Skip to contentSkip to conversation

tune preempt_[dis|en]able()

PATCH In review Juergen Gross 3 participants 20 messages May 22, 2019 – May 24, 2019 2 messages not in the archive

lore
Share & tools
lore
#0 lore raw

cover tune preempt_[dis|en]able()

There is no user of the preempt_count for non-debug builds. This series removes it for that case.

Juergen Gross (3):
xen: drop in_atomic()
xen: drop preempt_count() for non-debug builds
xen: build common/preempt.c only with CONFIG_DEBUG

 xen/arch/x86/hvm/hvm.c    | 16 ----------------
 xen/common/Makefile       |  2 +-
 xen/common/preempt.c      |  7 -------
 xen/include/xen/preempt.h |  8 +++++---
 4 files changed, 6 insertions(+), 27 deletions(-)
Signature · 7 lines
2.16.4


_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
#1 lore raw

PATCH 1/3 xen: drop in_atomic()

Currently there is only one user of in_atomic(), and that is in an #ifdef 0 section. This has been so since Xen 4.1, so chances are rather slim we suddenly want to enable it again.

Dropping in_atomic() will remove the last user of preempt_count() in non-debug builds enabling further optimizations.

---
 xen/arch/x86/hvm/hvm.c    | 16 ----------------
 xen/common/preempt.c      |  5 -----
 xen/include/xen/preempt.h |  2 --
 3 files changed, 23 deletions(-)
xen/arch/x86/hvm/hvm.c +0 -16
index 8993c2aa57..7b36765b97 100644
@@ -3185,22 +3185,6 @@ static enum hvm_translation_result __hvm_copy(
 
     ASSERT(is_hvm_vcpu(v));
 
-    /*
-     * XXX Disable for 4.1.0: PV-on-HVM drivers will do grant-table ops
-     * such as query_size. Grant-table code currently does copy_to/from_guest
-     * accesses under the big per-domain lock, which this test would disallow.
-     * The test is not needed until we implement sleeping-on-waitqueue when
-     * we access a paged-out frame, and that's post 4.1.0 now.
-     */
-#if 0
-    /*
-     * If the required guest memory is paged out, this function may sleep.
-     * Hence we bail immediately if called from atomic context.
-     */
-    if ( in_atomic() )
-        return HVMTRANS_unhandleable;
-#endif
-
     while ( todo > 0 )
     {
         enum hvm_translation_result res;
xen/common/preempt.c +0 -5
index 3b4178fd44..20913e20d3 100644
@@ -25,11 +25,6 @@
 
 DEFINE_PER_CPU(unsigned int, __preempt_count);
 
-bool_t in_atomic(void)
-{
-    return preempt_count() || in_irq() || !local_irq_is_enabled();
-}
-
 #ifndef NDEBUG
 void ASSERT_NOT_IN_ATOMIC(void)
 {
xen/include/xen/preempt.h +0 -2
index bef83135a1..f715ca09bc 100644
@@ -26,8 +26,6 @@ DECLARE_PER_CPU(unsigned int, __preempt_count);
     preempt_count()--;                          \
 } while (0)
 
-bool_t in_atomic(void);
-
 #ifndef NDEBUG
 void ASSERT_NOT_IN_ATOMIC(void);
 #else
Signature · 7 lines
2.16.4


_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Reviewed-by on v1 1/3
#2 lore raw

Re: [PATCH 1/3] xen: drop in_atomic()

Quoted 8 lines · Juergen Gross

Currently there is only one user of in_atomic(), and that is in an #ifdef 0 section. This has been so since Xen 4.1, so chances are rather slim we suddenly want to enable it again.

Dropping in_atomic() will remove the last user of preempt_count() in non-debug builds enabling further optimizations.

This code has never been used and is obviously not doing anything useful.

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#3 lore raw

Re: [PATCH 1/3] xen: drop in_atomic()

In reply to Juergen Gross
Quoted 12 lines
Quoted 1 line
Quoted 1 line · On 22.05.19 at 11:45, <[email protected]> wrote:
xen/arch/x86/hvm/hvm.c +0 -5
@@ -3185,22 +3185,6 @@ static enum hvm_translation_result __hvm_copy(
 
     ASSERT(is_hvm_vcpu(v));
 
-    /*
-     * XXX Disable for 4.1.0: PV-on-HVM drivers will do grant-table ops
-     * such as query_size. Grant-table code currently does copy_to/from_guest
-     * accesses under the big per-domain lock, which this test would disallow.
-     * The test is not needed until we implement sleeping-on-waitqueue when

xen/arch/x86/hvm/hvm.c static enum hvm_translation_result __hvm_copy( · line 3202

