* [PATCH 1/1] virt: arm-cca-guest: fix error check for RSI_INCOMPLETE
@ 2026-04-10 16:36 Sami Mujawar
2026-04-10 16:51 ` Yeoreum Yun
` (2 more replies)
0 siblings, 3 replies; 4+ messages in thread
From: Sami Mujawar @ 2026-04-10 16:36 UTC (permalink / raw)
To: linux-arm-kernel, linux-kernel
Cc: catalin.marinas, will, steven.price, suzuki.poulose, sami.mujawar,
gshan, YeoReum.Yun, Jagdish Gediya
The RSI interface can return RSI_INCOMPLETE when a report spans
multiple granules. This is an expected condition and should not be
treated as a fatal error.
Currently, arm_cca_report_new() checks for `info.result != RSI_SUCCESS`
and bails out, which incorrectly flags RSI_INCOMPLETE as a failure.
Fix the check to only break out on results other than RSI_SUCCESS or
RSI_INCOMPLETE.
This ensures partial reports are handled correctly and avoids spurious
-ENXIO errors when generating attestation reports.
Fixes: 7999edc484ca ("virt: arm-cca-guest: TSM_REPORT support for realms")
Signed-off-by: Sami Mujawar <sami.mujawar@arm.com>
Reported-by: Jagdish Gediya <Jagdish.Gediya@arm.com>
---
drivers/virt/coco/arm-cca-guest/arm-cca-guest.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/virt/coco/arm-cca-guest/arm-cca-guest.c b/drivers/virt/coco/arm-cca-guest/arm-cca-guest.c
index 0c9ea24a200c..66d00b6ceb78 100644
--- a/drivers/virt/coco/arm-cca-guest/arm-cca-guest.c
+++ b/drivers/virt/coco/arm-cca-guest/arm-cca-guest.c
@@ -157,7 +157,8 @@ static int arm_cca_report_new(struct tsm_report *report, void *data)
} while (info.result == RSI_INCOMPLETE &&
info.offset < RSI_GRANULE_SIZE);
- if (info.result != RSI_SUCCESS) {
+ /* Break out in case of failure */
+ if (info.result != RSI_SUCCESS && info.result != RSI_INCOMPLETE) {
ret = -ENXIO;
token_size = 0;
goto exit_free_granule_page;
--
SAMI:{C3F47F37-75D8-414A-A8BA-3980EC8A46D7}
^ permalink raw reply related [flat|nested] 4+ messages in thread* Re: [PATCH 1/1] virt: arm-cca-guest: fix error check for RSI_INCOMPLETE
2026-04-10 16:36 [PATCH 1/1] virt: arm-cca-guest: fix error check for RSI_INCOMPLETE Sami Mujawar
@ 2026-04-10 16:51 ` Yeoreum Yun
2026-04-10 17:29 ` Suzuki K Poulose
2026-04-10 20:50 ` Gavin Shan
2 siblings, 0 replies; 4+ messages in thread
From: Yeoreum Yun @ 2026-04-10 16:51 UTC (permalink / raw)
To: Sami Mujawar
Cc: linux-arm-kernel, linux-kernel, catalin.marinas, will,
steven.price, suzuki.poulose, gshan, Jagdish Gediya
Look good to me.
Reviewed-by: Yeoreum Yun <yeoreum.yun@arm.com>
On Fri, Apr 10, 2026 at 05:36:36PM +0100, Sami Mujawar wrote:
> The RSI interface can return RSI_INCOMPLETE when a report spans
> multiple granules. This is an expected condition and should not be
> treated as a fatal error.
>
> Currently, arm_cca_report_new() checks for `info.result != RSI_SUCCESS`
> and bails out, which incorrectly flags RSI_INCOMPLETE as a failure.
> Fix the check to only break out on results other than RSI_SUCCESS or
> RSI_INCOMPLETE.
>
> This ensures partial reports are handled correctly and avoids spurious
> -ENXIO errors when generating attestation reports.
>
> Fixes: 7999edc484ca ("virt: arm-cca-guest: TSM_REPORT support for realms")
> Signed-off-by: Sami Mujawar <sami.mujawar@arm.com>
> Reported-by: Jagdish Gediya <Jagdish.Gediya@arm.com>
> ---
> drivers/virt/coco/arm-cca-guest/arm-cca-guest.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/virt/coco/arm-cca-guest/arm-cca-guest.c b/drivers/virt/coco/arm-cca-guest/arm-cca-guest.c
> index 0c9ea24a200c..66d00b6ceb78 100644
> --- a/drivers/virt/coco/arm-cca-guest/arm-cca-guest.c
> +++ b/drivers/virt/coco/arm-cca-guest/arm-cca-guest.c
> @@ -157,7 +157,8 @@ static int arm_cca_report_new(struct tsm_report *report, void *data)
> } while (info.result == RSI_INCOMPLETE &&
> info.offset < RSI_GRANULE_SIZE);
>
> - if (info.result != RSI_SUCCESS) {
> + /* Break out in case of failure */
> + if (info.result != RSI_SUCCESS && info.result != RSI_INCOMPLETE) {
> ret = -ENXIO;
> token_size = 0;
> goto exit_free_granule_page;
> --
> SAMI:{C3F47F37-75D8-414A-A8BA-3980EC8A46D7}
>
--
Sincerely,
Yeoreum Yun
^ permalink raw reply [flat|nested] 4+ messages in thread* Re: [PATCH 1/1] virt: arm-cca-guest: fix error check for RSI_INCOMPLETE
2026-04-10 16:36 [PATCH 1/1] virt: arm-cca-guest: fix error check for RSI_INCOMPLETE Sami Mujawar
2026-04-10 16:51 ` Yeoreum Yun
@ 2026-04-10 17:29 ` Suzuki K Poulose
2026-04-10 20:50 ` Gavin Shan
2 siblings, 0 replies; 4+ messages in thread
From: Suzuki K Poulose @ 2026-04-10 17:29 UTC (permalink / raw)
To: Sami Mujawar, linux-arm-kernel, linux-kernel
Cc: catalin.marinas, will, steven.price, gshan, YeoReum.Yun,
Jagdish Gediya
On 10/04/2026 17:36, Sami Mujawar wrote:
> The RSI interface can return RSI_INCOMPLETE when a report spans
> multiple granules. This is an expected condition and should not be
> treated as a fatal error.
