Re: [PATCH] mtd: dataflash: replace msleep with usleep_range

2018-04-26 Thread Boris Brezillon
On Wed, 18 Apr 2018 11:26:26 +0200 Luca Ellero wrote: > Since msleep is based on jiffies, this 3 ms sleep becomes actually 20 ms. > Worst of all, since this sleep is used in a loop when writing, a single page > write (256 to 1024 bytes) causes 17 ms extra time. >

Re: [PATCH] mtd: dataflash: replace msleep with usleep_range

2018-04-26 Thread Boris Brezillon
On Wed, 18 Apr 2018 11:26:26 +0200 Luca Ellero wrote: > Since msleep is based on jiffies, this 3 ms sleep becomes actually 20 ms. > Worst of all, since this sleep is used in a loop when writing, a single page > write (256 to 1024 bytes) causes 17 ms extra time. > When writing large files (for

[PATCH] mtd: dataflash: replace msleep with usleep_range

2018-04-18 Thread Luca Ellero
Since msleep is based on jiffies, this 3 ms sleep becomes actually 20 ms. Worst of all, since this sleep is used in a loop when writing, a single page write (256 to 1024 bytes) causes 17 ms extra time. When writing large files (for example u-boot is usually 512 KB) this delay adds up to minutes.

[PATCH] mtd: dataflash: replace msleep with usleep_range

2018-04-18 Thread Luca Ellero
Since msleep is based on jiffies, this 3 ms sleep becomes actually 20 ms. Worst of all, since this sleep is used in a loop when writing, a single page write (256 to 1024 bytes) causes 17 ms extra time. When writing large files (for example u-boot is usually 512 KB) this delay adds up to minutes.