Re: [PATCH v4 1/4] lib: bitmap: introduce bitmap_print_to_buf

From: Yury Norov
Date: Thu Jun 17 2021 - 16:54:08 EST


On Thu, Jun 17, 2021 at 10:19:07PM +1200, Barry Song wrote:
> From: Tian Tao <tiantao6@xxxxxxxxxxxxx>
>
> New API bitmap_print_to_buf() with bin_attribute to avoid maskp
> exceeding PAGE_SIZE. bitmap_print_to_pagebuf() is a special case
> of bitmap_print_to_buf(), so in bitmap_print_to_pagebuf() call
> bitmap_print_to_buf().
>
> Signed-off-by: Tian Tao <tiantao6@xxxxxxxxxxxxx>
> Reviewed-by: Jonathan Cameron <jonathan.cameron@xxxxxxxxxx>
> Cc: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
> Cc: Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx>
> Cc: Randy Dunlap <rdunlap@xxxxxxxxxxxxx>
> Cc: Stefano Brivio <sbrivio@xxxxxxxxxx>
> Cc: Alexander Gordeev <agordeev@xxxxxxxxxxxxx>
> Cc: "Ma, Jianpeng" <jianpeng.ma@xxxxxxxxx>
> Cc: Yury Norov <yury.norov@xxxxxxxxx>
> Cc: Valentin Schneider <valentin.schneider@xxxxxxx>
> Cc: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
> Cc: Daniel Bristot de Oliveira <bristot@xxxxxxxxxx>
> ---
> include/linux/bitmap.h | 2 ++
> include/linux/cpumask.h | 21 +++++++++++++++++++++
> lib/bitmap.c | 37 +++++++++++++++++++++++++++++++++++--
> 3 files changed, 58 insertions(+), 2 deletions(-)
>
> diff --git a/include/linux/bitmap.h b/include/linux/bitmap.h
> index a36cfcec4e77..0de6effa2797 100644
> --- a/include/linux/bitmap.h
> +++ b/include/linux/bitmap.h
> @@ -226,6 +226,8 @@ void bitmap_copy_le(unsigned long *dst, const unsigned long *src, unsigned int n
> unsigned int bitmap_ord_to_pos(const unsigned long *bitmap, unsigned int ord, unsigned int nbits);
> int bitmap_print_to_pagebuf(bool list, char *buf,
> const unsigned long *maskp, int nmaskbits);
> +int bitmap_print_to_buf(bool list, char *buf, const unsigned long *maskp,
> + int nmaskbits, loff_t off, size_t count);
>
> #define BITMAP_FIRST_WORD_MASK(start) (~0UL << ((start) & (BITS_PER_LONG - 1)))
> #define BITMAP_LAST_WORD_MASK(nbits) (~0UL >> (-(nbits) & (BITS_PER_LONG - 1)))
> diff --git a/include/linux/cpumask.h b/include/linux/cpumask.h
> index bfc4690de4f4..1bef69e4b950 100644
> --- a/include/linux/cpumask.h
> +++ b/include/linux/cpumask.h
> @@ -983,6 +983,27 @@ cpumap_print_to_pagebuf(bool list, char *buf, const struct cpumask *mask)
> nr_cpu_ids);
> }
>
> +/**
> + * cpumap_print_to_buf - copies the cpumask into the buffer either
> + * as comma-separated list of cpus or hex values of cpumask
> + * @list: indicates whether the cpumap must be list
> + * @mask: the cpumask to copy
> + * @buf: the buffer to copy into
> + * @off: in the string from which we are copying, We copy to @buf + off
> + * @count: the maximum number of bytes to print
> + *
> + * The role of cpumap_print_to_buf() and cpumap_print_to_pagebuf() is
> + * the same, the difference is that buf of bitmap_print_to_buf()
> + * can be more than one pagesize.
> + */
> +static inline ssize_t
> +cpumap_print_to_buf(bool list, char *buf, const struct cpumask *mask,
> + loff_t off, size_t count)
> +{
> + return bitmap_print_to_buf(list, buf, cpumask_bits(mask),
> + nr_cpu_ids, off, count);
> +}
> +
> #if NR_CPUS <= BITS_PER_LONG
> #define CPU_MASK_ALL \
> (cpumask_t) { { \
> diff --git a/lib/bitmap.c b/lib/bitmap.c
> index 74ceb02f45e3..ed7ef6e2edca 100644
> --- a/lib/bitmap.c
> +++ b/lib/bitmap.c
> @@ -461,6 +461,40 @@ int bitmap_parse_user(const char __user *ubuf,
> }
> EXPORT_SYMBOL(bitmap_parse_user);
>
> +/**
> + * bitmap_print_to_buf - convert bitmap to list or hex format ASCII string
> + * @list: indicates whether the bitmap must be list
> + * @buf: the kernel space buffer to read to
> + * @maskp: pointer to bitmap to convert
> + * @nmaskbits: size of bitmap, in bits
> + * @off: offset in data buffer below

Below there's a @count parameter. Maybe above?

> + * @count: the maximum number of bytes to print
> + *
> + * The role of bitmap_print_to_buf() and bitmap_print_to_pagebuf() is
> + * the same, the difference is that buf of bitmap_print_to_buf()
> + * can be more than one pagesize.

You're introducing a function that duplicates an existing function.
I can understand it only if there is important performance and/or
security consideration. Otherwise, I'd recommend you to extend
bitmap_print_to_pagebuf() with one more parameter.

> + */
> +int bitmap_print_to_buf(bool list, char *buf, const unsigned long *maskp,
> + int nmaskbits, loff_t off, size_t count)
> +{
> + const char *fmt = list ? "%*pbl\n" : "%*pb\n";
> + ssize_t size;
> + void *data;
> +
> + if (off == LLONG_MAX && count == PAGE_SIZE - offset_in_page(buf))
> + return scnprintf(buf, count, fmt, nmaskbits, maskp);

Agree with Greg. This requires explanation.

> +
> + data = kasprintf(GFP_KERNEL, fmt, nmaskbits, maskp);
> + if (!data)
> + return -ENOMEM;

So here:
- you make user of your function to guess about how much memory he needs
to provide to the function;
- then kasprintf() allocates enough memory for you automatically;
- then you copy the data to potentially insufficient buffer;
- and then free a valid 'data' buffer provided by kasprintf(), even if
the following memory_read_from_buffer() fails;
- goto 'So here'.

It doesn't look like a good design.

kasprintf() does everything you need. Why don't you use it directly in
your driver?

> +
> + size = memory_read_from_buffer(buf, count, &off, data, strlen(data) + 1);

strlen() is O(n), just as memory_read_from_buffer(). In this line you introduce
a 2*O(n) operation. Consider strncpy().

> + kfree(data);
> +
> + return size;
> +}
> +EXPORT_SYMBOL(bitmap_print_to_buf);
> +
> /**
> * bitmap_print_to_pagebuf - convert bitmap to list or hex format ASCII string
> * @list: indicates whether the bitmap must be list
> @@ -482,8 +516,7 @@ int bitmap_print_to_pagebuf(bool list, char *buf, const unsigned long *maskp,
> {
> ptrdiff_t len = PAGE_SIZE - offset_in_page(buf);
>
> - return list ? scnprintf(buf, len, "%*pbl\n", nmaskbits, maskp) :
> - scnprintf(buf, len, "%*pb\n", nmaskbits, maskp);
> + return bitmap_print_to_buf(list, buf, maskp, nmaskbits, LLONG_MAX, len);
> }
> EXPORT_SYMBOL(bitmap_print_to_pagebuf);
>
> --
> 2.25.1