Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > linux.kernel > #1465162
| From | Paolo Bonzini <pbonzini@redhat.com> |
|---|---|
| Newsgroups | linux.kernel |
| Subject | Re: [PATCH 1/4] KVM-S390: Improve determination of sizes in kvm_s390_import_bp_data() |
| Date | 2016-08-18 13:10 +0200 |
| Message-ID | <s7tmV-7oo-15@gated-at.bofh.it> (permalink) |
| References | (1 earlier) <s7dBw-4DV-21@gated-at.bofh.it> <s7pW1-4X1-11@gated-at.bofh.it> <s7ruO-66w-23@gated-at.bofh.it> <s7sqS-6Jo-21@gated-at.bofh.it> <s7tdh-75c-59@gated-at.bofh.it> |
| Organization | linux.* mail to news gateway |
On 18/08/2016 12:52, walter harms wrote:
>
>
> Am 18.08.2016 11:48, schrieb Paolo Bonzini:
>>
>>
>> On 18/08/2016 11:02, Julia Lawall wrote:
>>>
>>>
>>> On Thu, 18 Aug 2016, walter harms wrote:
>>>
>>>>
>>>>
>>>> Am 17.08.2016 20:06, schrieb SF Markus Elfring:
>>>>> From: Markus Elfring <elfring@users.sourceforge.net>
>>>>> Date: Wed, 17 Aug 2016 18:29:04 +0200
>>>>>
>>>>> Replace the specification of data structures by pointer dereferences
>>>>> to make the corresponding size determination a bit safer according to
>>>>> the Linux coding style convention.
>>>>>
>>>>> Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
>>>>> ---
>>>>> arch/s390/kvm/guestdbg.c | 6 +++---
>>>>> 1 file changed, 3 insertions(+), 3 deletions(-)
>>>>>
>>>>> diff --git a/arch/s390/kvm/guestdbg.c b/arch/s390/kvm/guestdbg.c
>>>>> index d1f8241..b68db4b 100644
>>>>> --- a/arch/s390/kvm/guestdbg.c
>>>>> +++ b/arch/s390/kvm/guestdbg.c
>>>>> @@ -216,7 +216,7 @@ int kvm_s390_import_bp_data(struct kvm_vcpu *vcpu,
>>>>> else if (dbg->arch.nr_hw_bp > MAX_BP_COUNT)
>>>>> return -EINVAL;
>>>>>
>>>>> - size = dbg->arch.nr_hw_bp * sizeof(struct kvm_hw_breakpoint);
>>>>> + size = dbg->arch.nr_hw_bp * sizeof(*bp_data);
>>>>> bp_data = kmalloc(size, GFP_KERNEL);
>>>>> if (!bp_data) {
>>>>> ret = -ENOMEM;
>>>>> @@ -241,7 +241,7 @@ int kvm_s390_import_bp_data(struct kvm_vcpu *vcpu,
>>>>> }
>>>>> }
>>>>>
>>>>> - size = nr_wp * sizeof(struct kvm_hw_wp_info_arch);
>>>>> + size = nr_wp * sizeof(*wp_info);
>>>>> if (size > 0) {
>>>>> wp_info = kmalloc(size, GFP_KERNEL);
>>>>> if (!wp_info) {
>>>>> @@ -249,7 +249,7 @@ int kvm_s390_import_bp_data(struct kvm_vcpu *vcpu,
>>>>> goto error;
>>>>> }
>>>>> }
>>>>> - size = nr_bp * sizeof(struct kvm_hw_bp_info_arch);
>>>>> + size = nr_bp * sizeof(*bp_info);
>>>>> if (size > 0) {
>>>>> bp_info = kmalloc(size, GFP_KERNEL);
>>>>> if (!bp_info) {
>>>>
>>>>
>>>> IMHO the common pattern for kmalloc is
>>>> bp_info = kmalloc( nr_bp * sizeof(*bp_info), GFP_KERNEL);
>>>>
>>>> i can not remember code with a check for size < 0, i guess it is here
>>>> to avoid an overflow ? since kmalloc takes a size_t argument this would cause
>>>> a malloc failure an can be ignored.
>>>
>>> Shoudn't it be kcalloc?
>>
>> Or kmalloc_array, since zeroing is not necessary. Might be an idea for
>> a new Coccinelle script, like
>>
>> - kmalloc (N * sizeof T, GFP)
>> + kmalloc_array(N, sizeof T, GFP)
>>
>
>
> my personal taste is to stay close to the libc functions.
> technical there is no difference
>
> static inline void *kcalloc(size_t n, size_t size, gfp_t flags)
> {
> return kmalloc_array(n, size, flags | __GFP_ZERO);
> }
>
> and i do not see any time critical things here,
This is _not_ premature optimization. (k)calloc tells the reader that
it's safe not to initialize part of the array. kmalloc_array says the
opposite. Using the right function adds important hints in the
code---which unlike comments cannot get stale without also introducing
visible bugs.
