From: Trent Piepho <tpiepho@impinj.com>
To: "agust@denx.de" <agust@denx.de>,
"linux-spi@vger.kernel.org" <linux-spi@vger.kernel.org>,
"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
"linux-fpga@vger.kernel.org" <linux-fpga@vger.kernel.org>,
"linux-usb@vger.kernel.org" <linux-usb@vger.kernel.org>
Cc: "gregkh@linuxfoundation.org" <gregkh@linuxfoundation.org>,
"broonie@kernel.org" <broonie@kernel.org>,
"atull@kernel.org" <atull@kernel.org>,
"mdf@kernel.org" <mdf@kernel.org>
Subject: Re: [PATCH v2 1/3] usb: misc: add driver for FT232H based FPGA configuration devices
Date: Tue, 20 Nov 2018 00:56:13 +0000 [thread overview]
Message-ID: <1542675372.30311.573.camel@impinj.com> (raw)
In-Reply-To: <20181120002821.12794-2-agust@denx.de>
On Tue, 2018-11-20 at 01:28 +0100, Anatolij Gustschin wrote:
> Add USB interface driver for ARRI FPGA configuration devices based on
> FTDI FT232H chip. Depending on USB PID the driver registers different
> platform devices describing an FPGA configuration interface.
Is ARRI different than Arria?
> +/* Use baudrate calculation borrowed from libftdi */
> +static int ftdi_to_clkbits(int baudrate, unsigned int clk, int clk_div,
Linux uses unsigned values for clocks. Does it make any sense to mix
the unsigned clk with signed values? Seems like baudrate and clk_div
should also be unsigned.
> + unsigned long *encoded_divisor)
unsigned long is an odd choice here. Is there any to reason to use an
unsigned long to store the result of right shifting a signed int
(best_div)? It can't be longer than a int, but it can be negative.
> +{
> + static const char frac_code[8] = { 0, 3, 2, 4, 1, 5, 6, 7 };
> + int best_baud = 0;
> + int div, best_div;
> +
> + if (baudrate >= clk / clk_div) {
> + *encoded_divisor = 0;
> + best_baud = clk / clk_div;
> + } else if (baudrate >= clk / (clk_div + clk_div / 2)) {
> + *encoded_divisor = 1;
> + best_baud = clk / (clk_div + clk_div / 2);
> + } else if (baudrate >= clk / (2 * clk_div)) {
> + *encoded_divisor = 2;
> + best_baud = clk / (2 * clk_div);
> + } else {
> + /*
> + * Divide by 16 to have 3 fractional bits and
> + * one bit for rounding
> + */
> + div = clk * 16 / clk_div / baudrate;
>
> + if (div & 1) /* Decide if to round up or down */
> + best_div = div / 2 + 1;
> + else
> + best_div = div / 2;
In Linux we would write:
best_div = DIV_ROUND_UP(div, 2);
Though I think you can combine that with the above to get:
best_div = DIV_ROUND_CLOSEST(clk * 8 / clk_div, baudrate);
That what the above is trying to accomplish in a round about way
> + if (best_div > 0x20000)
> + best_div = 0x1ffff;
Looks like the above was probably supposed to be >=
> + best_baud = clk * 16 / clk_div / best_div;
> + if (best_baud & 1) /* Decide if to round up or down */
> + best_baud = best_baud / 2 + 1;
> + else
> + best_baud = best_baud / 2;
Again, looks like a complicated way to round to the nearest.
> + *encoded_divisor = (best_div >> 3) |
> + (frac_code[best_div & 0x7] << 14);
> + }
> + return best_baud;
>
WARNING: multiple messages have this Message-ID (diff)
From: Trent Piepho <tpiepho@impinj.com>
To: "agust@denx.de" <agust@denx.de>,
"linux-spi@vger.kernel.org" <linux-spi@vger.kernel.org>,
"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
"linux-fpga@vger.kernel.org" <linux-fpga@vger.kernel.org>,
"linux-usb@vger.kernel.org" <linux-usb@vger.kernel.org>
Cc: "gregkh@linuxfoundation.org" <gregkh@linuxfoundation.org>,
"broonie@kernel.org" <broonie@kernel.org>,
"atull@kernel.org" <atull@kernel.org>,
"mdf@kernel.org" <mdf@kernel.org>
Subject: [v2,1/3] usb: misc: add driver for FT232H based FPGA configuration devices
Date: Tue, 20 Nov 2018 00:56:13 +0000 [thread overview]
Message-ID: <1542675372.30311.573.camel@impinj.com> (raw)
On Tue, 2018-11-20 at 01:28 +0100, Anatolij Gustschin wrote:
> Add USB interface driver for ARRI FPGA configuration devices based on
> FTDI FT232H chip. Depending on USB PID the driver registers different
> platform devices describing an FPGA configuration interface.
