Skip to contentSkip to conversation

Fix crash with illegal "-net nic, model=xxx" option

PATCH Applied Thomas Huth 8 participants 12 messages Apr 9, 2015 – Apr 27, 2015

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 Fix crash with illegal "-net nic, model=xxx" option

Current QEMU crashes when specifying an illegal model with the "-net nic,model=xxx" option, e.g.:

 $ qemu-system-x86_64 -net nic,model=n/a
 qemu-system-x86_64: Unsupported NIC model: n/a

Program received signal SIGSEGV, Segmentation fault.

The gdb backtrace looks like this:

0x0000555555965fe0 in error_get_pretty (err=0x0) at util/error.c:152
152 return err->msg;
(gdb) bt

 0  0x0000555555965fe0 in error_get_pretty (err=0x0) at util/error.c:152
 1  0x0000555555965ffd in error_report_err (err=0x0) at util/error.c:157
 2  0x0000555555809c90 in pci_nic_init_nofail (nd=0x555555e49860 <nd_table>, rootbus=0x5555564409b0,
    default_model=0x55555598c37b "e1000", default_devaddr=0x0) at hw/pci/pci.c:1663
 3  0x0000555555691e42 in pc_nic_init (isa_bus=0x555556f71900, pci_bus=0x5555564409b0)
    at hw/i386/pc.c:1506
 4  0x000055555569396b in pc_init1 (machine=0x5555562abbf0, pci_enabled=1, kvmclock_enabled=1)
    at hw/i386/pc_piix.c:248
 5  0x0000555555693d27 in pc_init_pci (machine=0x5555562abbf0) at hw/i386/pc_piix.c:310
 6  0x000055555572ddf5 in main (argc=3, argv=0x7fffffffe018, envp=0x7fffffffe038) at vl.c:4226

The problem is that pci_nic_init_nofail() does not check whether the err parameter from pci_nic_init has been set up and thus passes a NULL pointer to error_report_err(). Fix it by correctly checking the err parameter.

---
 hw/pci/pci.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
hw/pci/pci.c +3 -1
index 6941a82..b3d5100 100644
@@ -1660,7 +1660,9 @@ PCIDevice *pci_nic_init_nofail(NICInfo *nd, PCIBus *rootbus,
 
     res = pci_nic_init(nd, rootbus, default_model, default_devaddr, &err);
     if (!res) {
-        error_report_err(err);
+        if (err) {
+            error_report_err(err);
+        }
         exit(1);
     }
     return res;
Signature · 1 line
1.8.3.1
#1 lore raw
Quoted 24 lines · Thomas Huth

Current QEMU crashes when specifying an illegal model with the "-net nic,model=xxx" option, e.g.:

 $ qemu-system-x86_64 -net nic,model=n/a
 qemu-system-x86_64: Unsupported NIC model: n/a

Program received signal SIGSEGV, Segmentation fault.

The gdb backtrace looks like this:

0x0000555555965fe0 in error_get_pretty (err=0x0) at util/error.c:152
152 return err->msg;
(gdb) bt

 0  0x0000555555965fe0 in error_get_pretty (err=0x0) at util/error.c:152
 1  0x0000555555965ffd in error_report_err (err=0x0) at util/error.c:157
 2  0x0000555555809c90 in pci_nic_init_nofail (nd=0x555555e49860 <nd_table>, rootbus=0x5555564409b0,
    default_model=0x55555598c37b "e1000", default_devaddr=0x0) at hw/pci/pci.c:1663
 3  0x0000555555691e42 in pc_nic_init (isa_bus=0x555556f71900, pci_bus=0x5555564409b0)
    at hw/i386/pc.c:1506
 4  0x000055555569396b in pc_init1 (machine=0x5555562abbf0, pci_enabled=1, kvmclock_enabled=1)
    at hw/i386/pc_piix.c:248
 5  0x0000555555693d27 in pc_init_pci (machine=0x5555562abbf0) at hw/i386/pc_piix.c:310
 6  0x000055555572ddf5 in main (argc=3, argv=0x7fffffffe018, envp=0x7fffffffe038) at vl.c:4226
Quoted 5 lines

The problem is that pci_nic_init_nofail() does not check whether the err parameter from pci_nic_init has been set up and thus passes a NULL pointer to error_report_err(). Fix it by correctly checking the err parameter.

Thanks!
Given that this is a legacy -net option, I'm inclined to fix it post-2.3, and Cc stable.
Unfortunately I won't be able to do a pull request before rc3.

