Skip to contentSkip to conversation

xen/arm: vgic-v3: fix GICD_ISACTIVER range

PATCH RESEND In review Stefano Stabellini 3 participants 6 messages Apr 17, 2020 – Apr 21, 2020

lore
Share & tools
lore
#0 lore raw

PATCH RESEND 1/1 xen/arm: vgic-v3: fix GICD_ISACTIVER range

From: Peng Fan <[email protected]>

The end should be GICD_ISACTIVERN not GICD_ISACTIVER.

(See https://marc.info/?l=xen-devel&m=158527653730795 for a discussion on what it would take to implement GICD_ISACTIVER/GICD_ICACTIVER properly.)

---
 xen/arch/arm/vgic-v3.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
xen/arch/arm/vgic-v3.c +1 -1
index 4e60ba15cc..fd8cfc156d 100644
@@ -713,7 +713,7 @@ static int __vgic_v3_distr_common_mmio_read(const char *name, struct vcpu *v,
         goto read_as_zero;
 
     /* Read the active status of an IRQ via GICD/GICR is not supported */
-    case VRANGE32(GICD_ISACTIVER, GICD_ISACTIVER):
+    case VRANGE32(GICD_ISACTIVER, GICD_ISACTIVERN):
     case VRANGE32(GICD_ICACTIVER, GICD_ICACTIVERN):
         goto read_as_zero;
 
Signature · 1 line
2.17.1
#1 lore raw

Hi,

The title claim this is a resend, but this is actually not the latest version of this patch. Can you explain why you decided to use v1 rather than v2?

Quoted 11 lines · Stefano Stabellini

From: Peng Fan <[email protected]>

The end should be GICD_ISACTIVERN not GICD_ISACTIVER.

(See https://marc.info/?l=xen-devel&m=158527653730795 for a discussion on what it would take to implement GICD_ISACTIVER/GICD_ICACTIVER properly.)

I don't think you can be at the same time an author of the patch and a reviewer. Otherwise, I could review my own patch ;).

Cheers,

#2 lore raw
Quoted 5 lines · Julien Grall

Hi,

The title claim this is a resend, but this is actually not the latest version of this patch. Can you explain why you decided to use v1 rather than v2?

Because v2 added a printk for every read, and I thought you only wanted the range fix. Also, the printk is not a "warn once" so after the discussion where it became apparent that the register can be read many times, it sounded undesirable.

Nonetheless I don't have a strong preference between the two. If you prefer v2 it is here:

https://marc.info/?l=xen-devel&m=157440872522065

Do you need me to resent it? If it is OK for you as-is, you can give your ack here, and I'll go ahead and commit it.

Quoted 9 lines · Stefano Stabellini
Quoted 8 lines

From: Peng Fan <[email protected]>

The end should be GICD_ISACTIVERN not GICD_ISACTIVER.

(See https://marc.info/?l=xen-devel&m=158527653730795 for a discussion on what it would take to implement GICD_ISACTIVER/GICD_ICACTIVER properly.)

Quoted 6 lines
Quoted 3 lines

I don't think you can be at the same time an author of the patch and a reviewer. Otherwise, I could review my own patch ;).

Yeah ... that was not the intention. I changed the commit message so I had to add my signed-off-by for copyright reasons. At the same time I reviewed it even before changing the commit message so I added the reviewed-by. I agree it looks kind of weird.

#3 lore raw

Hi,

Quoted 11 lines · Stefano Stabellini
Quoted 5 lines · Julien Grall

Hi,

The title claim this is a resend, but this is actually not the latest version of this patch. Can you explain why you decided to use v1 rather than v2?

Because v2 added a printk for every read, and I thought you only wanted the range fix. Also, the printk is not a "warn once" so after the discussion where it became apparent that the register can be read many times, it sounded undesirable.

You previously mentionned that you will use Peng's patch. From my perspective, this meant you are using the latest version of a patch not a previous one.

So this was a bit of a surprised to me.

Quoted 5 lines

Nonetheless I don't have a strong preference between the two. If you prefer v2 it is here:

https://marc.info/?l=xen-devel&m=157440872522065

I understand the "warn" issue but we did agree with it back then. I am ok to drop it. However, may I request to be more forthcoming in your patch in the future?

For instance in this case, this a link to the original patch and an explanation why you selected v1 would have been really useful.

Quoted 18 lines

Do you need me to resent it? If it is OK for you as-is, you can give your ack here, and I'll go ahead and commit it.

Quoted 13 lines · Stefano Stabellini
Quoted 11 lines

From: Peng Fan <[email protected]>

The end should be GICD_ISACTIVERN not GICD_ISACTIVER.