Quoted 10 lines
xen/arch/x86/hvm/hvm.c +0 -10
-     * we access a paged-out frame, and that's post 4.1.0 now.
-     */
-#if 0
-    /*
-     * If the required guest memory is paged out, this function may sleep.
-     * Hence we bail immediately if called from atomic context.
-     */
-    if ( in_atomic() )
-        return HVMTRANS_unhandleable;
-#endif

Dealing with this TODO item is of course much appreciated, but should it really be deleted altogether? The big-domain-lock issue is gone afair, in which case dropping the #if 0 would seem possible to me, even if it's not strictly needed without the sleep- on-waitqueue behavior mentioned.

Jan

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#4 lore raw

Re: [PATCH 1/3] xen: drop in_atomic()

Quoted 20 lines · Jan Beulich
Quoted 20 lines
Quoted 1 line
Quoted 1 line · On 22.05.19 at 11:45, <[email protected]> wrote:
xen/arch/x86/hvm/hvm.c +0 -13
@@ -3185,22 +3185,6 @@ static enum hvm_translation_result __hvm_copy(
 
     ASSERT(is_hvm_vcpu(v));
 
-    /*
-     * XXX Disable for 4.1.0: PV-on-HVM drivers will do grant-table ops
-     * such as query_size. Grant-table code currently does copy_to/from_guest
-     * accesses under the big per-domain lock, which this test would disallow.
-     * The test is not needed until we implement sleeping-on-waitqueue when
-     * we access a paged-out frame, and that's post 4.1.0 now.
-     */
-#if 0
-    /*
-     * If the required guest memory is paged out, this function may sleep.
-     * Hence we bail immediately if called from atomic context.
-     */
-    if ( in_atomic() )
Quoted 8 lines

xen/arch/x86/hvm/hvm.c static enum hvm_translation_result __hvm_copy( · line 3202

Quoted 2 lines
xen/arch/x86/hvm/hvm.c +0 -2
-        return HVMTRANS_unhandleable;
-#endif

Dealing with this TODO item is of course much appreciated, but should it really be deleted altogether? The big-domain-lock issue is gone afair, in which case dropping the #if 0 would seem possible to me, even if it's not strictly needed without the sleep- on-waitqueue behavior mentioned.

I just had a look and found the following path:

do_domctl() (takes domctl_lock and hypercall_deadlock_mutex)
arch_do_domctl()

    raw_copy_from_guest()
      copy_from_user_hvm()
        hvm_copy_from_guest_linear()
          __hvm_copy()

So no, we can't do the in_atomic() test IMO.

Juergen

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#5 lore raw

Re: [PATCH 1/3] xen: drop in_atomic()

Quoted 33 lines
Quoted 1 line
Quoted 1 line · On 24.05.19 at 07:41, <[email protected]> wrote:
Quoted 20 lines · Jan Beulich
Quoted 20 lines
Quoted 1 line
Quoted 1 line · On 22.05.19 at 11:45, <[email protected]> wrote:
xen/arch/x86/hvm/hvm.c +0 -13
@@ -3185,22 +3185,6 @@ static enum hvm_translation_result __hvm_copy(
 
     ASSERT(is_hvm_vcpu(v));
 
-    /*
-     * XXX Disable for 4.1.0: PV-on-HVM drivers will do grant-table ops
-     * such as query_size. Grant-table code currently does copy_to/from_guest
-     * accesses under the big per-domain lock, which this test would disallow.
-     * The test is not needed until we implement sleeping-on-waitqueue when
-     * we access a paged-out frame, and that's post 4.1.0 now.
-     */
-#if 0
-    /*
-     * If the required guest memory is paged out, this function may sleep.
-     * Hence we bail immediately if called from atomic context.
-     */
-    if ( in_atomic() )
Quoted 8 lines

xen/arch/x86/hvm/hvm.c static enum hvm_translation_result __hvm_copy( · line 3202

Quoted 2 lines
xen/arch/x86/hvm/hvm.c +0 -2
-        return HVMTRANS_unhandleable;
-#endif

Dealing with this TODO item is of course much appreciated, but should it really be deleted altogether? The big-domain-lock issue is gone afair, in which case dropping the #if 0 would seem possible to me, even if it's not strictly needed without the sleep- on-waitqueue behavior mentioned.

