2021-11-16 00:12:17

by Stefano Stabellini

[permalink] [raw]
Subject: [PATCH v2] xen: don't continue xenstore initialization in case of errors

From: Stefano Stabellini <[email protected]>

In case of errors in xenbus_init (e.g. missing xen_store_gfn parameter),
we goto out_error but we forget to reset xen_store_domain_type to
XS_UNKNOWN. As a consequence xenbus_probe_initcall and other initcalls
will still try to initialize xenstore resulting into a crash at boot.

[ 2.479830] Call trace:
[ 2.482314] xb_init_comms+0x18/0x150
[ 2.486354] xs_init+0x34/0x138
[ 2.489786] xenbus_probe+0x4c/0x70
[ 2.498432] xenbus_probe_initcall+0x2c/0x7c
[ 2.503944] do_one_initcall+0x54/0x1b8
[ 2.507358] kernel_init_freeable+0x1ac/0x210
[ 2.511617] kernel_init+0x28/0x130
[ 2.516112] ret_from_fork+0x10/0x20

Cc: <[email protected]>
Cc: [email protected]
Signed-off-by: Stefano Stabellini <[email protected]>
---
Changes in v2:
- use return 0 for the success case
- remove err initializer as it is not useful any longer
---
drivers/xen/xenbus/xenbus_probe.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/xen/xenbus/xenbus_probe.c b/drivers/xen/xenbus/xenbus_probe.c
index bd003ca8acbe..5967aa937255 100644
--- a/drivers/xen/xenbus/xenbus_probe.c
+++ b/drivers/xen/xenbus/xenbus_probe.c
@@ -909,7 +909,7 @@ static struct notifier_block xenbus_resume_nb = {

static int __init xenbus_init(void)
{
- int err = 0;
+ int err;
uint64_t v = 0;
xen_store_domain_type = XS_UNKNOWN;

@@ -983,8 +983,10 @@ static int __init xenbus_init(void)
*/
proc_create_mount_point("xen");
#endif
+ return 0;

out_error:
+ xen_store_domain_type = XS_UNKNOWN;
return err;
}

--
2.25.1



2021-11-16 07:45:01

by Jan Beulich

[permalink] [raw]
Subject: Re: [PATCH v2] xen: don't continue xenstore initialization in case of errors

On 15.11.2021 23:27, Stefano Stabellini wrote:
> From: Stefano Stabellini <[email protected]>
>
> In case of errors in xenbus_init (e.g. missing xen_store_gfn parameter),
> we goto out_error but we forget to reset xen_store_domain_type to
> XS_UNKNOWN. As a consequence xenbus_probe_initcall and other initcalls
> will still try to initialize xenstore resulting into a crash at boot.
>
> [ 2.479830] Call trace:
> [ 2.482314] xb_init_comms+0x18/0x150
> [ 2.486354] xs_init+0x34/0x138
> [ 2.489786] xenbus_probe+0x4c/0x70
> [ 2.498432] xenbus_probe_initcall+0x2c/0x7c
> [ 2.503944] do_one_initcall+0x54/0x1b8
> [ 2.507358] kernel_init_freeable+0x1ac/0x210
> [ 2.511617] kernel_init+0x28/0x130
> [ 2.516112] ret_from_fork+0x10/0x20
>
> Cc: <[email protected]>
> Cc: [email protected]
> Signed-off-by: Stefano Stabellini <[email protected]>

For the immediate purpose as described this looks okay, so
Reviewed-by: Jan Beulich <[email protected]>

However, aren't there further pieces missing on this error patch:
- clearing of xenstored_ready in case it got set,
- rolling back xenstored_local_init() (XS_LOCAL) and xen_remap()
(XS_HVM).
And shouldn't xs_init() failure when called from xenbus_probe()
also result in the driver not giving the appearance of being usable?

> --- a/drivers/xen/xenbus/xenbus_probe.c
> +++ b/drivers/xen/xenbus/xenbus_probe.c
> @@ -909,7 +909,7 @@ static struct notifier_block xenbus_resume_nb = {
>
> static int __init xenbus_init(void)
> {
> - int err = 0;
> + int err;
> uint64_t v = 0;
> xen_store_domain_type = XS_UNKNOWN;
>

Minor remark: You may want to take the opportunity and add the
missing blank line here to visually separate the assignment from
the declarations.

