Skip to contentSkip to conversation

vfio: do not use memcmp() to compare PCI address

PATCH In review Andrew Rybchenko 2 participants 4 messages Mar 31, 2017

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 vfio: do not use memcmp() to compare PCI address

PCI address structure has padding which may have garbage.

Fixes 2f4adfad0a69 ("vfio: add multiprocess support")

---
It is a real bug which I've hit during multi-process debugging.

 lib/librte_eal/linuxapp/eal/eal_pci_vfio.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
lib/librte_eal/linuxapp/eal/eal_pci_vfio.c +2 -1
index 5f478c5..7d8b9fb 100644
@@ -355,7 +355,8 @@ pci_vfio_map_resource(struct rte_pci_device *dev)
 	} else {
 		/* if we're in a secondary process, just find our tailq entry */
 		TAILQ_FOREACH(vfio_res, vfio_res_list, next) {
-			if (memcmp(&vfio_res->pci_addr, &dev->addr, sizeof(dev->addr)))
+			if (rte_eal_compare_pci_addr(&vfio_res->pci_addr,
+						     &dev->addr))
 				continue;
 			break;
 		}
Signature · 1 line
2.9.3
#1 lore raw
Quoted 17 lines · Andrew Rybchenko

PCI address structure has padding which may have garbage.

Fixes 2f4adfad0a69 ("vfio: add multiprocess support")

---
It is a real bug which I've hit during multi-process debugging.

  lib/librte_eal/linuxapp/eal/eal_pci_vfio.c | 3 ++-
  1 file changed, 2 insertions(+), 1 deletion(-)
lib/librte_eal/linuxapp/eal/eal_pci_vfio.c +0 -0
index 5f478c5..7d8b9fb 100644

lib/librte_eal/linuxapp/eal/eal_pci_vfio.c pci_vfio_map_resource(struct rte_pci_device *dev) · line 362

Quoted 9 lines
lib/librte_eal/linuxapp/eal/eal_pci_vfio.c +2 -1
@@ -355,7 +355,8 @@ pci_vfio_map_resource(struct rte_pci_device *dev)
  	} else {
  		/* if we're in a secondary process, just find our tailq entry */
  		TAILQ_FOREACH(vfio_res, vfio_res_list, next) {
-			if (memcmp(&vfio_res->pci_addr, &dev->addr, sizeof(dev->addr)))
+			if (rte_eal_compare_pci_addr(&vfio_res->pci_addr,
+						     &dev->addr))
  				continue;
  			break;
  		}

Different commit, same patch :)
http://dpdk.org/dev/patchwork/patch/21828/

Regards,
Sergio

#2 lore raw
Quoted 24 lines · Sergio Gonzalez Monroy
Quoted 23 lines · Andrew Rybchenko

PCI address structure has padding which may have garbage.

Fixes 2f4adfad0a69 ("vfio: add multiprocess support")

---
It is a real bug which I've hit during multi-process debugging.

  lib/librte_eal/linuxapp/eal/eal_pci_vfio.c | 3 ++-
  1 file changed, 2 insertions(+), 1 deletion(-)
lib/librte_eal/linuxapp/eal/eal_pci_vfio.c +0 -0
index 5f478c5..7d8b9fb 100644
@@ -355,7 +355,8 @@ pci_vfio_map_resource(struct rte_pci_device *dev)
      } else {
          /* if we're in a secondary process, just find our tailq 

entry */

          TAILQ_FOREACH(vfio_res, vfio_res_list, next) {

- if (memcmp(&vfio_res->pci_addr, &dev->addr,
sizeof(dev->addr)))

Quoted 8 lines

lib/librte_eal/linuxapp/eal/eal_pci_vfio.c pci_vfio_map_resource(struct rte_pci_device *dev) · line 362

Quoted 5 lines
lib/librte_eal/linuxapp/eal/eal_pci_vfio.c +2 -0
+            if (rte_eal_compare_pci_addr(&vfio_res->pci_addr,
+                             &dev->addr))
                  continue;
              break;
          }

Different commit, same patch :)
https://urldefense.proofpoint.com/v2/url?u=http-3A__dpdk.org_dev_patchwork_patch_21828_&d=DQICaQ&c=euGZstcaTDllvimEN8b7jXrwqOf-v5A_CdpgnVfiiMM&r=flTOx6Av679My7o_iScb5sOlLD68bpUyE2RUtfW3SWQ&m=uvsF8SJaLsRNp3l01Ji5iD4EB1hkHVPOX_dFKqTu_mU&s=HYxvqVvNsbEzQgMJmwQuH6G7vqOkyg7vJ1PrzsxKr5c&e=

True, sorry, lost your patch from my view.
I'm not sure which state should be set on my patch in the patchwork. Thomas, please, resolve it properly.

Thanks,
Andrew.

#3 lore raw
Quoted 35 lines · Andrew Rybchenko
Quoted 24 lines · Sergio Gonzalez Monroy
Quoted 23 lines · Andrew Rybchenko

PCI address structure has padding which may have garbage.

Fixes 2f4adfad0a69 ("vfio: add multiprocess support")

---
It is a real bug which I've hit during multi-process debugging.

  lib/librte_eal/linuxapp/eal/eal_pci_vfio.c | 3 ++-
  1 file changed, 2 insertions(+), 1 deletion(-)
lib/librte_eal/linuxapp/eal/eal_pci_vfio.c +0 -0
index 5f478c5..7d8b9fb 100644
@@ -355,7 +355,8 @@ pci_vfio_map_resource(struct rte_pci_device *dev)
      } else {
          /* if we're in a secondary process, just find our tailq 

entry */

          TAILQ_FOREACH(vfio_res, vfio_res_list, next) {

- if (memcmp(&vfio_res->pci_addr, &dev->addr,
sizeof(dev->addr)))

Quoted 8 lines

lib/librte_eal/linuxapp/eal/eal_pci_vfio.c pci_vfio_map_resource(struct rte_pci_device *dev) · line 362

Quoted 5 lines
lib/librte_eal/linuxapp/eal/eal_pci_vfio.c +2 -0
+            if (rte_eal_compare_pci_addr(&vfio_res->pci_addr,
+                             &dev->addr))
                  continue;
              break;
          }

Different commit, same patch :)
https://urldefense.proofpoint.com/v2/url?u=http-3A__dpdk.org_dev_patchwork_patch_21828_&d=DQICaQ&c=euGZstcaTDllvimEN8b7jXrwqOf-v5A_CdpgnVfiiMM&r=flTOx6Av679My7o_iScb5sOlLD68bpUyE2RUtfW3SWQ&m=uvsF8SJaLsRNp3l01Ji5iD4EB1hkHVPOX_dFKqTu_mU&s=HYxvqVvNsbEzQgMJmwQuH6G7vqOkyg7vJ1PrzsxKr5c&e=

Quoted 3 lines

True, sorry, lost your patch from my view.
I'm not sure which state should be set on my patch in the patchwork. Thomas, please, resolve it properly.

No worries. Probably the title on the other patch was a bit misleading.

You could just set it to the "Not Applicable" state.

Sergio

Quoted 2 lines

Thanks,
Andrew.