I just had a look and found the following path:

Quoted 8 lines

do_domctl() (takes domctl_lock and hypercall_deadlock_mutex)
arch_do_domctl()

    raw_copy_from_guest()
      copy_from_user_hvm()
        hvm_copy_from_guest_linear()
          __hvm_copy()

So no, we can't do the in_atomic() test IMO.

Oh, right - that's a PVH constraint that could probably not even be thought of that the time the comment was written. I'm still of the opinion though that at least the still applicable part of the comment should be kept in place. Whether this means also keeping in_atomic() itself is then an independent question, i.e. I wouldn't consider it overly bad if there was no implementation in the tree, but the above still served as documentation of what would need to be re-added. Still my preference would be for it to be kept.

Jan

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#6 lore raw

Re: [PATCH 1/3] xen: drop in_atomic()

Quoted 42 lines · Jan Beulich
Quoted 33 lines
Quoted 1 line
Quoted 1 line · On 24.05.19 at 07:41, <[email protected]> wrote:
Quoted 20 lines · Jan Beulich
Quoted 20 lines
Quoted 1 line
Quoted 1 line · On 22.05.19 at 11:45, <[email protected]> wrote:
xen/arch/x86/hvm/hvm.c +0 -13
@@ -3185,22 +3185,6 @@ static enum hvm_translation_result __hvm_copy(
 
     ASSERT(is_hvm_vcpu(v));
 
-    /*
-     * XXX Disable for 4.1.0: PV-on-HVM drivers will do grant-table ops
-     * such as query_size. Grant-table code currently does copy_to/from_guest
-     * accesses under the big per-domain lock, which this test would disallow.
-     * The test is not needed until we implement sleeping-on-waitqueue when
-     * we access a paged-out frame, and that's post 4.1.0 now.
-     */
-#if 0
-    /*
-     * If the required guest memory is paged out, this function may sleep.
-     * Hence we bail immediately if called from atomic context.
-     */
-    if ( in_atomic() )
Quoted 8 lines

xen/arch/x86/hvm/hvm.c static enum hvm_translation_result __hvm_copy( · line 3202

Quoted 2 lines
xen/arch/x86/hvm/hvm.c +0 -2
-        return HVMTRANS_unhandleable;
-#endif

Dealing with this TODO item is of course much appreciated, but should it really be deleted altogether? The big-domain-lock issue is gone afair, in which case dropping the #if 0 would seem possible to me, even if it's not strictly needed without the sleep- on-waitqueue behavior mentioned.

I just had a look and found the following path:

Quoted 8 lines

do_domctl() (takes domctl_lock and hypercall_deadlock_mutex)
arch_do_domctl()

    raw_copy_from_guest()
      copy_from_user_hvm()
        hvm_copy_from_guest_linear()
          __hvm_copy()

So no, we can't do the in_atomic() test IMO.

Quoted 9 lines

Oh, right - that's a PVH constraint that could probably not even be thought of that the time the comment was written. I'm still of the opinion though that at least the still applicable part of the comment should be kept in place. Whether this means also keeping in_atomic() itself is then an independent question, i.e. I wouldn't consider it overly bad if there was no implementation in the tree, but the above still served as documentation of what would need to be re-added. Still my preference would be for it to be kept.

Would you be okay with replacing the removed stuff above with:

/*
* If the required guest memory is paged out this function may sleep.
* So in theory we should bail out if called in atomic context.
* Unfortunately this is true for PVH dom0 doing domctl calls which
* holds the domctl lock when accessing dom0 memory. OTOH dom0 memory
* should never be paged out, so we are fine without testing for
* atomic context.
*/

Juergen

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#7 lore raw

Re: [PATCH 1/3] xen: drop in_atomic()

Quoted 54 lines
Quoted 1 line
Quoted 1 line · On 24.05.19 at 10:34, <[email protected]> wrote:
Quoted 42 lines · Jan Beulich
Quoted 33 lines
Quoted 1 line
Quoted 1 line · On 24.05.19 at 07:41, <[email protected]> wrote:
Quoted 20 lines · Jan Beulich
Quoted 20 lines
Quoted 1 line
Quoted 1 line · On 22.05.19 at 11:45, <[email protected]> wrote:
xen/arch/x86/hvm/hvm.c +0 -13
@@ -3185,22 +3185,6 @@ static enum hvm_translation_result __hvm_copy(
 
     ASSERT(is_hvm_vcpu(v));
 
