Skip to contentSkip to conversation

common: XSA-327 follow-up

PATCH In review Jan Beulich 2 participants 8 messages Dec 22, 2020 – Apr 29, 2021

lore
Share & tools
lore
#0 lore raw

v2 cover common: XSA-327 follow-up

There are a few largely cosmetic things that were discussed in the context of the XSA, but which weren't really XSA material.

1: common: map_vcpu_info() cosmetics
2: evtchn/fifo: don't enforce higher than necessary alignment

I realize both changes are controversial. For the first one discussion was about the choice of error code. Neither EINVAL nor EFAULT represent the fact that it is a choice of implementation to not support mis-aligned structures. If ENXIO isn't liked, the best I can suggest are EOPNOTSUPP or (as previously suggested) EPERM. I think it ought to be possible to settle on an error code everyone can live with.

For the second one the question was whether the relaxation is useful to do. The original reason for wanting to make the change
remains: The original code here should not be used as an excuse to introduce similar over-alignment requirements elsewhere. I can live with the change getting rejected, but if so I'd like to request that some alternative be submitted to ensure that the immediate goal can still be reached.

Jan

#1 lore raw

PATCH v2 1/2 common: map_vcpu_info() cosmetics

Use ENXIO instead of EINVAL to cover the two cases of the address not satisfying the requirements. This will make an issue here better stand out at the call site.

Also add a missing compat-mode related size check: If the sizes differed, other code in the function would need changing. Accompany this by a change to the initial sizeof() expression, tying it to the type of the variable we're actually after (matching e.g. the alignof() added by XSA-327).

xen/common/domain.c +4 -3
@@ -1241,17 +1241,18 @@ int map_vcpu_info(struct vcpu *v, unsign
     struct page_info *page;
     unsigned int align;
 
-    if ( offset > (PAGE_SIZE - sizeof(vcpu_info_t)) )
-        return -EINVAL;
+    if ( offset > (PAGE_SIZE - sizeof(*new_info)) )
+        return -ENXIO;
 
 #ifdef CONFIG_COMPAT
+    BUILD_BUG_ON(sizeof(*new_info) != sizeof(new_info->compat));
     if ( has_32bit_shinfo(d) )
         align = alignof(new_info->compat);
     else
 #endif
         align = alignof(*new_info);
     if ( offset & (align - 1) )
-        return -EINVAL;
+        return -ENXIO;
 
     if ( !mfn_eq(v->vcpu_info_mfn, INVALID_MFN) )
         return -EINVAL;
Acked-by on v2 1/2
#2 lore raw

Re: [PATCH v2 1/2] common: map_vcpu_info() cosmetics

Hi Jan,

Quoted 4 lines · Jan Beulich

Use ENXIO instead of EINVAL to cover the two cases of the address not satisfying the requirements. This will make an issue here better stand out at the call site.

Quoted 8 lines

Also add a missing compat-mode related size check: If the sizes differed, other code in the function would need changing. Accompany this by a change to the initial sizeof() expression, tying it to the type of the variable we're actually after (matching e.g. the alignof() added by XSA-327).

Cheers,

Signature · 1 line
Julien Grall
#3 lore raw

PATCH v2 2/2 evtchn/fifo: don't enforce higher than necessary alignment

In reply to Jan Beulich

Neither the code nor the original commit provide any justification for the need to 8-byte align the struct in all cases. Enforce just as much alignment as the structure actually needs - 4 bytes - by using alignof() instead of a literal number.

While relaxation of the requirements is intended here, the primary goal is to simply get rid of the hard coded number as well its lack of connection to the structure that is is meant to apply to.

Take the opportunity and also
- add so far missing validation that native and compat mode layouts of the structures actually match,
- tie sizeof() expressions to the types of the fields we're actually after, rather than specifying the type explicitly (which in the general case risks a disconnect, even if there's close to zero risk in this particular case),
- use ENXIO instead of EINVAL for the two cases of the address not satisfying the requirements, which will make an issue here better stand out at the call site.

---
v2: Add comment to public header. Re-base.
---
I question the need for the array_index_nospec() here. Or else I'd expect map_vcpu_info() would also need the same.

xen/common/event_fifo.c +17 -6
@@ -567,6 +567,16 @@ static void setup_ports(struct domain *d
     }
 }
 
+#ifdef CONFIG_COMPAT
+
+#include <compat/event_channel.h>
+
+#define xen_evtchn_fifo_control_block evtchn_fifo_control_block
+CHECK_evtchn_fifo_control_block;
+#undef xen_evtchn_fifo_control_block
+
+#endif
+
 int evtchn_fifo_init_control(struct evtchn_init_control *init_control)
 {
     struct domain *d = current->domain;
@@ -586,19 +596,20 @@ int evtchn_fifo_init_control(struct evtc
         return -ENOENT;
 
