0_area *sei_area)
> case 2: /* i/o resource accessibility */
> chsc_process_sei_res_acc(sei_area);
> break;
> + case 3: /* ap config changed */
> + chsc_process_sei_ap_cfg_chg(sei_area);
> + break;
> case 7: /* channel-path-availability information */
> chsc_process_sei_chp_avail(sei_area);
> break;
> diff --git a/drivers/s390/cio/chsc.h b/drivers/s390/cio/chsc.h
> index 78aba8d94eec..5651066c46e3 100644
> --- a/drivers/s390/cio/chsc.h
> +++ b/drivers/s390/cio/chsc.h
> @@ -9,6 +9,7 @@
> #include <asm/chsc.h>
> #include <asm/schid.h>
> #include <asm/qdio.h>
> +#include <asm/ap.h>

I would probably include this in chsc.c instead.

>
> #define CHSC_SDA_OC_MSS 0x2
>
> diff --git a/drivers/s390/crypto/ap_bus.c b/drivers/s390/crypto/ap_bus.c
> index 48ea0004a56d..94f621783d6b 100644
> --- a/drivers/s390/crypto/ap_bus.c
> +++ b/drivers/s390/crypto/ap_bus.c
> @@ -35,6 +35,7 @@
> #include <linux/mod_devicetable.h>
> #include <linux/debugfs.h>
> #include <linux/ctype.h>
> +#include <asm/crw.h>

Why are you adding this #include?

>
> #include "ap_bus.h"
> #include "ap_debug.h"
> @@ -860,6 +861,17 @@ void ap_bus_force_rescan(void)
> EXPORT_SYMBOL(ap_bus_force_rescan);
>
> /*
> +* A config change has happened, Force an ap bus rescan.

s/Force/force/

> +*/
> +void ap_bus_cfg_chg(void)
> +{
> + AP_DBF(DBF_INFO, "%s config change, forcing bus rescan\n", __func__);
> +
> + ap_bus_force_rescan();
> +}
> +EXPORT_SYMBOL(ap_bus_cfg_chg);
> +
> +/*
> * hex2bitmap() - parse hex mask string and set bitmap.
> * Valid strings are "0x012345678" with at least one valid hex number.
> * Rest of the bitmap to the right is padded with 0. No spaces allowed