-    /*
-     * XXX Disable for 4.1.0: PV-on-HVM drivers will do grant-table ops
-     * such as query_size. Grant-table code currently does copy_to/from_guest
-     * accesses under the big per-domain lock, which this test would disallow.
-     * The test is not needed until we implement sleeping-on-waitqueue when
-     * we access a paged-out frame, and that's post 4.1.0 now.
-     */
-#if 0
-    /*
-     * If the required guest memory is paged out, this function may sleep.
-     * Hence we bail immediately if called from atomic context.
-     */
-    if ( in_atomic() )
Quoted 8 lines

xen/arch/x86/hvm/hvm.c static enum hvm_translation_result __hvm_copy( · line 3202

Quoted 2 lines
xen/arch/x86/hvm/hvm.c +0 -2
-        return HVMTRANS_unhandleable;
-#endif

Dealing with this TODO item is of course much appreciated, but should it really be deleted altogether? The big-domain-lock issue is gone afair, in which case dropping the #if 0 would seem possible to me, even if it's not strictly needed without the sleep- on-waitqueue behavior mentioned.

I just had a look and found the following path:

Quoted 8 lines

do_domctl() (takes domctl_lock and hypercall_deadlock_mutex)
arch_do_domctl()

    raw_copy_from_guest()
      copy_from_user_hvm()
        hvm_copy_from_guest_linear()
          __hvm_copy()

So no, we can't do the in_atomic() test IMO.

Quoted 9 lines

Oh, right - that's a PVH constraint that could probably not even be thought of that the time the comment was written. I'm still of the opinion though that at least the still applicable part of the comment should be kept in place. Whether this means also keeping in_atomic() itself is then an independent question, i.e. I wouldn't consider it overly bad if there was no implementation in the tree, but the above still served as documentation of what would need to be re-added. Still my preference would be for it to be kept.

Quoted 6 lines

Would you be okay with replacing the removed stuff above with:

/*
* If the required guest memory is paged out this function may sleep.
* So in theory we should bail out if called in atomic context.
* Unfortunately this is true for PVH dom0 doing domctl calls which

... this is true at least for ...

Quoted 4 lines

* holds the domctl lock when accessing dom0 memory. OTOH dom0 memory
* should never be paged out, so we are fine without testing for
* atomic context.
*/

Not sure about this Dom0-specific remark: Are we certain there are no other paths, similar to the gnttab one having been mentioned till now?

Jan

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#8 lore raw

Re: [PATCH 1/3] xen: drop in_atomic()

Quoted 58 lines · Jan Beulich
Quoted 50 lines
Quoted 1 line
Quoted 1 line · On 24.05.19 at 10:34, <[email protected]> wrote:
Quoted 39 lines · Jan Beulich
Quoted 31 lines
Quoted 1 line
Quoted 1 line · On 24.05.19 at 07:41, <[email protected]> wrote:
Quoted 19 lines · Jan Beulich
Quoted 19 lines
Quoted 1 line
Quoted 1 line · On 22.05.19 at 11:45, <[email protected]> wrote:
xen/arch/x86/hvm/hvm.c +0 -12
@@ -3185,22 +3185,6 @@ static enum hvm_translation_result __hvm_copy(
 
     ASSERT(is_hvm_vcpu(v));
 
-    /*
-     * XXX Disable for 4.1.0: PV-on-HVM drivers will do grant-table ops
-     * such as query_size. Grant-table code currently does copy_to/from_guest
-     * accesses under the big per-domain lock, which this test would disallow.
-     * The test is not needed until we implement sleeping-on-waitqueue when
-     * we access a paged-out frame, and that's post 4.1.0 now.
-     */
-#if 0
-    /*
-     * If the required guest memory is paged out, this function may sleep.
-     * Hence we bail immediately if called from atomic context.
-     */
Quoted 8 lines

xen/arch/x86/hvm/hvm.c static enum hvm_translation_result __hvm_copy( · line 3202

Quoted 3 lines
xen/arch/x86/hvm/hvm.c +0 -3
-    if ( in_atomic() )
-        return HVMTRANS_unhandleable;
-#endif

Dealing with this TODO item is of course much appreciated, but should it really be deleted altogether? The big-domain-lock issue is gone afair, in which case dropping the #if 0 would seem possible to me, even if it's not strictly needed without the sleep- on-waitqueue behavior mentioned.

I just had a look and found the following path:

Quoted 8 lines

do_domctl() (takes domctl_lock and hypercall_deadlock_mutex)
arch_do_domctl()

    raw_copy_from_guest()
      copy_from_user_hvm()
        hvm_copy_from_guest_linear()
          __hvm_copy()

So no, we can't do the in_atomic() test IMO.