Quoted 21 lines
---
 hw/pci/pci.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
hw/pci/pci.c +3 -1
index 6941a82..b3d5100 100644
@@ -1660,7 +1660,9 @@ PCIDevice *pci_nic_init_nofail(NICInfo *nd, PCIBus *rootbus,
 
     res = pci_nic_init(nd, rootbus, default_model, default_devaddr, &err);
     if (!res) {
-        error_report_err(err);
+        if (err) {
+            error_report_err(err);
+        }
         exit(1);
     }
     return res;
Signature · 1 line
1.8.3.1
#2 lore raw
Quoted 29 lines · Michael S. Tsirkin
Quoted 28 lines · Thomas Huth

Current QEMU crashes when specifying an illegal model with the "-net nic,model=xxx" option, e.g.:

 $ qemu-system-x86_64 -net nic,model=n/a
 qemu-system-x86_64: Unsupported NIC model: n/a

Program received signal SIGSEGV, Segmentation fault.

The gdb backtrace looks like this:

0x0000555555965fe0 in error_get_pretty (err=0x0) at util/error.c:152
152 return err->msg;
(gdb) bt

 0  0x0000555555965fe0 in error_get_pretty (err=0x0) at util/error.c:152
 1  0x0000555555965ffd in error_report_err (err=0x0) at util/error.c:157
 2  0x0000555555809c90 in pci_nic_init_nofail (nd=0x555555e49860 <nd_table>, rootbus=0x5555564409b0,
    default_model=0x55555598c37b "e1000", default_devaddr=0x0) at hw/pci/pci.c:1663
 3  0x0000555555691e42 in pc_nic_init (isa_bus=0x555556f71900, pci_bus=0x5555564409b0)
    at hw/i386/pc.c:1506
 4  0x000055555569396b in pc_init1 (machine=0x5555562abbf0, pci_enabled=1, kvmclock_enabled=1)
    at hw/i386/pc_piix.c:248
 5  0x0000555555693d27 in pc_init_pci (machine=0x5555562abbf0) at hw/i386/pc_piix.c:310
 6  0x000055555572ddf5 in main (argc=3, argv=0x7fffffffe018, envp=0x7fffffffe038) at vl.c:4226

The problem is that pci_nic_init_nofail() does not check whether the err parameter from pci_nic_init has been set up and thus passes a NULL pointer to error_report_err(). Fix it by correctly checking the err parameter.

Quoted 6 lines
Quoted 1 line

Thanks!
Given that this is a legacy -net option, I'm inclined to fix it post-2.3, and Cc stable.
Unfortunately I won't be able to do a pull request before rc3.

Since this is a pretty safe and simple change I'm happy to apply it direct to master if you like. Do you want to provide a reviewed-by tag?

thanks
-- PMM

#3 lore raw
Quoted 37 lines · Peter Maydell
Quoted 29 lines · Michael S. Tsirkin
Quoted 28 lines · Thomas Huth

Current QEMU crashes when specifying an illegal model with the "-net nic,model=xxx" option, e.g.:

 $ qemu-system-x86_64 -net nic,model=n/a
 qemu-system-x86_64: Unsupported NIC model: n/a

Program received signal SIGSEGV, Segmentation fault.

The gdb backtrace looks like this:

0x0000555555965fe0 in error_get_pretty (err=0x0) at util/error.c:152
152 return err->msg;
(gdb) bt

 0  0x0000555555965fe0 in error_get_pretty (err=0x0) at util/error.c:152
 1  0x0000555555965ffd in error_report_err (err=0x0) at util/error.c:157
 2  0x0000555555809c90 in pci_nic_init_nofail (nd=0x555555e49860 <nd_table>, rootbus=0x5555564409b0,
    default_model=0x55555598c37b "e1000", default_devaddr=0x0) at hw/pci/pci.c:1663
 3  0x0000555555691e42 in pc_nic_init (isa_bus=0x555556f71900, pci_bus=0x5555564409b0)
    at hw/i386/pc.c:1506
 4  0x000055555569396b in pc_init1 (machine=0x5555562abbf0, pci_enabled=1, kvmclock_enabled=1)
    at hw/i386/pc_piix.c:248
 5  0x0000555555693d27 in pc_init_pci (machine=0x5555562abbf0) at hw/i386/pc_piix.c:310
 6  0x000055555572ddf5 in main (argc=3, argv=0x7fffffffe018, envp=0x7fffffffe038) at vl.c:4226

The problem is that pci_nic_init_nofail() does not check whether the err parameter from pci_nic_init has been set up and thus passes a NULL pointer to error_report_err(). Fix it by correctly checking the err parameter.

