| Message ID | 20220521063534.138930-1-yiyang13@huawei.com |
|---|---|
| State | New |
| Headers | show |
| Series | [-next] cpufreq: Fix reserved space in cpufreq_show_cpus() | expand |
On 14-06-22, 15:48, Rafael J. Wysocki wrote: > On Tue, Jun 14, 2022 at 3:37 PM Rafael J. Wysocki <rafael@kernel.org> wrote: > > > > On Tue, May 24, 2022 at 9:10 AM Viresh Kumar <viresh.kumar@linaro.org> wrote: > > > > > > On 21-05-22, 14:35, Yi Yang wrote: > > > > Function scnprintf() would reserve space for the trailing '\0' and return > > > > value is the number of characters written into buf not including the > > > > trailing '\0'. internally meaning the next scnprintf() would write begin > > > > the trailing '\0'. The code specifying "PAGE_SIZE - i - 2" here is trying > > > > to reserve space for "\n\0" which would cause scnprintf() to reserve an > > > > additional byte making the tail of the buf looks like this: "\n\0\0". > > > > Thus. we should reserve only the space for one '\0'. passing in > > > > "PAGE_SIZE - i - 1". > > > > > > > > Additionally, each iteration would replace the trailing '\0' from the last > > > > iteration with a space, and append 4 additional bytes to the string making > > > > it a total of 5 additional bytes. That means we should stop printing into > > > > the buffer if the remaining size is less than 7 bytes(1 for the ' ', 4 for > > > > the %u and 2 for the tailing "\n\0") > > > > > > > > Signed-off-by: Yi Yang <yiyang13@huawei.com> > > > > --- > > > > drivers/cpufreq/cpufreq.c | 6 +++--- > > > > 1 file changed, 3 insertions(+), 3 deletions(-) > > > > > > > > diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c > > > > index 1f6667ce43bd..60c005c9961e 100644 > > > > --- a/drivers/cpufreq/cpufreq.c > > > > +++ b/drivers/cpufreq/cpufreq.c > > > > @@ -844,9 +844,9 @@ ssize_t cpufreq_show_cpus(const struct cpumask *mask, char *buf) > > > > > > > > for_each_cpu(cpu, mask) { > > > > if (i) > > > > - i += scnprintf(&buf[i], (PAGE_SIZE - i - 2), " "); > > > > - i += scnprintf(&buf[i], (PAGE_SIZE - i - 2), "%u", cpu); > > > > - if (i >= (PAGE_SIZE - 5)) > > > > + i += scnprintf(&buf[i], (PAGE_SIZE - i - 1), " "); > > > > + i += scnprintf(&buf[i], (PAGE_SIZE - i - 1), "%u", cpu); > > > > + if (i >= (PAGE_SIZE - 6)) > > > > break; > > > > } > > > > i += sprintf(&buf[i], "\n"); > > > > > > Acked-by: Viresh Kumar <viresh.kumar@linaro.org> > > > > Applied as 5.20 material, thanks! > > And dropped, because it has been superseded by this one: > > https://patchwork.kernel.org/project/linux-pm/patch/b9fa08171c09343ace94a7343553a4bee4695c90.1653565641.git.viresh.kumar@linaro.org/ The $Subject patch is still required AFAICT, it is fixing a different problem. Though it needs to be rebased on top of your branch now.