Quoted 9 lines

Oh, right - that's a PVH constraint that could probably not even be thought of that the time the comment was written. I'm still of the opinion though that at least the still applicable part of the comment should be kept in place. Whether this means also keeping in_atomic() itself is then an independent question, i.e. I wouldn't consider it overly bad if there was no implementation in the tree, but the above still served as documentation of what would need to be re-added. Still my preference would be for it to be kept.

Quoted 6 lines

Would you be okay with replacing the removed stuff above with:

/*
* If the required guest memory is paged out this function may sleep.
* So in theory we should bail out if called in atomic context.
* Unfortunately this is true for PVH dom0 doing domctl calls which

... this is true at least for ...

Quoted 7 lines
Quoted 4 lines

* holds the domctl lock when accessing dom0 memory. OTOH dom0 memory
* should never be paged out, so we are fine without testing for
* atomic context.
*/

Not sure about this Dom0-specific remark: Are we certain there are no other paths, similar to the gnttab one having been mentioned till now?

Why is __hvm_copy() so special?  It is just one of many places which can end up touching guest memory.

A comment here isn't going to help anyone who might find themselves with problems.

Given that the test has never been used, and no issues have been raised, and this path isn't AFAICT special, I don't see why it should be special-cased.

~Andrew

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#9 lore raw

Re: [PATCH 1/3] xen: drop in_atomic()

Quoted 66 lines
Quoted 1 line
Quoted 1 line · On 24.05.19 at 14:30, <[email protected]> wrote:
Quoted 9 lines · Jan Beulich
Quoted 9 lines
Quoted 1 line
Quoted 1 line · On 24.05.19 at 10:34, <[email protected]> wrote:
Quoted 7 lines · Jan Beulich
Quoted 7 lines
Quoted 1 line
Quoted 1 line · On 24.05.19 at 07:41, <[email protected]> wrote:
Quoted 5 lines · Jan Beulich
Quoted 5 lines
Quoted 1 line
Quoted 1 line · On 22.05.19 at 11:45, <[email protected]> wrote:
xen/arch/x86/hvm/hvm.c +0 -0
@@ -3185,22 +3185,6 @@ static enum hvm_translation_result __hvm_copy(
 
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines

xen/arch/x86/hvm/hvm.c static enum hvm_translation_result __hvm_copy( · line 3190

Quoted 5 lines
xen/arch/x86/hvm/hvm.c +0 -3
     ASSERT(is_hvm_vcpu(v));
 
-    /*
-     * XXX Disable for 4.1.0: PV-on-HVM drivers will do grant-table ops
-     * such as query_size. Grant-table code currently does 

copy_to/from_guest

Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

xen/arch/x86/hvm/hvm.c static enum hvm_translation_result __hvm_copy( · line 3191

Quoted 1 line
xen/arch/x86/hvm/hvm.c +0 -1
-     * accesses under the big per-domain lock, which this test would 

disallow.

Quoted 47 lines
Quoted 35 lines
Quoted 26 lines
Quoted 18 lines
Quoted 8 lines
Quoted 8 lines
xen/arch/x86/hvm/hvm.c +0 -8
-     * The test is not needed until we implement sleeping-on-waitqueue when
-     * we access a paged-out frame, and that's post 4.1.0 now.
-     */
-#if 0
-    /*
-     * If the required guest memory is paged out, this function may sleep.
-     * Hence we bail immediately if called from atomic context.
-     */
Quoted 8 lines

xen/arch/x86/hvm/hvm.c static enum hvm_translation_result __hvm_copy( · line 3202

Quoted 3 lines
xen/arch/x86/hvm/hvm.c +0 -3
-    if ( in_atomic() )
-        return HVMTRANS_unhandleable;
-#endif

Dealing with this TODO item is of course much appreciated, but should it really be deleted altogether? The big-domain-lock issue is gone afair, in which case dropping the #if 0 would seem possible to me, even if it's not strictly needed without the sleep- on-waitqueue behavior mentioned.

I just had a look and found the following path:

Quoted 8 lines

do_domctl() (takes domctl_lock and hypercall_deadlock_mutex)
arch_do_domctl()

    raw_copy_from_guest()
      copy_from_user_hvm()
        hvm_copy_from_guest_linear()
          __hvm_copy()

So no, we can't do the in_atomic() test IMO.