>
> Currently, arm_cca_report_new() checks for `info.result != RSI_SUCCESS`
> and bails out, which incorrectly flags RSI_INCOMPLETE as a failure.
> Fix the check to only break out on results other than RSI_SUCCESS or
> RSI_INCOMPLETE.
>
> This ensures partial reports are handled correctly and avoids spurious
> -ENXIO errors when generating attestation reports.
>
> Fixes: 7999edc484ca ("virt: arm-cca-guest: TSM_REPORT support for realms")
> Signed-off-by: Sami Mujawar <sami.mujawar@arm.com>
> Reported-by: Jagdish Gediya <Jagdish.Gediya@arm.com>
Reviewed-by: Suzuki K Poulose <suzuki.poulose@arm.com>
> ---
> drivers/virt/coco/arm-cca-guest/arm-cca-guest.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/virt/coco/arm-cca-guest/arm-cca-guest.c b/drivers/virt/coco/arm-cca-guest/arm-cca-guest.c
> index 0c9ea24a200c..66d00b6ceb78 100644
> --- a/drivers/virt/coco/arm-cca-guest/arm-cca-guest.c
> +++ b/drivers/virt/coco/arm-cca-guest/arm-cca-guest.c
> @@ -157,7 +157,8 @@ static int arm_cca_report_new(struct tsm_report *report, void *data)
> } while (info.result == RSI_INCOMPLETE &&
> info.offset < RSI_GRANULE_SIZE);
>
> - if (info.result != RSI_SUCCESS) {
> + /* Break out in case of failure */
> + if (info.result != RSI_SUCCESS && info.result != RSI_INCOMPLETE) {
> ret = -ENXIO;
> token_size = 0;
> goto exit_free_granule_page;
^ permalink raw reply [flat|nested] 4+ messages in thread* Re: [PATCH 1/1] virt: arm-cca-guest: fix error check for RSI_INCOMPLETE
2026-04-10 16:36 [PATCH 1/1] virt: arm-cca-guest: fix error check for RSI_INCOMPLETE Sami Mujawar
2026-04-10 16:51 ` Yeoreum Yun
2026-04-10 17:29 ` Suzuki K Poulose
@ 2026-04-10 20:50 ` Gavin Shan
2 siblings, 0 replies; 4+ messages in thread
From: Gavin Shan @ 2026-04-10 20:50 UTC (permalink / raw)
To: Sami Mujawar, linux-arm-kernel, linux-kernel
Cc: catalin.marinas, will, steven.price, suzuki.poulose, YeoReum.Yun,
Jagdish Gediya
On 4/11/26 2:36 AM, Sami Mujawar wrote:
> The RSI interface can return RSI_INCOMPLETE when a report spans
> multiple granules. This is an expected condition and should not be
> treated as a fatal error.
>
> Currently, arm_cca_report_new() checks for `info.result != RSI_SUCCESS`
> and bails out, which incorrectly flags RSI_INCOMPLETE as a failure.
> Fix the check to only break out on results other than RSI_SUCCESS or
> RSI_INCOMPLETE.
>
> This ensures partial reports are handled correctly and avoids spurious
> -ENXIO errors when generating attestation reports.
>
> Fixes: 7999edc484ca ("virt: arm-cca-guest: TSM_REPORT support for realms")
> Signed-off-by: Sami Mujawar <sami.mujawar@arm.com>
> Reported-by: Jagdish Gediya <Jagdish.Gediya@arm.com>
> ---
> drivers/virt/coco/arm-cca-guest/arm-cca-guest.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
Reviewed-by: Gavin Shan <gshan@redhat.com>
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2026-04-10 20:50 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-04-10 16:36 [PATCH 1/1] virt: arm-cca-guest: fix error check for RSI_INCOMPLETE Sami Mujawar
2026-04-10 16:51 ` Yeoreum Yun
2026-04-10 17:29 ` Suzuki K Poulose
2026-04-10 20:50 ` Gavin Shan
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox