Re: [Intel-wired-lan] [PATCH] overflow: Add struct_size_t() helper

From: Tony Nguyen
Date: Wed May 24 2023 - 11:47:58 EST


On 5/24/2023 7:17 AM, Alexander Lobakin wrote:
From: Jakub Kicinski <kuba@xxxxxxxxxx>
Date: Tue, 23 May 2023 20:53:54 -0700

On Mon, 22 May 2023 14:18:13 -0700 Kees Cook wrote:
diff --git a/drivers/net/ethernet/intel/ice/ice_ddp.h b/drivers/net/ethernet/intel/ice/ice_ddp.h
index 37eadb3d27a8..41acfe26df1c 100644
--- a/drivers/net/ethernet/intel/ice/ice_ddp.h
+++ b/drivers/net/ethernet/intel/ice/ice_ddp.h
@@ -185,7 +185,7 @@ struct ice_buf_hdr {
#define ICE_MAX_ENTRIES_IN_BUF(hd_sz, ent_sz) \
((ICE_PKG_BUF_SIZE - \
- struct_size((struct ice_buf_hdr *)0, section_entry, 1) - (hd_sz)) / \
+ struct_size_t(struct ice_buf_hdr, section_entry, 1) - (hd_sz)) / \
(ent_sz))
/* ice package section IDs */
@@ -297,7 +297,7 @@ struct ice_label_section {
};
#define ICE_MAX_LABELS_IN_BUF \
- ICE_MAX_ENTRIES_IN_BUF(struct_size((struct ice_label_section *)0, \
+ ICE_MAX_ENTRIES_IN_BUF(struct_size_t(struct ice_label_section, \
label, 1) - \
sizeof(struct ice_label), \
sizeof(struct ice_label))
@@ -352,7 +352,7 @@ struct ice_boost_tcam_section {
};
#define ICE_MAX_BST_TCAMS_IN_BUF \
- ICE_MAX_ENTRIES_IN_BUF(struct_size((struct ice_boost_tcam_section *)0, \
+ ICE_MAX_ENTRIES_IN_BUF(struct_size_t(struct ice_boost_tcam_section, \
tcam, 1) - \
sizeof(struct ice_boost_tcam_entry), \
sizeof(struct ice_boost_tcam_entry))
@@ -372,8 +372,7 @@ struct ice_marker_ptype_tcam_section {
};
#define ICE_MAX_MARKER_PTYPE_TCAMS_IN_BUF \
- ICE_MAX_ENTRIES_IN_BUF( \
- struct_size((struct ice_marker_ptype_tcam_section *)0, tcam, \
+ ICE_MAX_ENTRIES_IN_BUF(struct_size_t(struct ice_marker_ptype_tcam_section, tcam, \
1) - \
sizeof(struct ice_marker_ptype_tcam_entry), \
sizeof(struct ice_marker_ptype_tcam_entry))

Acked-by: Jakub Kicinski <kuba@xxxxxxxxxx>

but Intel ICE folks please speak up if this has a high chance of
conflicts, I think I've seen some ICE DDP patches flying around :(

I haven't found anything that would conflict with this, esp. since it
implies no functional changes.

Same here. I'm not seeing any conflicts with the patches I'm aware of.

Thanks,
Tony