Is ARRI different than Arria?
> +/* Use baudrate calculation borrowed from libftdi */
> +static int ftdi_to_clkbits(int baudrate, unsigned int clk, int clk_div,
Linux uses unsigned values for clocks. Does it make any sense to mix
the unsigned clk with signed values? Seems like baudrate and clk_div
should also be unsigned.
> + unsigned long *encoded_divisor)
unsigned long is an odd choice here. Is there any to reason to use an
unsigned long to store the result of right shifting a signed int
(best_div)? It can't be longer than a int, but it can be negative.
> +{
> + static const char frac_code[8] = { 0, 3, 2, 4, 1, 5, 6, 7 };
> + int best_baud = 0;
> + int div, best_div;
> +
> + if (baudrate >= clk / clk_div) {
> + *encoded_divisor = 0;
> + best_baud = clk / clk_div;
> + } else if (baudrate >= clk / (clk_div + clk_div / 2)) {
> + *encoded_divisor = 1;
> + best_baud = clk / (clk_div + clk_div / 2);
> + } else if (baudrate >= clk / (2 * clk_div)) {
> + *encoded_divisor = 2;
> + best_baud = clk / (2 * clk_div);
> + } else {
> + /*
> + * Divide by 16 to have 3 fractional bits and
> + * one bit for rounding
> + */
> + div = clk * 16 / clk_div / baudrate;
>
> + if (div & 1) /* Decide if to round up or down */
> + best_div = div / 2 + 1;
> + else
> + best_div = div / 2;
In Linux we would write:
best_div = DIV_ROUND_UP(div, 2);
Though I think you can combine that with the above to get:
best_div = DIV_ROUND_CLOSEST(clk * 8 / clk_div, baudrate);
That what the above is trying to accomplish in a round about way
> + if (best_div > 0x20000)
> + best_div = 0x1ffff;
Looks like the above was probably supposed to be >=
> + best_baud = clk * 16 / clk_div / best_div;
> + if (best_baud & 1) /* Decide if to round up or down */
> + best_baud = best_baud / 2 + 1;
> + else
> + best_baud = best_baud / 2;
Again, looks like a complicated way to round to the nearest.
> + *encoded_divisor = (best_div >> 3) |
> + (frac_code[best_div & 0x7] << 14);
> + }
> + return best_baud;
>
next prev parent reply other threads:[~2018-11-20 11:22 UTC|newest]
Thread overview: 17+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-11-20 0:28 [PATCH v2 0/3] Add support for ARRI FPGA configuration Anatolij Gustschin
2018-11-20 0:28 ` [PATCH v2 1/3] usb: misc: add driver for FT232H based FPGA configuration devices Anatolij Gustschin
2018-11-20 0:28 ` [v2,1/3] " Anatolij Gustschin
2018-11-20 0:56 ` Trent Piepho [this message]
2018-11-20 0:56 ` Trent Piepho
2018-11-20 14:49 ` [PATCH v2 1/3] " Anatolij Gustschin
2018-11-20 14:49 ` [v2,1/3] " Anatolij Gustschin
2018-11-20 0:28 ` [PATCH v2 2/3] spi: add FTDI MPSSE SPI controller driver Anatolij Gustschin
2018-11-20 0:28 ` [v2,2/3] " Anatolij Gustschin
2018-11-21 12:42 ` [PATCH v2 2/3] " Mark Brown
2018-11-21 12:42 ` [v2,2/3] " Mark Brown
2018-11-27 0:21 ` [PATCH v2 2/3] " Anatolij Gustschin
2018-11-27 0:21 ` [v2,2/3] " Anatolij Gustschin
2018-11-28 9:14 ` [PATCH v2 2/3] " Mark Brown
2018-11-28 9:14 ` [v2,2/3] " Mark Brown
2018-11-20 0:28 ` [PATCH v2 3/3] fpga: Add fpga manager driver for ARRI Altera FPP Anatolij Gustschin
2018-11-20 0:28 ` [v2,3/3] " Anatolij Gustschin
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1542675372.30311.573.camel@impinj.com \
--to=tpiepho@impinj.com \
--cc=agust@denx.de \
--cc=atull@kernel.org \
--cc=broonie@kernel.org \
--cc=gregkh@linuxfoundation.org \
--cc=linux-fpga@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-spi@vger.kernel.org \
--cc=linux-usb@vger.kernel.org \
--cc=mdf@kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.