Jan


2021-11-16 20:41:15

by Stefano Stabellini

[permalink] [raw]
Subject: Re: [PATCH v2] xen: don't continue xenstore initialization in case of errors

On Tue, 16 Nov 2021, Jan Beulich wrote:
> On 15.11.2021 23:27, Stefano Stabellini wrote:
> > From: Stefano Stabellini <[email protected]>
> >
> > In case of errors in xenbus_init (e.g. missing xen_store_gfn parameter),
> > we goto out_error but we forget to reset xen_store_domain_type to
> > XS_UNKNOWN. As a consequence xenbus_probe_initcall and other initcalls
> > will still try to initialize xenstore resulting into a crash at boot.
> >
> > [ 2.479830] Call trace:
> > [ 2.482314] xb_init_comms+0x18/0x150
> > [ 2.486354] xs_init+0x34/0x138
> > [ 2.489786] xenbus_probe+0x4c/0x70
> > [ 2.498432] xenbus_probe_initcall+0x2c/0x7c
> > [ 2.503944] do_one_initcall+0x54/0x1b8
> > [ 2.507358] kernel_init_freeable+0x1ac/0x210
> > [ 2.511617] kernel_init+0x28/0x130
> > [ 2.516112] ret_from_fork+0x10/0x20
> >
> > Cc: <[email protected]>
> > Cc: [email protected]
> > Signed-off-by: Stefano Stabellini <[email protected]>
>
> For the immediate purpose as described this looks okay, so
> Reviewed-by: Jan Beulich <[email protected]>

Thank you!


> However, aren't there further pieces missing on this error patch:
> - clearing of xenstored_ready in case it got set,
> - rolling back xenstored_local_init() (XS_LOCAL) and xen_remap()
> (XS_HVM).
> And shouldn't xs_init() failure when called from xenbus_probe()
> also result in the driver not giving the appearance of being usable?

I prioritized this patch because I have a direct test case for this
issue and I can see that this patch solves it. But what you wrote is
true, and I can have a look in the following weeks.


> > --- a/drivers/xen/xenbus/xenbus_probe.c
> > +++ b/drivers/xen/xenbus/xenbus_probe.c
> > @@ -909,7 +909,7 @@ static struct notifier_block xenbus_resume_nb = {
> >
> > static int __init xenbus_init(void)
> > {
> > - int err = 0;
> > + int err;
> > uint64_t v = 0;
> > xen_store_domain_type = XS_UNKNOWN;
> >
>
> Minor remark: You may want to take the opportunity and add the
> missing blank line here to visually separate the assignment from
> the declarations.
>
> Jan
>

2021-11-17 16:06:23

by Boris Ostrovsky

[permalink] [raw]
Subject: Re: [PATCH v2] xen: don't continue xenstore initialization in case of errors


On 11/15/21 5:27 PM, Stefano Stabellini wrote:
> From: Stefano Stabellini <[email protected]>
>
> In case of errors in xenbus_init (e.g. missing xen_store_gfn parameter),
> we goto out_error but we forget to reset xen_store_domain_type to
> XS_UNKNOWN. As a consequence xenbus_probe_initcall and other initcalls
> will still try to initialize xenstore resulting into a crash at boot.
>
> [ 2.479830] Call trace:
> [ 2.482314] xb_init_comms+0x18/0x150
> [ 2.486354] xs_init+0x34/0x138
> [ 2.489786] xenbus_probe+0x4c/0x70
> [ 2.498432] xenbus_probe_initcall+0x2c/0x7c
> [ 2.503944] do_one_initcall+0x54/0x1b8
> [ 2.507358] kernel_init_freeable+0x1ac/0x210
> [ 2.511617] kernel_init+0x28/0x130
> [ 2.516112] ret_from_fork+0x10/0x20
>
> Cc: <[email protected]>
> Cc: [email protected]
> Signed-off-by: Stefano Stabellini <[email protected]>



Applied to for-linus-5.16c


-boris