Quoted 9 lines

Oh, right - that's a PVH constraint that could probably not even be thought of that the time the comment was written. I'm still of the opinion though that at least the still applicable part of the comment should be kept in place. Whether this means also keeping in_atomic() itself is then an independent question, i.e. I wouldn't consider it overly bad if there was no implementation in the tree, but the above still served as documentation of what would need to be re-added. Still my preference would be for it to be kept.

Quoted 6 lines

Would you be okay with replacing the removed stuff above with:

/*
* If the required guest memory is paged out this function may sleep.
* So in theory we should bail out if called in atomic context.
* Unfortunately this is true for PVH dom0 doing domctl calls which

... this is true at least for ...

Quoted 4 lines

* holds the domctl lock when accessing dom0 memory. OTOH dom0 memory
* should never be paged out, so we are fine without testing for
* atomic context.
*/

Quoted 6 lines
Quoted 3 lines

Not sure about this Dom0-specific remark: Are we certain there are no other paths, similar to the gnttab one having been mentioned till now?

Why is __hvm_copy() so special? It is just one of many places which can end up touching guest memory.

Are you sure? I think everything that can touch guest (HVM) memory actually ends up calling into this function.

Jan

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

A message that is not in the archive (sent off-list or not yet received): [email protected]

#11 lore raw

Re: [PATCH 1/3] xen: drop in_atomic()

Quoted 20 lines · Jan Beulich
Quoted 20 lines
Quoted 1 line
Quoted 1 line · On 22.05.19 at 11:45, <[email protected]> wrote:
xen/arch/x86/hvm/hvm.c +0 -13
@@ -3185,22 +3185,6 @@ static enum hvm_translation_result __hvm_copy(
 
     ASSERT(is_hvm_vcpu(v));
 
-    /*
-     * XXX Disable for 4.1.0: PV-on-HVM drivers will do grant-table ops
-     * such as query_size. Grant-table code currently does copy_to/from_guest
-     * accesses under the big per-domain lock, which this test would disallow.
-     * The test is not needed until we implement sleeping-on-waitqueue when
-     * we access a paged-out frame, and that's post 4.1.0 now.
-     */
-#if 0
-    /*
-     * If the required guest memory is paged out, this function may sleep.
-     * Hence we bail immediately if called from atomic context.
-     */
-    if ( in_atomic() )
Quoted 8 lines

xen/arch/x86/hvm/hvm.c static enum hvm_translation_result __hvm_copy( · line 3202

Quoted 2 lines
xen/arch/x86/hvm/hvm.c +0 -2
-        return HVMTRANS_unhandleable;
-#endif

Dealing with this TODO item is of course much appreciated, but should it really be deleted altogether? The big-domain-lock issue is gone afair, in which case dropping the #if 0 would seem possible to me, even if it's not strictly needed without the sleep- on-waitqueue behavior mentioned.

Question is whether it is worth to keep it resulting in the need to keep preempt_count() as well.

Juergen

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#12 lore raw

Re: [PATCH 1/3] xen: drop in_atomic()

Quoted 33 lines
Quoted 1 line
Quoted 1 line · On 22.05.19 at 12:19, <[email protected]> wrote:
Quoted 10 lines · Jan Beulich

xen/arch/x86/hvm/hvm.c static enum hvm_translation_result __hvm_copy( · line 3190

Quoted 10 lines
Quoted 1 line
Quoted 1 line · On 22.05.19 at 11:45, <[email protected]> wrote:
xen/arch/x86/hvm/hvm.c +0 -3
@@ -3185,22 +3185,6 @@ static enum hvm_translation_result __hvm_copy(
 
     ASSERT(is_hvm_vcpu(v));
 
-    /*
-     * XXX Disable for 4.1.0: PV-on-HVM drivers will do grant-table ops
-     * such as query_size. Grant-table code currently does 

copy_to/from_guest

Quoted 1 line

xen/arch/x86/hvm/hvm.c static enum hvm_translation_result __hvm_copy( · line 3191

Quoted 1 line
xen/arch/x86/hvm/hvm.c +0 -1
-     * accesses under the big per-domain lock, which this test would 

disallow.