On 2022/6/15 12:56, Viresh Kumar wrote: > On 14-06-22, 15:48, Rafael J. Wysocki wrote: >> On Tue, Jun 14, 2022 at 3:37 PM Rafael J. Wysocki <rafael@kernel.org> wrote: >>> >>> On Tue, May 24, 2022 at 9:10 AM Viresh Kumar <viresh.kumar@linaro.org> wrote: >>>> >>>> On 21-05-22, 14:35, Yi Yang wrote: >>>>> Function scnprintf() would reserve space for the trailing '\0' and return >>>>> value is the number of characters written into buf not including the >>>>> trailing '\0'. internally meaning the next scnprintf() would write begin >>>>> the trailing '\0'. The code specifying "PAGE_SIZE - i - 2" here is trying >>>>> to reserve space for "\n\0" which would cause scnprintf() to reserve an >>>>> additional byte making the tail of the buf looks like this: "\n\0\0". >>>>> Thus. we should reserve only the space for one '\0'. passing in >>>>> "PAGE_SIZE - i - 1". >>>>> >>>>> Additionally, each iteration would replace the trailing '\0' from the last >>>>> iteration with a space, and append 4 additional bytes to the string making >>>>> it a total of 5 additional bytes. That means we should stop printing into >>>>> the buffer if the remaining size is less than 7 bytes(1 for the ' ', 4 for >>>>> the %u and 2 for the tailing "\n\0") >>>>> >>>>> Signed-off-by: Yi Yang <yiyang13@huawei.com> >>>>> --- >>>>> drivers/cpufreq/cpufreq.c | 6 +++--- >>>>> 1 file changed, 3 insertions(+), 3 deletions(-) >>>>> >>>>> diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c >>>>> index 1f6667ce43bd..60c005c9961e 100644 >>>>> --- a/drivers/cpufreq/cpufreq.c >>>>> +++ b/drivers/cpufreq/cpufreq.c >>>>> @@ -844,9 +844,9 @@ ssize_t cpufreq_show_cpus(const struct cpumask *mask, char *buf) >>>>> >>>>> for_each_cpu(cpu, mask) { >>>>> if (i) >>>>> - i += scnprintf(&buf[i], (PAGE_SIZE - i - 2), " "); >>>>> - i += scnprintf(&buf[i], (PAGE_SIZE - i - 2), "%u", cpu); >>>>> - if (i >= (PAGE_SIZE - 5)) >>>>> + i += scnprintf(&buf[i], (PAGE_SIZE - i - 1), " "); >>>>> + i += scnprintf(&buf[i], (PAGE_SIZE - i - 1), "%u", cpu); >>>>> + if (i >= (PAGE_SIZE - 6)) >>>>> break; >>>>> } >>>>> i += sprintf(&buf[i], "\n"); >>>> >>>> Acked-by: Viresh Kumar <viresh.kumar@linaro.org> >>> >>> Applied as 5.20 material, thanks! >> >> And dropped, because it has been superseded by this one: >> >> https://patchwork.kernel.org/project/linux-pm/patch/b9fa08171c09343ace94a7343553a4bee4695c90.1653565641.git.viresh.kumar@linaro.org/ > > The $Subject patch is still required AFAICT, it is fixing a different problem. > Though it needs to be rebased on top of your branch now. > What can i do for this patch? -- Yi
On 18-06-22, 18:32, yiyang (D) wrote:
> What can i do for this patch?
Resend it after rebasing on pm/linux-next.
diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c index 1f6667ce43bd..60c005c9961e 100644 --- a/drivers/cpufreq/cpufreq.c +++ b/drivers/cpufreq/cpufreq.c @@ -844,9 +844,9 @@ ssize_t cpufreq_show_cpus(const struct cpumask *mask, char *buf) for_each_cpu(cpu, mask) { if (i) - i += scnprintf(&buf[i], (PAGE_SIZE - i - 2), " "); - i += scnprintf(&buf[i], (PAGE_SIZE - i - 2), "%u", cpu); - if (i >= (PAGE_SIZE - 5)) + i += scnprintf(&buf[i], (PAGE_SIZE - i - 1), " "); + i += scnprintf(&buf[i], (PAGE_SIZE - i - 1), "%u", cpu); + if (i >= (PAGE_SIZE - 6)) break; } i += sprintf(&buf[i], "\n");
Function scnprintf() would reserve space for the trailing '\0' and return value is the number of characters written into buf not including the trailing '\0'. internally meaning the next scnprintf() would write begin the trailing '\0'. The code specifying "PAGE_SIZE - i - 2" here is trying to reserve space for "\n\0" which would cause scnprintf() to reserve an additional byte making the tail of the buf looks like this: "\n\0\0". Thus. we should reserve only the space for one '\0'. passing in "PAGE_SIZE - i - 1". Additionally, each iteration would replace the trailing '\0' from the last iteration with a space, and append 4 additional bytes to the string making it a total of 5 additional bytes. That means we should stop printing into the buffer if the remaining size is less than 7 bytes(1 for the ' ', 4 for the %u and 2 for the tailing "\n\0") Signed-off-by: Yi Yang <yiyang13@huawei.com> --- drivers/cpufreq/cpufreq.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)