| Message ID | 20231130191050.3165862-2-hugo@hugovil.com |
|---|---|
| State | New |
| Headers | show |
| Series | serial: sc16is7xx and max310x: regmap fixes and improvements | expand |
On Thu, 7 Dec 2023 10:44:33 +0900 Greg KH <gregkh@linuxfoundation.org> wrote: > On Thu, Nov 30, 2023 at 02:10:43PM -0500, Hugo Villeneuve wrote: > > From: Hugo Villeneuve <hvilleneuve@dimonoff.com> > > > > Change snprint format specifier from %d to %u since port_id is unsigned. > > > > Fixes: 3837a0379533 ("serial: sc16is7xx: improve regmap debugfs by using one regmap per port") > > Cc: stable@vger.kernel.org # 6.1.x: 3837a03 serial: sc16is7xx: improve regmap debugfs by using one regmap per port > > Suggested-by: Andy Shevchenko <andy.shevchenko@gmail.com> > > Signed-off-by: Hugo Villeneuve <hvilleneuve@dimonoff.com> > > --- > > I did not originally add a "Cc: stable" tag for commit 3837a0379533 ("serial: sc16is7xx: improve regmap debugfs by using one regmap per port") > > as it was intended only to improve debugging using debugfs. But > > since then, I have been able to confirm that it also fixes a long standing > > bug in our system where the Tx interrupt are no longer enabled at some > > point when transmitting large RS-485 paquets (> 64 bytes, which is the size > > of the FIFO). I have been investigating why, but so far I haven't found the > > exact cause, altough I suspect it has something to do with regmap caching. > > Therefore, I have added it as a prerequisite for this patch so that it is > > automatically added to the stable kernels. > > Looks like the 0-day test bot found problems with this, so I'll hold off > on taking this patch and the rest of the series until that's fixed up > with a new version of this series. No problem, I am on it. Hugo
On Thu, Dec 7, 2023 at 7:52 PM Hugo Villeneuve <hugo@hugovil.com> wrote: > On Wed, 6 Dec 2023 14:29:39 +0800 > kernel test robot <lkp@intel.com> wrote: ... > > drivers/tty/serial/sc16is7xx.c: In function 'sc16is7xx_i2c_probe': > > >> drivers/tty/serial/sc16is7xx.c:1703:41: warning: '%u' directive output may be truncated writing between 1 and 10 bytes into a region of size 2 [-Wformat-truncation=] > > 1703 | snprintf(buf, sizeof(buf), "port%u", port_id); > > | ^~ > > In function 'sc16is7xx_regmap_name', > > inlined from 'sc16is7xx_i2c_probe' at drivers/tty/serial/sc16is7xx.c:1805:17: > > drivers/tty/serial/sc16is7xx.c:1703:36: note: directive argument in the range [0, 4294967294] > > 1703 | snprintf(buf, sizeof(buf), "port%u", port_id); > > | ^~~~~~~~ > > drivers/tty/serial/sc16is7xx.c:1703:9: note: 'snprintf' output between 6 and 15 bytes into a destination of size 6 > > 1703 | snprintf(buf, sizeof(buf), "port%u", port_id); > > | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ > > Hi, > the only solution I could find is to add this line just before snprintf: > > BUG_ON(port_id > MAX310X_MAX_PORTS); > > it allows us to have the smallest buffer size possible. > > One other solution would be to change port_id from "unsigned int" > to "u8", and increase the buffer by an additional 2 bytes to silence > the warning, but then wasting 2 bytes for each channel, like so: I didn't get this. It's a buffer that is rewritten on each port (why is it even static?). Just make sure it's enough for any given number and drop the static. ... While at it, can you look at the following items to improve? - sc16is7xx_alloc_line() can be updated to use IDA framework - move return xxx; to the default cases in a few functions - if (div > 0xffff) { --> if (div >= BIT(16)) { as it better shows why the limit is that (we have only 16 bits for the divider) - do {} while (0) in the sc16is7xx_port_irq, WTH?! - while (1) { -- do { } while (keep_polling); in sc16is7xx_irq() - use in_range() in sc16is7xx_setup_mctrl_ports() ? (maybe not, dunno) - for (i--; i >= 0; i--) { --> while (i--) { - use spi_get_device_match_data() and i2c_get_match_data() - 15000000 --> 15 * HZ_PER_MHZ ? - dropping MODULE_ALIAS (and fix the ID tables, _if_ needed) - split the code to the core / main + SPI + I2C glue drivers * These just come on the first glance at the code, perhaps there is more room to improve. -- With Best Regards, Andy Shevchenko