(See https://marc.info/?l=xen-devel&m=158527653730795 for a discussion on what it would take to implement GICD_ISACTIVER/GICD_ICACTIVER properly.)

Quoted 8 lines
Quoted 2 lines

I don't think you can be at the same time an author of the patch and a reviewer. Otherwise, I could review my own patch ;).

Yeah ... that was not the intention. I changed the commit message so I had to add my signed-off-by for copyright reasons.
At the same time I
reviewed it even before changing the commit message so I added the reviewed-by. I agree it looks kind of weird.

I don't feel this should go as-is. It is not clear in the commit message the changes you did and could lead confusion once merged. One may think you reviewed your own patch.

In general when I tweak a commit message, I will not add my signed-off-by but just add [julieng: Tweak commit message] above my reviewed-by/acked-by tag.

Alternatively, you can remove your reviewed-by and let another maintainer reviewing it.

I will let you decide your preference and resend the patch appropriately.

Cheers,

Signature · 1 line
Julien Grall
author Acked-by Linus Torvalds Reviewed-by Catalin Marinas Reviewed-by Kees Cook Reviewed-by Catalin Marinas Reviewed-by Vincenzo Frascino Reviewed-by Edgar E. Iglesias Reviewed-by Edgar E. Iglesias
#4 lore raw

+ George

Quoted 24 lines · Julien Grall

Hi,

Quoted 11 lines · Stefano Stabellini
Quoted 5 lines · Julien Grall

Hi,

The title claim this is a resend, but this is actually not the latest version of this patch. Can you explain why you decided to use v1 rather than v2?

Because v2 added a printk for every read, and I thought you only wanted the range fix. Also, the printk is not a "warn once" so after the discussion where it became apparent that the register can be read many times, it sounded undesirable.

You previously mentionned that you will use Peng's patch. From my perspective, this meant you are using the latest version of a patch not a previous one.

So this was a bit of a surprised to me.

Quoted 4 lines

Nonetheless I don't have a strong preference between the two. If you prefer v2 it is here:

Quoted 7 lines
Quoted 1 line

https://marc.info/?l=xen-devel&m=157440872522065

I understand the "warn" issue but we did agree with it back then. I am ok to drop it. However, may I request to be more forthcoming in your patch in the future?

For instance in this case, this a link to the original patch and an explanation why you selected v1 would have been really useful.

That's a good point, I can add it. So, for clarity, I'll keep v1 but expand on the commit message to say that we kept v1 to avoid spamming the console.

Quoted 32 lines
Quoted 19 lines

Do you need me to resent it? If it is OK for you as-is, you can give your ack here, and I'll go ahead and commit it.

Quoted 15 lines
Quoted 12 lines · Stefano Stabellini

From: Peng Fan <[email protected]>

The end should be GICD_ISACTIVERN not GICD_ISACTIVER.

(See https://marc.info/?l=xen-devel&m=158527653730795 for a discussion on
what it would take to implement GICD_ISACTIVER/GICD_ICACTIVER properly.)

Quoted 8 lines
Quoted 2 lines

I don't think you can be at the same time an author of the patch and a reviewer. Otherwise, I could review my own patch ;).

Yeah ... that was not the intention. I changed the commit message so I had to add my signed-off-by for copyright reasons.
At the same time I
reviewed it even before changing the commit message so I added the reviewed-by. I agree it looks kind of weird.

I don't feel this should go as-is. It is not clear in the commit message the changes you did and could lead confusion once merged. One may think you reviewed your own patch.

Quoted 7 lines

In general when I tweak a commit message, I will not add my signed-off-by but just add [julieng: Tweak commit message] above my reviewed-by/acked-by tag.

Alternatively, you can remove your reviewed-by and let another maintainer reviewing it.

I will let you decide your preference and resend the patch appropriately.

The Linux policy on Signed-off-by is really strict: basically any time a person touches a patch, even just to commit the patch (no changes to it), they add a Signed-off-by. So it is common there and other projects to have both Signed-off-by and Reviewed-by from the same individual. For instance on Linux:

commit b2a84de2a2deb76a6a51609845341f508c518c03

commit 33e45234987ea3ed4b05fc512f4441696478f12d

On QEMU:

commit 22cd0945b8429f818a2d90f06f02bb526bfb319d

Message-id: [email protected]

commit 133d23b3ad1be53105b9950fb18858cf059f2da6

Your suggestion of adding something between brackets like:

[stefano: update commit message]

is good because it clarifies things and I'd like to do that. But still, I think it would require the addition of my Signed-off-by. At the same time it doesn't look like we want to avoiding adding a Reviewed-by because a reviewer made a change to the commit message too?