Quoted 9 lines
Quoted 9 lines
xen/arch/x86/hvm/hvm.c +0 -9
-     * The test is not needed until we implement sleeping-on-waitqueue when
-     * we access a paged-out frame, and that's post 4.1.0 now.
-     */
-#if 0
-    /*
-     * If the required guest memory is paged out, this function may sleep.
-     * Hence we bail immediately if called from atomic context.
-     */
-    if ( in_atomic() )
Quoted 8 lines

xen/arch/x86/hvm/hvm.c static enum hvm_translation_result __hvm_copy( · line 3202

Quoted 2 lines
xen/arch/x86/hvm/hvm.c +0 -2
-        return HVMTRANS_unhandleable;
-#endif

Dealing with this TODO item is of course much appreciated, but should it really be deleted altogether? The big-domain-lock issue is gone afair, in which case dropping the #if 0 would seem possible to me, even if it's not strictly needed without the sleep- on-waitqueue behavior mentioned.

Quoted 2 lines

Question is whether it is worth to keep it resulting in the need to keep preempt_count() as well.

Well, personally I think keeping it is a small price to pay. But seeing Andrew's R-b he clearly thinks different. And just to be clear - I don't really want to veto this change, as at the same time it's also easy enough to put back if need be. But I'd like this to be give a 2nd consideration at least.

Jan

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#13 lore raw

PATCH 2/3 xen: drop preempt_count() for non-debug builds

In reply to Juergen Gross

preempt_count() and the associated per-cpu variable __preempt_count are tested in debug build only. So drop them for non-debug builds.

---
 xen/common/preempt.c      | 2 +-
 xen/include/xen/preempt.h | 6 +++++-
 2 files changed, 6 insertions(+), 2 deletions(-)
xen/common/preempt.c +1 -1
index 20913e20d3..3077c51d52 100644
@@ -23,9 +23,9 @@
 #include <xen/irq.h>
 #include <asm/system.h>
 
+#ifndef NDEBUG
 DEFINE_PER_CPU(unsigned int, __preempt_count);
 
-#ifndef NDEBUG
 void ASSERT_NOT_IN_ATOMIC(void)
 {
     ASSERT(!preempt_count());
xen/include/xen/preempt.h +5 -1
index f715ca09bc..0bf49cc979 100644
@@ -12,6 +12,8 @@
 #include <xen/types.h>
 #include <xen/percpu.h>
 
+#ifndef NDEBUG
+
 DECLARE_PER_CPU(unsigned int, __preempt_count);
 
 #define preempt_count() (this_cpu(__preempt_count))
@@ -26,9 +28,11 @@ DECLARE_PER_CPU(unsigned int, __preempt_count);
     preempt_count()--;                          \
 } while (0)
 
-#ifndef NDEBUG
 void ASSERT_NOT_IN_ATOMIC(void);
+
 #else
+#define preempt_disable()    barrier();
+#define preempt_enable()     barrier();
 #define ASSERT_NOT_IN_ATOMIC() ((void)0)
 #endif
 
Signature · 7 lines
2.16.4


_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Reviewed-by on v1 2/3
#14 lore raw

Re: [PATCH 2/3] xen: drop preempt_count() for non-debug builds

Quoted 4 lines · Juergen Gross

preempt_count() and the associated per-cpu variable __preempt_count are tested in debug build only. So drop them for non-debug builds.

I'd be tempted to fold patches 2 and 3 together, because they are both the same change, and it would reduce the churn.

two folded into one.

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#15 lore raw

Re: [PATCH 2/3] xen: drop preempt_count() for non-debug builds

Quoted 8 lines · Andrew Cooper
Quoted 4 lines · Juergen Gross

preempt_count() and the associated per-cpu variable __preempt_count are tested in debug build only. So drop them for non-debug builds.

I'd be tempted to fold patches 2 and 3 together, because they are both the same change, and it would reduce the churn.

Fine with me.

Quoted 2 lines

two folded into one.

Thanks,

Juergen

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#16 lore raw

Re: [PATCH 2/3] xen: drop preempt_count() for non-debug builds

In reply to Andrew Cooper
Quoted 5 lines
Quoted 1 line
Quoted 1 line · On 22.05.19 at 12:00, <[email protected]> wrote:
Quoted 3 lines · Juergen Gross

preempt_count() and the associated per-cpu variable __preempt_count are tested in debug build only. So drop them for non-debug builds.

Quoted 7 lines
Quoted 1 line

I'd be tempted to fold patches 2 and 3 together, because they are both the same change, and it would reduce the churn.

two folded into one.

I'm a little surprised by this: Wasn't it you who generally wanted what ASSERT() expands to (controlled by NDEBUG) be independent of CONFIG_DEBUG, at some point down the road? Aren't you even having ASSERT()s enabled in release builds of XenServer, or am I misremembering? If so patch 3 would move us in the wrong direction.