From: Hugo Villeneuve > Sent: 07 December 2023 17:53 ... > > kernel test robot noticed the following build warnings: > > > > [auto build test WARNING on d804987153e7bedf503f8e4ba649afe52cfd7f6d] > > > > url: https://github.com/intel-lab-lkp/linux/commits/Hugo-Villeneuve/serial-sc16is7xx-fix- > snprintf-format-specifier-in-sc16is7xx_regmap_name/20231201-031413 > > base: d804987153e7bedf503f8e4ba649afe52cfd7f6d > > patch link: https://lore.kernel.org/r/20231130191050.3165862-2-hugo%40hugovil.com > > patch subject: [PATCH 1/7] serial: sc16is7xx: fix snprintf format specifier in > sc16is7xx_regmap_name() > > config: x86_64-buildonly-randconfig-001-20231201 (https://download.01.org/0day- > ci/archive/20231206/202312061443.Cknef7Uq-lkp@intel.com/config) > > compiler: gcc-12 (Debian 12.2.0-14) 12.2.0 > > reproduce (this is a W=1 build): (https://download.01.org/0day- > ci/archive/20231206/202312061443.Cknef7Uq-lkp@intel.com/reproduce) > > > > If you fix the issue in a separate patch/commit (i.e. not just a new version of > > the same patch/commit), kindly add following tags > > | Reported-by: kernel test robot <lkp@intel.com> > > | Closes: https://lore.kernel.org/oe-kbuild-all/202312061443.Cknef7Uq-lkp@intel.com/ > > > > All warnings (new ones prefixed by >>): > > > > drivers/tty/serial/sc16is7xx.c: In function 'sc16is7xx_i2c_probe': > > >> drivers/tty/serial/sc16is7xx.c:1703:41: warning: '%u' directive output may be truncated writing > between 1 and 10 bytes into a region of size 2 [-Wformat-truncation=] > > 1703 | snprintf(buf, sizeof(buf), "port%u", port_id); > > | ^~ > > In function 'sc16is7xx_regmap_name', > > inlined from 'sc16is7xx_i2c_probe' at drivers/tty/serial/sc16is7xx.c:1805:17: > > drivers/tty/serial/sc16is7xx.c:1703:36: note: directive argument in the range [0, 4294967294] > > 1703 | snprintf(buf, sizeof(buf), "port%u", port_id); > > | ^~~~~~~~ > > drivers/tty/serial/sc16is7xx.c:1703:9: note: 'snprintf' output between 6 and 15 bytes into a > destination of size 6 > > 1703 | snprintf(buf, sizeof(buf), "port%u", port_id); > > | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ > > Hi, > the only solution I could find is to add this line just before snprintf: > > BUG_ON(port_id > MAX310X_MAX_PORTS); > > it allows us to have the smallest buffer size possible. Or "port%c", '0' + port_id); Or maybe: size_t buflen = sizeof (buf); OPTIMIZER_HIDE_VAR(buflen); snprintf(buf, buflen, fmt, args); See https://godbolt.org/z/Wjz3xG5c4 Maybe there should be snprintf_may_truncate() (etc) in one of the headers. David - Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK Registration No: 1397386 (Wales)
On Thu, 7 Dec 2023 20:24:45 +0200 Andy Shevchenko <andy.shevchenko@gmail.com> wrote: > On Thu, Dec 7, 2023 at 7:52 PM Hugo Villeneuve <hugo@hugovil.com> wrote: > > On Wed, 6 Dec 2023 14:29:39 +0800 > > kernel test robot <lkp@intel.com> wrote: > > ... > > > > drivers/tty/serial/sc16is7xx.c: In function 'sc16is7xx_i2c_probe': > > > >> drivers/tty/serial/sc16is7xx.c:1703:41: warning: '%u' directive output may be truncated writing between 1 and 10 bytes into a region of size 2 [-Wformat-truncation=] > > > 1703 | snprintf(buf, sizeof(buf), "port%u", port_id); > > > | ^~ > > > In function 'sc16is7xx_regmap_name', > > > inlined from 'sc16is7xx_i2c_probe' at drivers/tty/serial/sc16is7xx.c:1805:17: > > > drivers/tty/serial/sc16is7xx.c:1703:36: note: directive argument in the range [0, 4294967294] > > > 1703 | snprintf(buf, sizeof(buf), "port%u", port_id); > > > | ^~~~~~~~ > > > drivers/tty/serial/sc16is7xx.c:1703:9: note: 'snprintf' output between 6 and 15 bytes into a destination of size 6 > > > 1703 | snprintf(buf, sizeof(buf), "port%u", port_id); > > > | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ > > > > Hi, > > the only solution I could find is to add this line just before snprintf: > > > > BUG_ON(port_id > MAX310X_MAX_PORTS); > > > > it allows us to have the smallest buffer size possible. > > > > One other solution would be to change port_id from "unsigned int" > > to "u8", and increase the buffer by an additional 2 bytes to silence > > the warning, but then wasting 2 bytes for each channel, like so: > > I didn't get this. It's a buffer that is rewritten on each port (why > is it even static?). Just make sure it's enough for any given number > and drop the static. > > ... > > While at it, can you look at the following items to improve? > - sc16is7xx_alloc_line() can be updated to use IDA framework > - move return xxx; to the default cases in a few functions > - if (div > 0xffff) { --> if (div >= BIT(16)) { as it better shows why > the limit is that (we have only 16 bits for the divider) > - do {} while (0) in the sc16is7xx_port_irq, WTH?! > - while (1) { -- do { } while (keep_polling); in sc16is7xx_irq() > - use in_range() in sc16is7xx_setup_mctrl_ports() ? (maybe not, dunno) > - for (i--; i >= 0; i--) { --> while (i--) { > - use spi_get_device_match_data() and i2c_get_match_data() > - 15000000 --> 15 * HZ_PER_MHZ ? > - dropping MODULE_ALIAS (and fix the ID tables, _if_ needed) > - split the code to the core / main + SPI + I2C glue drivers > > * These just come on the first glance at the code, perhaps there is > more room to improve. Hi Andy, just to let you know that I have implemented almost all of the fixes / improvements. I will submit them once V2 of this current series lands in Greg's next tree. However, for sc16is7xx_alloc_line(), I looked at using the IDA framework but it doesn't seem possible because there is no IDA function to search if a bit is set, which is a needed functionality. Hugo Villeneuve
On Tue, Dec 12, 2023 at 10:03 PM Hugo Villeneuve <hugo@hugovil.com> wrote: > On Thu, 7 Dec 2023 20:24:45 +0200 > Andy Shevchenko <andy.shevchenko@gmail.com> wrote: > > On Thu, Dec 7, 2023 at 7:52 PM Hugo Villeneuve <hugo@hugovil.com> wrote: ... > > While at it, can you look at the following items to improve? > > - sc16is7xx_alloc_line() can be updated to use IDA framework > > - move return xxx; to the default cases in a few functions > > - if (div > 0xffff) { --> if (div >= BIT(16)) { as it better shows why > > the limit is that (we have only 16 bits for the divider) > > - do {} while (0) in the sc16is7xx_port_irq, WTH?! > > - while (1) { -- do { } while (keep_polling); in sc16is7xx_irq() > > - use in_range() in sc16is7xx_setup_mctrl_ports() ? (maybe not, dunno) > > - for (i--; i >= 0; i--) { --> while (i--) { > > - use spi_get_device_match_data() and i2c_get_match_data() > > - 15000000 --> 15 * HZ_PER_MHZ ? > > - dropping MODULE_ALIAS (and fix the ID tables, _if_ needed) > > - split the code to the core / main + SPI + I2C glue drivers > > > > * These just come on the first glance at the code, perhaps there is > > more room to improve. > > Hi Andy, > just to let you know that I have implemented almost all of the fixes / > improvements. I will submit them once V2 of this current series > lands in Greg's next tree. Hooray! > However, for sc16is7xx_alloc_line(), I looked at using the IDA framework > but it doesn't seem possible because there is no IDA function > to search if a bit is set, which is a needed functionality. It can be done via trying to get it, but probably it's uglier than current behaviour. Okay, let's leave it as is for now.
diff --git a/drivers/tty/serial/sc16is7xx.c b/drivers/tty/serial/sc16is7xx.c index 10e90a7774f0..8e5baf2f6ec6 100644 --- a/drivers/tty/serial/sc16is7xx.c +++ b/drivers/tty/serial/sc16is7xx.c @@ -1700,7 +1700,7 @@ static const char *sc16is7xx_regmap_name(unsigned int port_id) { static char buf[6]; - snprintf(buf, sizeof(buf), "port%d", port_id); + snprintf(buf, sizeof(buf), "port%u", port_id); return buf; }