Quoted 6 lines
Quoted 1 line

Thanks!
Given that this is a legacy -net option, I'm inclined to fix it post-2.3, and Cc stable.
Unfortunately I won't be able to do a pull request before rc3.

Quoted 6 lines

Since this is a pretty safe and simple change I'm happy to apply it direct to master if you like. Do you want to provide a reviewed-by tag?

thanks
-- PMM

I reviewed the patch, and sent a tag.
I'm fine with you making the decision on whether it's appropriate for 2.3.

Signature · 1 line
MST
#4 lore raw

Am 12.04.2015 um 13:14 schrieb Michael S. Tsirkin:

Quoted 38 lines · Peter Maydell
Quoted 37 lines · Michael S. Tsirkin
Quoted 35 lines · Thomas Huth
Quoted 24 lines

Current QEMU crashes when specifying an illegal model with the "-net nic,model=xxx" option, e.g.:

 $ qemu-system-x86_64 -net nic,model=n/a
 qemu-system-x86_64: Unsupported NIC model: n/a

Program received signal SIGSEGV, Segmentation fault.

The gdb backtrace looks like this:

0x0000555555965fe0 in error_get_pretty (err=0x0) at util/error.c:152
152 return err->msg;
(gdb) bt

 0  0x0000555555965fe0 in error_get_pretty (err=0x0) at util/error.c:152
 1  0x0000555555965ffd in error_report_err (err=0x0) at util/error.c:157
 2  0x0000555555809c90 in pci_nic_init_nofail (nd=0x555555e49860 <nd_table>, rootbus=0x5555564409b0,
    default_model=0x55555598c37b "e1000", default_devaddr=0x0) at hw/pci/pci.c:1663
 3  0x0000555555691e42 in pc_nic_init (isa_bus=0x555556f71900, pci_bus=0x5555564409b0)
    at hw/i386/pc.c:1506
 4  0x000055555569396b in pc_init1 (machine=0x5555562abbf0, pci_enabled=1, kvmclock_enabled=1)
    at hw/i386/pc_piix.c:248
 5  0x0000555555693d27 in pc_init_pci (machine=0x5555562abbf0) at hw/i386/pc_piix.c:310
 6  0x000055555572ddf5 in main (argc=3, argv=0x7fffffffe018, envp=0x7fffffffe038) at vl.c:4226
Quoted 5 lines

The problem is that pci_nic_init_nofail() does not check whether the err parameter from pci_nic_init has been set up and thus passes a NULL pointer to error_report_err(). Fix it by correctly checking the err parameter.

Thanks!
Given that this is a legacy -net option, I'm inclined to fix it post-2.3, and Cc stable.
Unfortunately I won't be able to do a pull request before rc3.

Quoted 7 lines
Quoted 3 lines

Since this is a pretty safe and simple change I'm happy to apply it direct to master if you like. Do you want to provide a reviewed-by tag?

I reviewed the patch, and sent a tag.
I'm fine with you making the decision on whether it's appropriate for 2.3.

Hope you added a "pci:" topic? :)

Andreas

Signature · 3 lines
SUSE Linux GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
GF: Felix Imendörffer, Jane Smithard, Jennifer Guild, Dilip Upmanyu,
Graham Norton; HRB 21284 (AG Nürnberg)
#5 lore raw
In reply to Thomas Huth
Quoted 21 lines · Thomas Huth

Current QEMU crashes when specifying an illegal model with the "-net nic,model=xxx" option, e.g.:

 $ qemu-system-x86_64 -net nic,model=n/a
 qemu-system-x86_64: Unsupported NIC model: n/a

The problem is that pci_nic_init_nofail() does not check whether the err parameter from pci_nic_init has been set up and thus passes a NULL pointer to error_report_err(). Fix it by correctly checking the err parameter.

---
 hw/pci/pci.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
hw/pci/pci.c +0 -0
index 6941a82..b3d5100 100644

hw/pci/pci.c PCIDevice *pci_nic_init_nofail(NICInfo *nd, PCIBus *rootbus, · line 1666

Quoted 8 lines
hw/pci/pci.c +3 -1
@@ -1660,7 +1660,9 @@ PCIDevice *pci_nic_init_nofail(NICInfo *nd, PCIBus *rootbus,
 
     res = pci_nic_init(nd, rootbus, default_model, default_devaddr, &err);
     if (!res) {
-        error_report_err(err);
+        if (err) {
+            error_report_err(err);
+        }
         exit(1);

Doesn't this mean the program can exit without an error message, if pci_nic_init returns failure but failed to set err? Shouldn't you at least print something in that case as an else branch?