     /* Must not cross page boundary. */
-    if ( offset > (PAGE_SIZE - sizeof(evtchn_fifo_control_block_t)) )
-        return -EINVAL;
+    if ( offset > (PAGE_SIZE - sizeof(*v->evtchn_fifo->control_block)) )
+        return -ENXIO;
 
     /*
      * Make sure the guest controlled value offset is bounded even during
      * speculative execution.
      */
     offset = array_index_nospec(offset,
-                           PAGE_SIZE - sizeof(evtchn_fifo_control_block_t) + 1);
+                                PAGE_SIZE -
+                                sizeof(*v->evtchn_fifo->control_block) + 1);
 
-    /* Must be 8-bytes aligned. */
-    if ( offset & (8 - 1) )
-        return -EINVAL;
+    /* Must be suitably aligned. */
+    if ( offset & (alignof(*v->evtchn_fifo->control_block) - 1) )
+        return -ENXIO;
 
     spin_lock(&d->event_lock);
 
xen/include/public/event_channel.h +5 -0
@@ -368,6 +368,11 @@ typedef uint32_t event_word_t;
 
 #define EVTCHN_FIFO_NR_CHANNELS (1 << EVTCHN_FIFO_LINK_BITS)
 
+/*
+ * While this structure only requires 4-byte alignment, Xen versions 4.14 and
+ * earlier reject offset values (in struct evtchn_init_control) that aren't a
+ * multiple of 8.
+ */
 struct evtchn_fifo_control_block {
     uint32_t ready;
     uint32_t _rsvd;
xen/include/xlat.lst +1 -0
@@ -67,6 +67,7 @@
 ?	evtchn_bind_virq		event_channel.h
 ?	evtchn_close			event_channel.h
 ?	evtchn_expand_array		event_channel.h
+?	evtchn_fifo_control_block	event_channel.h
 ?	evtchn_init_control		event_channel.h
 ?	evtchn_op			event_channel.h
 ?	evtchn_reset			event_channel.h
#4 lore raw

PATCH v3 1/1 evtchn/fifo: don't enforce higher than necessary alignment

In reply to Jan Beulich

Neither the code nor the original commit provide any justification for the need to 8-byte align the struct in all cases. Enforce just as much alignment as the structure actually needs - 4 bytes - by using alignof() instead of a literal number.

While relaxation of the requirements is intended here, the primary goal is to simply get rid of the hard coded number as well its lack of connection to the structure that is is meant to apply to.

Take the opportunity and also
- add so far missing validation that native and compat mode layouts of the structures actually match,
- tie sizeof() expressions to the types of the fields we're actually after, rather than specifying the type explicitly (which in the general case risks a disconnect, even if there's close to zero risk in this particular case),
- use ENXIO instead of EINVAL for the two cases of the address not satisfying the requirements, which will make an issue here better stand out at the call site.

---
v3: Adjust version in public header comment.
v2: Add comment to public header. Re-base.
---
I question the need for the array_index_nospec() here. Or else I'd expect map_vcpu_info() would also need the same.

xen/common/event_fifo.c +17 -6
@@ -567,6 +567,16 @@ static void setup_ports(struct domain *d
     }
 }
 
+#ifdef CONFIG_COMPAT
+
+#include <compat/event_channel.h>
+
+#define xen_evtchn_fifo_control_block evtchn_fifo_control_block
+CHECK_evtchn_fifo_control_block;
+#undef xen_evtchn_fifo_control_block
+
+#endif
+
 int evtchn_fifo_init_control(struct evtchn_init_control *init_control)
 {
     struct domain *d = current->domain;
@@ -586,19 +596,20 @@ int evtchn_fifo_init_control(struct evtc
         return -ENOENT;
 
     /* Must not cross page boundary. */
-    if ( offset > (PAGE_SIZE - sizeof(evtchn_fifo_control_block_t)) )
-        return -EINVAL;
+    if ( offset > (PAGE_SIZE - sizeof(*v->evtchn_fifo->control_block)) )
+        return -ENXIO;
 
     /*
      * Make sure the guest controlled value offset is bounded even during
      * speculative execution.
      */
     offset = array_index_nospec(offset,
-                           PAGE_SIZE - sizeof(evtchn_fifo_control_block_t) + 1);
+                                PAGE_SIZE -
+                                sizeof(*v->evtchn_fifo->control_block) + 1);
 
-    /* Must be 8-bytes aligned. */
-    if ( offset & (8 - 1) )
-        return -EINVAL;
+    /* Must be suitably aligned. */
+    if ( offset & (alignof(*v->evtchn_fifo->control_block) - 1) )
+        return -ENXIO;
 
     spin_lock(&d->event_lock);
 
xen/include/public/event_channel.h +5 -0
@@ -368,6 +368,11 @@ typedef uint32_t event_word_t;
 
 #define EVTCHN_FIFO_NR_CHANNELS (1 << EVTCHN_FIFO_LINK_BITS)
 