Paolo
Back to linux.kernel | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread
[PATCH 0/4] KVM-S390: Fine-tuning for kvm_s390_import_bp_data() SF Markus Elfring <elfring@users.sourceforge.net> - 2016-08-17 20:10 +0200
[PATCH 2/4] KVM-S390: Use memdup_user() rather than duplicating its implementation SF Markus Elfring <elfring@users.sourceforge.net> - 2016-08-17 20:20 +0200
Re: [PATCH 2/4] KVM-S390: Use memdup_user() rather than duplicating its implementation Cornelia Huck <cornelia.huck@de.ibm.com> - 2016-08-22 15:10 +0200
Re: [PATCH 2/4] KVM-S390: Use memdup_user() rather than duplicating its implementation David Hildenbrand <dahi@linux.vnet.ibm.com> - 2016-08-22 15:10 +0200
Re: [PATCH 2/4] KVM-S390: Use memdup_user() rather than duplicating its implementation Christian Borntraeger <borntraeger@de.ibm.com> - 2016-08-24 17:20 +0200
[PATCH v2 0/2] KVM: s390: Fine-tuning for kvm_s390_import_bp_data() SF Markus Elfring <elfring@users.sourceforge.net> - 2016-08-24 20:40 +0200
[PATCH v2 1/2] KVM: s390: Improve determination of sizes in kvm_s390_import_bp_data() SF Markus Elfring <elfring@users.sourceforge.net> - 2016-08-24 20:50 +0200
[PATCH v2 2/2] KVM: s390: Use memdup_user() rather than duplicating code SF Markus Elfring <elfring@users.sourceforge.net> - 2016-08-24 20:50 +0200
[PATCH 3/4] KVM-S390: Less function calls in kvm_s390_import_bp_data() after error detection SF Markus Elfring <elfring@users.sourceforge.net> - 2016-08-17 20:20 +0200
Re: [PATCH 3/4] KVM-S390: Less function calls in kvm_s390_import_bp_data() after error detection David Hildenbrand <dahi@linux.vnet.ibm.com> - 2016-08-22 15:00 +0200
Re: [PATCH 3/4] KVM-S390: Less function calls in kvm_s390_import_bp_data() after error detection Cornelia Huck <cornelia.huck@de.ibm.com> - 2016-08-22 15:10 +0200
Re: [PATCH 3/4] KVM-S390: Less function calls in kvm_s390_import_bp_data() after error detection SF Markus Elfring <elfring@users.sourceforge.net> - 2016-08-22 19:00 +0200
Re: [PATCH 3/4] KVM-S390: Less function calls in kvm_s390_import_bp_data() after error detection Cornelia Huck <cornelia.huck@de.ibm.com> - 2016-08-22 21:40 +0200
Re: [PATCH 3/4] KVM-S390: Less function calls in kvm_s390_import_bp_data() after error detection SF Markus Elfring <elfring@users.sourceforge.net> - 2016-08-22 23:20 +0200
Re: [PATCH 3/4] KVM-S390: Less function calls in kvm_s390_import_bp_data() after error detection Cornelia Huck <cornelia.huck@de.ibm.com> - 2016-08-22 23:30 +0200
Re: [PATCH 3/4] KVM-S390: Less function calls in kvm_s390_import_bp_data() after error detection Christian Borntraeger <borntraeger@de.ibm.com> - 2016-08-24 17:20 +0200
[PATCH 4/4] KVM-S390: Delete an unnecessary initialisation for a buffer variable SF Markus Elfring <elfring@users.sourceforge.net> - 2016-08-17 20:20 +0200
Re: [PATCH 4/4] KVM-S390: Delete an unnecessary initialisation for a buffer variable Cornelia Huck <cornelia.huck@de.ibm.com> - 2016-08-22 15:10 +0200
Re: [PATCH 4/4] KVM-S390: Delete an unnecessary initialisation for a buffer variable David Hildenbrand <dahi@linux.vnet.ibm.com> - 2016-08-22 15:10 +0200
[PATCH 1/4] KVM-S390: Improve determination of sizes in kvm_s390_import_bp_data() SF Markus Elfring <elfring@users.sourceforge.net> - 2016-08-17 20:20 +0200
Re: [PATCH 1/4] KVM-S390: Improve determination of sizes in kvm_s390_import_bp_data() walter harms <wharms@bfs.de> - 2016-08-18 09:30 +0200
Re: [PATCH 1/4] KVM-S390: Improve determination of sizes in kvm_s390_import_bp_data() Julia Lawall <julia.lawall@lip6.fr> - 2016-08-18 11:10 +0200
Re: [PATCH 1/4] KVM-S390: Improve determination of sizes in kvm_s390_import_bp_data() Paolo Bonzini <pbonzini@redhat.com> - 2016-08-18 12:10 +0200
Re: [PATCH 1/4] KVM-S390: Improve determination of sizes in kvm_s390_import_bp_data() walter harms <wharms@bfs.de> - 2016-08-18 13:00 +0200
Re: [PATCH 1/4] KVM-S390: Improve determination of sizes in kvm_s390_import_bp_data() Paolo Bonzini <pbonzini@redhat.com> - 2016-08-18 13:10 +0200
Re: [PATCH 1/4] KVM-S390: Improve determination of sizes in kvm_s390_import_bp_data() Cornelia Huck <cornelia.huck@de.ibm.com> - 2016-08-22 15:00 +0200
Re: Replacing specific kmalloc() calls by kmalloc_array()? SF Markus Elfring <elfring@users.sourceforge.net> - 2016-08-24 14:20 +0200
Re: Replacing specific kmalloc() calls by kmalloc_array()? Paolo Bonzini <pbonzini@redhat.com> - 2016-08-24 17:10 +0200
csiph-web