Of course, for this patch, I am happy to drop my Reviewed-by and resend and I'll do that. But I think it would be worth clarifying this point at the project level. George, do we or the LinuxFoundation have a policy on whether we can or cannot have reviewed-by and signed-off-by from the same individual?

Acked-by
#5 lore raw
Quoted 32 lines · Stefano Stabellini

+ George

Quoted 29 lines · Julien Grall

Hi,

Quoted 11 lines · Stefano Stabellini
Quoted 5 lines · Julien Grall

Hi,

The title claim this is a resend, but this is actually not the latest version of this patch. Can you explain why you decided to use v1 rather than v2?

Because v2 added a printk for every read, and I thought you only wanted the range fix. Also, the printk is not a "warn once" so after the discussion where it became apparent that the register can be read many times, it sounded undesirable.

You previously mentionned that you will use Peng's patch. From my perspective, this meant you are using the latest version of a patch not a previous one.

So this was a bit of a surprised to me.

Quoted 5 lines

Nonetheless I don't have a strong preference between the two. If you prefer v2 it is here:

https://marc.info/?l=xen-devel&m=157440872522065

I understand the "warn" issue but we did agree with it back then. I am ok to drop it. However, may I request to be more forthcoming in your patch in the future?

Quoted 6 lines
Quoted 2 lines

For instance in this case, this a link to the original patch and an explanation why you selected v1 would have been really useful.

That's a good point, I can add it. So, for clarity, I'll keep v1 but expand on the commit message to say that we kept v1 to avoid spamming the console.

I am fine with that:

Quoted 40 lines
Quoted 40 lines
Quoted 21 lines

Do you need me to resent it? If it is OK for you as-is, you can give your ack here, and I'll go ahead and commit it.

Quoted 15 lines
Quoted 12 lines · Stefano Stabellini

From: Peng Fan <[email protected]>

The end should be GICD_ISACTIVERN not GICD_ISACTIVER.

(See https://marc.info/?l=xen-devel&m=158527653730795 for a discussion on
what it would take to implement GICD_ISACTIVER/GICD_ICACTIVER properly.)

Quoted 8 lines
Quoted 2 lines

I don't think you can be at the same time an author of the patch and a reviewer. Otherwise, I could review my own patch ;).

Yeah ... that was not the intention. I changed the commit message so I had to add my signed-off-by for copyright reasons.
At the same time I
reviewed it even before changing the commit message so I added the reviewed-by. I agree it looks kind of weird.

I don't feel this should go as-is. It is not clear in the commit message the changes you did and could lead confusion once merged. One may think you reviewed your own patch.

In general when I tweak a commit message, I will not add my signed-off-by but just add [julieng: Tweak commit message] above my reviewed-by/acked-by tag.

Alternatively, you can remove your reviewed-by and let another maintainer reviewing it.

Quoted 7 lines
Quoted 1 line

I will let you decide your preference and resend the patch appropriately.

The Linux policy on Signed-off-by is really strict: basically any time a person touches a patch, even just to commit the patch (no changes to it), they add a Signed-off-by. So it is common there and other projects to have both Signed-off-by and Reviewed-by from the same individual. For instance on Linux:

I don't think we used this policy so far for Xen Project. Before pointing out the Signed-off-by vs Reviewed-by, I actually looked online but I wasn't able to find why Signed-off-by and Reviewed-by was added together.

Quoted 39 lines

commit b2a84de2a2deb76a6a51609845341f508c518c03

commit 33e45234987ea3ed4b05fc512f4441696478f12d

On QEMU:

commit 22cd0945b8429f818a2d90f06f02bb526bfb319d

Message-id: [email protected]

commit 133d23b3ad1be53105b9950fb18858cf059f2da6

Quoted 8 lines

Your suggestion of adding something between brackets like:

[stefano: update commit message]

is good because it clarifies things and I'd like to do that. But still, I think it would require the addition of my Signed-off-by. At the same time it doesn't look like we want to avoiding adding a Reviewed-by because a reviewer made a change to the commit message too?

Agree, I also think we need to clarify in this case as it is more difficult to understand if the signed-off-by was by a contributor or someone who merged it.

Quoted 7 lines

Of course, for this patch, I am happy to drop my Reviewed-by and resend and I'll do that. But I think it would be worth clarifying this point at the project level. George, do we or the LinuxFoundation have a policy on whether we can or cannot have reviewed-by and signed-off-by from the same individual?

Cheers,

Signature · 1 line
Julien Grall