+/*
+ * While this structure only requires 4-byte alignment, Xen versions 4.15 and
+ * earlier reject offset values (in struct evtchn_init_control) that aren't a
+ * multiple of 8.
+ */
 struct evtchn_fifo_control_block {
     uint32_t ready;
     uint32_t _rsvd;
xen/include/xlat.lst +1 -0
@@ -67,6 +67,7 @@
 ?	evtchn_bind_virq		event_channel.h
 ?	evtchn_close			event_channel.h
 ?	evtchn_expand_array		event_channel.h
+?	evtchn_fifo_control_block	event_channel.h
 ?	evtchn_init_control		event_channel.h
 ?	evtchn_op			event_channel.h
 ?	evtchn_reset			event_channel.h
#5 lore raw

Re: [PATCH v3] evtchn/fifo: don't enforce higher than necessary alignment

Hi,

Quoted 4 lines · Jan Beulich

Neither the code nor the original commit provide any justification for the need to 8-byte align the struct in all cases. Enforce just as much alignment as the structure actually needs - 4 bytes - by using alignof() instead of a literal number.

I had another fresh look today at this patch. The 32-bit padding is right after the field 'ready'.

I can't for sure tell how the second half is going to ever be used and how.

However, one possibility would be to extend the field 'ready' to 64-bit. With the current code, we could easily make a single 64-bit access without having to know whether the guest is able to interpret the top half.

With your approach, we may need to have different path depending on the padding and ensure the new extension cannot be enabled if the padding is 4-byte. Otherwise, the atomicity would be broken.

Quoted 3 lines

While relaxation of the requirements is intended here, the primary goal is to simply get rid of the hard coded number as well its lack of connection to the structure that is is meant to apply to.

Based on what I wrote above, I think the relaxation should not be done to give us more flexibility about possible extension to the structure.

Although, I would be worth documenting the reasoning in the code.

Cheers,

Signature · 1 line
Julien Grall
#6 lore raw

Re: [PATCH v3] evtchn/fifo: don't enforce higher than necessary alignment

Quoted 8 lines · Julien Grall

Hi,

Quoted 4 lines · Jan Beulich

Neither the code nor the original commit provide any justification for the need to 8-byte align the struct in all cases. Enforce just as much alignment as the structure actually needs - 4 bytes - by using alignof() instead of a literal number.

Quoted 8 lines

I had another fresh look today at this patch. The 32-bit padding is right after the field 'ready'.

I can't for sure tell how the second half is going to ever be used and how.

However, one possibility would be to extend the field 'ready' to 64-bit. With the current code, we could easily make a single 64-bit access without having to know whether the guest is able to interpret the top half.

I don't think extending field sizes is generally to be considered ABI- compatible. I also don't think we can re-use the field at all, as I couldn't find any checking of it being zero (input) or it getting set to zero (output). struct evtchn_init_control, which in principle could be a way to convey respective controlling flags, similarly has no room for extension, as its _pad[] also doesn't look to get checked anywhere.

Jan

Quoted 15 lines

With your approach, we may need to have different path depending on the padding and ensure the new extension cannot be enabled if the padding is 4-byte. Otherwise, the atomicity would be broken.

Quoted 3 lines

While relaxation of the requirements is intended here, the primary goal is to simply get rid of the hard coded number as well its lack of connection to the structure that is is meant to apply to.

Based on what I wrote above, I think the relaxation should not be done to give us more flexibility about possible extension to the structure.

Although, I would be worth documenting the reasoning in the code.

Cheers,

#7 lore raw

Re: [PATCH v3] evtchn/fifo: don't enforce higher than necessary alignment

Hi Jan,

Quoted 14 lines · Jan Beulich
Quoted 13 lines · Julien Grall

Hi,

Quoted 4 lines · Jan Beulich

Neither the code nor the original commit provide any justification for the need to 8-byte align the struct in all cases. Enforce just as much alignment as the structure actually needs - 4 bytes - by using alignof() instead of a literal number.

I had another fresh look today at this patch. The 32-bit padding is right after the field 'ready'.

I can't for sure tell how the second half is going to ever be used and how.

Quoted 8 lines
Quoted 3 lines

However, one possibility would be to extend the field 'ready' to 64-bit. With the current code, we could easily make a single 64-bit access without having to know whether the guest is able to interpret the top half.

I don't think extending field sizes is generally to be considered ABI- compatible. I also don't think we can re-use the field at all, as I couldn't find any checking of it being zero (input) or it getting set to zero (output).

That's would be fine so long we have a flag to control it. We can still write unconditionally because a guest can't rely on the pad...

Quoted 3 lines

struct evtchn_init_control, which in principle could
be a way to convey respective controlling flags, similarly has no room for extension, as its _pad[] also doesn't look to get checked anywhere.

Right, we would need to have a different way to convey. Yet, I am still unconvinced of the benefits change offer in this patch.

I am not going to Nack. So another maintainer in "THE REST" can express support for your patch and ack it.

Cheers,

Signature · 1 line
Julien Grall