Signature · 2 lines
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org
#6 lore raw
Quoted 6 lines · Eric Blake
Quoted 6 lines
hw/pci/pci.c +0 -0
index 6941a82..b3d5100 100644
@@ -1660,7 +1660,9 @@ PCIDevice *pci_nic_init_nofail(NICInfo *nd, PCIBus *rootbus,
 
Quoted 11 lines

hw/pci/pci.c PCIDevice *pci_nic_init_nofail(NICInfo *nd, PCIBus *rootbus, · line 1666

Quoted 7 lines
hw/pci/pci.c +3 -1
     res = pci_nic_init(nd, rootbus, default_model, default_devaddr, &err);
     if (!res) {
-        error_report_err(err);
+        if (err) {
+            error_report_err(err);
+        }
         exit(1);

Doesn't this mean the program can exit without an error message, if pci_nic_init returns failure but failed to set err? Shouldn't you at least print something in that case as an else branch?

git grep 'Unsupported NIC model' shows that the error is printed with error_report; same for other errors produced by pci_nic_init.

This is not beautiful compared to correct propagation of Error*, but it's okay because -net is only used at startup. It's good enough for rc3.

Paolo

#7 lore raw
Quoted 19 lines · Paolo Bonzini
Quoted 6 lines · Eric Blake
Quoted 6 lines
hw/pci/pci.c +0 -0
index 6941a82..b3d5100 100644
@@ -1660,7 +1660,9 @@ PCIDevice *pci_nic_init_nofail(NICInfo *nd, PCIBus *rootbus,
 
Quoted 11 lines

hw/pci/pci.c PCIDevice *pci_nic_init_nofail(NICInfo *nd, PCIBus *rootbus, · line 1666

Quoted 7 lines
hw/pci/pci.c +3 -1
     res = pci_nic_init(nd, rootbus, default_model, default_devaddr, &err);
     if (!res) {
-        error_report_err(err);
+        if (err) {
+            error_report_err(err);
+        }
         exit(1);

Doesn't this mean the program can exit without an error message, if pci_nic_init returns failure but failed to set err? Shouldn't you at least print something in that case as an else branch?

Quoted 5 lines

git grep 'Unsupported NIC model' shows that the error is printed with error_report; same for other errors produced by pci_nic_init.

This is not beautiful compared to correct propagation of Error*, but it's okay because -net is only used at startup. It's good enough for rc3.

Thomas, can you clean this up now 2.3 is out?

#8 lore raw
Quoted 21 lines · On Mon, 27 Apr 2015 13:48:27 +0200 Markus Armbruster <[email protected]> wrote:
Quoted 19 lines · Paolo Bonzini
Quoted 17 lines · Eric Blake

hw/pci/pci.c PCIDevice *pci_nic_init_nofail(NICInfo *nd, PCIBus *rootbus, · line 1666

Quoted 13 lines
hw/pci/pci.c +3 -1
index 6941a82..b3d5100 100644
@@ -1660,7 +1660,9 @@ PCIDevice *pci_nic_init_nofail(NICInfo *nd, PCIBus *rootbus,
 
     res = pci_nic_init(nd, rootbus, default_model, default_devaddr, &err);
     if (!res) {
-        error_report_err(err);
+        if (err) {
+            error_report_err(err);
+        }
         exit(1);

Doesn't this mean the program can exit without an error message, if pci_nic_init returns failure but failed to set err? Shouldn't you at least print something in that case as an else branch?

Quoted 7 lines
Quoted 5 lines

git grep 'Unsupported NIC model' shows that the error is printed with error_report; same for other errors produced by pci_nic_init.

This is not beautiful compared to correct propagation of Error*, but it's okay because -net is only used at startup. It's good enough for rc3.

Thomas, can you clean this up now 2.3 is out?

Sure, I'll have a look.

Thomas

Reviewed-by on v1 1/1
#9 lore raw
In reply to Thomas Huth
Quoted 24 lines · Thomas Huth

Current QEMU crashes when specifying an illegal model with the "-net nic,model=xxx" option, e.g.:

 $ qemu-system-x86_64 -net nic,model=n/a
 qemu-system-x86_64: Unsupported NIC model: n/a

Program received signal SIGSEGV, Segmentation fault.

The gdb backtrace looks like this:

0x0000555555965fe0 in error_get_pretty (err=0x0) at util/error.c:152
152 return err->msg;
(gdb) bt

 0  0x0000555555965fe0 in error_get_pretty (err=0x0) at util/error.c:152
 1  0x0000555555965ffd in error_report_err (err=0x0) at util/error.c:157
 2  0x0000555555809c90 in pci_nic_init_nofail (nd=0x555555e49860 <nd_table>, rootbus=0x5555564409b0,
    default_model=0x55555598c37b "e1000", default_devaddr=0x0) at hw/pci/pci.c:1663
 3  0x0000555555691e42 in pc_nic_init (isa_bus=0x555556f71900, pci_bus=0x5555564409b0)
    at hw/i386/pc.c:1506
 4  0x000055555569396b in pc_init1 (machine=0x5555562abbf0, pci_enabled=1, kvmclock_enabled=1)
    at hw/i386/pc_piix.c:248
 5  0x0000555555693d27 in pc_init_pci (machine=0x5555562abbf0) at hw/i386/pc_piix.c:310
 6  0x000055555572ddf5 in main (argc=3, argv=0x7fffffffe018, envp=0x7fffffffe038) at vl.c:4226
Quoted 5 lines

The problem is that pci_nic_init_nofail() does not check whether the err parameter from pci_nic_init has been set up and thus passes a NULL pointer to error_report_err(). Fix it by correctly checking the err parameter.

Quoted 21 lines
---
 hw/pci/pci.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
hw/pci/pci.c +3 -1
index 6941a82..b3d5100 100644
@@ -1660,7 +1660,9 @@ PCIDevice *pci_nic_init_nofail(NICInfo *nd, PCIBus *rootbus,
 
     res = pci_nic_init(nd, rootbus, default_model, default_devaddr, &err);
     if (!res) {
-        error_report_err(err);
+        if (err) {
+            error_report_err(err);
+        }
         exit(1);
     }
     return res;
Signature · 1 line
1.8.3.1
Reviewed-by on v1 1/1
#10 lore raw
In reply to Thomas Huth
Quoted 56 lines · Thomas Huth

Current QEMU crashes when specifying an illegal model with the "-net nic,model=xxx" option, e.g.:

 $ qemu-system-x86_64 -net nic,model=n/a
 qemu-system-x86_64: Unsupported NIC model: n/a

Program received signal SIGSEGV, Segmentation fault.

The gdb backtrace looks like this:

0x0000555555965fe0 in error_get_pretty (err=0x0) at util/error.c:152
152 return err->msg;
(gdb) bt
0 0x0000555555965fe0 in error_get_pretty (err=0x0) at
util/error.c:152
1 0x0000555555965ffd in error_report_err (err=0x0) at
util/error.c:157
2 0x0000555555809c90 in pci_nic_init_nofail (nd=0x555555e49860
<nd_table>, rootbus=0x5555564409b0,

    default_model=0x55555598c37b "e1000", default_devaddr=0x0) at 
hw/pci/pci.c:1663
 3  0x0000555555691e42 in pc_nic_init (isa_bus=0x555556f71900, 
pci_bus=0x5555564409b0)
    at hw/i386/pc.c:1506
 4  0x000055555569396b in pc_init1 (machine=0x5555562abbf0, 
pci_enabled=1, kvmclock_enabled=1)
    at hw/i386/pc_piix.c:248
 5  0x0000555555693d27 in pc_init_pci (machine=0x5555562abbf0) at 

hw/i386/pc_piix.c:310
6 0x000055555572ddf5 in main (argc=3, argv=0x7fffffffe018,
envp=0x7fffffffe038) at vl.c:4226

The problem is that pci_nic_init_nofail() does not check whether the err parameter from pci_nic_init has been set up and thus passes a NULL pointer to error_report_err(). Fix it by correctly checking the err parameter.

---
 hw/pci/pci.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
hw/pci/pci.c +3 -1
index 6941a82..b3d5100 100644
@@ -1660,7 +1660,9 @@ PCIDevice *pci_nic_init_nofail(NICInfo *nd, 
 PCIBus *rootbus,
 
     res = pci_nic_init(nd, rootbus, default_model, default_devaddr, 
 &err);
     if (!res) {
-        error_report_err(err);
+        if (err) {
+            error_report_err(err);
+        }
         exit(1);
Quoted 4 lines
     }
     return res;
Signature · 1 line
1.8.3.1
#11 lore raw
Quoted 10 lines · Jason Wang
Quoted 6 lines · Thomas Huth

Current QEMU crashes when specifying an illegal model with the "-net nic,model=xxx" option, e.g.:

 $ qemu-system-x86_64 -net nic,model=n/a
 qemu-system-x86_64: Unsupported NIC model: n/a
Quoted 7 lines
Quoted 5 lines

The problem is that pci_nic_init_nofail() does not check whether the err parameter from pci_nic_init has been set up and thus passes a NULL pointer to error_report_err(). Fix it by correctly checking the err parameter.

Applied, thanks (with a 'pci:' subject prefix added).

-- PMM