Jan

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#17 lore raw

Re: [PATCH 2/3] xen: drop preempt_count() for non-debug builds

Quoted 9 lines · Jan Beulich
Quoted 9 lines
Quoted 1 line
Quoted 1 line · On 22.05.19 at 12:00, <[email protected]> wrote:
Quoted 4 lines · Juergen Gross

preempt_count() and the associated per-cpu variable __preempt_count are tested in debug build only. So drop them for non-debug builds.

I'd be tempted to fold patches 2 and 3 together, because they are both the same change, and it would reduce the churn.

Quoted 6 lines
Quoted 2 lines

two folded into one.

I'm a little surprised by this: Wasn't it you who generally wanted what ASSERT() expands to (controlled by NDEBUG) be independent of CONFIG_DEBUG, at some point down the road?

In some ideal world yes, but what is rather more important is actually having optimisation control unrelated to NDEBUG.

Quoted 3 lines

Aren't you even having ASSERT()s enabled in release builds of XenServer, or am I misremembering? If so patch 3 would move us in the wrong direction.

We build and ship a debug and a release hypervisor.

~Andrew

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#18 lore raw

Re: [PATCH 2/3] xen: drop preempt_count() for non-debug builds

In reply to Juergen Gross

xen/include/xen/preempt.h DECLARE_PER_CPU(unsigned int, __preempt_count); · line 35

Quoted 11 lines
Quoted 1 line
Quoted 1 line · On 22.05.19 at 11:45, <[email protected]> wrote:
xen/include/xen/preempt.h +3 -1
@@ -26,9 +28,11 @@ DECLARE_PER_CPU(unsigned int, __preempt_count);
     preempt_count()--;                          \
 } while (0)
 
-#ifndef NDEBUG
 void ASSERT_NOT_IN_ATOMIC(void);
+
 #else
+#define preempt_disable()    barrier();
+#define preempt_enable()     barrier();

Stray semicolons (could be dropped while committing if we really want to go this route).

Jan

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

A message that is not in the archive (sent off-list or not yet received): [email protected]

#20 lore raw

Re: [PATCH 2/3] xen: drop preempt_count() for non-debug builds

Quoted 5 lines · Jan Beulich
Quoted 5 lines
Quoted 1 line
Quoted 1 line · On 22.05.19 at 11:45, <[email protected]> wrote:
xen/include/xen/preempt.h +0 -0
@@ -26,9 +28,11 @@ DECLARE_PER_CPU(unsigned int, __preempt_count);
     preempt_count()--;                          \
 } while (0)
 
Quoted 9 lines

xen/include/xen/preempt.h DECLARE_PER_CPU(unsigned int, __preempt_count); · line 35

Quoted 6 lines
xen/include/xen/preempt.h +3 -1
-#ifndef NDEBUG
 void ASSERT_NOT_IN_ATOMIC(void);
+
 #else
+#define preempt_disable()    barrier();
+#define preempt_enable()     barrier();

Stray semicolons (could be dropped while committing if we really want to go this route).

Oh, right.

Will correct in V2.

Juergen

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#21 lore raw

PATCH 3/3 xen: build common/preempt.c only with CONFIG_DEBUG

In reply to Juergen Gross

There is nothing left in common/preempt.c in non-debug builds.

---
 xen/common/Makefile  | 2 +-
 xen/common/preempt.c | 2 --
 2 files changed, 1 insertion(+), 3 deletions(-)
xen/common/Makefile +1 -1
index 33d03b862f..d6576a3fc3 100644
@@ -30,7 +30,7 @@ obj-y += notifier.o
 obj-y += page_alloc.o
 obj-$(CONFIG_HAS_PDX) += pdx.o
 obj-$(CONFIG_PERF_COUNTERS) += perfc.o
-obj-y += preempt.o
+obj-$(CONFIG_DEBUG) += preempt.o
 obj-y += random.o
 obj-y += rangeset.o
 obj-y += radix-tree.o
xen/common/preempt.c +0 -2
index 3077c51d52..8bb099e8e8 100644
@@ -23,7 +23,6 @@
 #include <xen/irq.h>
 #include <asm/system.h>
 
-#ifndef NDEBUG
 DEFINE_PER_CPU(unsigned int, __preempt_count);
 
 void ASSERT_NOT_IN_ATOMIC(void)
@@ -32,4 +31,3 @@ void ASSERT_NOT_IN_ATOMIC(void)
     ASSERT(!in_irq());
     ASSERT(local_irq_is_enabled());
 }
-#endif
Signature · 7 lines
2.16.4


_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel