| Message ID | 20200807161906.6d119d2e@endymion |
|---|---|
| State | Superseded |
| Delegated to: | Bartosz Golaszewski |
| Headers | show |
| Series | [1/2] eeprom: at24: Add support for the Sony VAIO EEPROMs | expand |
On Fri, Aug 7, 2020 at 4:19 PM Jean Delvare <jdelvare@suse.de> wrote: > > The elegant code in at24_read() has the drawback that we now need > to make a copy of all parameters to pass them to the post-processing > callback function if there is one. Rewrite the loop in such a way that > the parameters are not modified, so saving them is no longer needed. > > Signed-off-by: Jean Delvare <jdelvare@suse.de> > Cc: Bartosz Golaszewski <bgolaszewski@baylibre.com> > Cc: Arnd Bergmann <arnd@arndb.de> > Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> > --- > This has the drawback of creating an asymetry with at24_write(), so > I'm not 100% if we want to apply this. If anyone has a better idea, > please let me know. > > drivers/misc/eeprom/at24.c | 13 +++++-------- > 1 file changed, 5 insertions(+), 8 deletions(-) > > --- linux-5.7.orig/drivers/misc/eeprom/at24.c 2020-08-07 14:23:39.882191500 +0200 > +++ linux-5.7/drivers/misc/eeprom/at24.c 2020-08-07 14:28:39.039360687 +0200 > @@ -423,10 +423,7 @@ static int at24_read(void *priv, unsigne > struct at24_data *at24; > struct device *dev; > char *buf = val; > - int ret; > - unsigned int orig_off = off; > - char *orig_buf = buf; > - size_t orig_count = count; > + int i, ret; > > at24 = priv; > dev = at24_base_client_dev(at24); > @@ -449,15 +446,15 @@ static int at24_read(void *priv, unsigne > */ > mutex_lock(&at24->lock); > > + i = 0; Hi Jean, I think doing: for (i = 0; count; i += ret, count -= ret) would be even more elegant, don't you think? Bartosz > while (count) { > - ret = at24_regmap_read(at24, buf, off, count); > + ret = at24_regmap_read(at24, buf + i, off + i, count); > if (ret < 0) { > mutex_unlock(&at24->lock); > pm_runtime_put(dev); > return ret; > } > - buf += ret; > - off += ret; > + i += ret; > count -= ret; > } > > @@ -466,7 +463,7 @@ static int at24_read(void *priv, unsigne > pm_runtime_put(dev); > > if (unlikely(at24->read_post)) > - at24->read_post(orig_off, orig_buf, orig_count); > + at24->read_post(off, buf, i); > > return 0; > } > > -- > Jean Delvare > SUSE L3 Support
Hi Bartosz, On Tue, 18 Aug 2020 15:46:36 +0200, Bartosz Golaszewski wrote: > On Fri, Aug 7, 2020 at 4:19 PM Jean Delvare <jdelvare@suse.de> wrote: > > The elegant code in at24_read() has the drawback that we now need > > to make a copy of all parameters to pass them to the post-processing > > callback function if there is one. Rewrite the loop in such a way that > > the parameters are not modified, so saving them is no longer needed. > > > > Signed-off-by: Jean Delvare <jdelvare@suse.de> > > Cc: Bartosz Golaszewski <bgolaszewski@baylibre.com> > > Cc: Arnd Bergmann <arnd@arndb.de> > > Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> > > --- > > This has the drawback of creating an asymetry with at24_write(), so > > I'm not 100% if we want to apply this. If anyone has a better idea, > > please let me know. > > > > drivers/misc/eeprom/at24.c | 13 +++++-------- > > 1 file changed, 5 insertions(+), 8 deletions(-) > > > > --- linux-5.7.orig/drivers/misc/eeprom/at24.c 2020-08-07 14:23:39.882191500 +0200 > > +++ linux-5.7/drivers/misc/eeprom/at24.c 2020-08-07 14:28:39.039360687 +0200 > > (...) > > @@ -449,15 +446,15 @@ static int at24_read(void *priv, unsigne > > */ > > mutex_lock(&at24->lock); > > > > + i = 0; > > Hi Jean, > > I think doing: > > for (i = 0; count; i += ret, count -= ret) > > would be even more elegant, don't you think? Definitely. I'll change the code that way and send v2 of the patch, thank you for the suggestion. > Bartosz > > > while (count) { > > - ret = at24_regmap_read(at24, buf, off, count); > > + ret = at24_regmap_read(at24, buf + i, off + i, count); > > if (ret < 0) { > > mutex_unlock(&at24->lock); > > pm_runtime_put(dev); > > return ret; > > } > > - buf += ret; > > - off += ret; > > + i += ret; > > count -= ret; > > } > >
--- linux-5.7.orig/drivers/misc/eeprom/at24.c 2020-08-07 14:23:39.882191500 +0200 +++ linux-5.7/drivers/misc/eeprom/at24.c 2020-08-07 14:28:39.039360687 +0200 @@ -423,10 +423,7 @@ static int at24_read(void *priv, unsigne struct at24_data *at24; struct device *dev; char *buf = val; - int ret; - unsigned int orig_off = off; - char *orig_buf = buf; - size_t orig_count = count; + int i, ret; at24 = priv; dev = at24_base_client_dev(at24); @@ -449,15 +446,15 @@ static int at24_read(void *priv, unsigne */ mutex_lock(&at24->lock); + i = 0; while (count) { - ret = at24_regmap_read(at24, buf, off, count); + ret = at24_regmap_read(at24, buf + i, off + i, count); if (ret < 0) { mutex_unlock(&at24->lock); pm_runtime_put(dev); return ret; } - buf += ret; - off += ret; + i += ret; count -= ret; } @@ -466,7 +463,7 @@ static int at24_read(void *priv, unsigne pm_runtime_put(dev); if (unlikely(at24->read_post)) - at24->read_post(orig_off, orig_buf, orig_count); + at24->read_post(off, buf, i); return 0; }
The elegant code in at24_read() has the drawback that we now need to make a copy of all parameters to pass them to the post-processing callback function if there is one. Rewrite the loop in such a way that the parameters are not modified, so saving them is no longer needed. Signed-off-by: Jean Delvare <jdelvare@suse.de> Cc: Bartosz Golaszewski <bgolaszewski@baylibre.com> Cc: Arnd Bergmann <arnd@arndb.de> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> --- This has the drawback of creating an asymetry with at24_write(), so I'm not 100% if we want to apply this. If anyone has a better idea, please let me know. drivers/misc/eeprom/at24.c | 13 +++++-------- 1 file changed, 5 insertions(+), 8 deletions(-)