Re: [PATCH 03/03] staging: dgap: remove more unneeded brd-state states
On Fri, Mar 28, 2014 at 09:08:34AM -0400, Mark Hounschell wrote: > On 03/28/2014 07:34 AM, Dan Carpenter wrote: > > These patches are fine and they were applied already. > > > > On Wed, Mar 12, 2014 at 12:50:55PM -0400, Mark Hounschell wrote: > >> @@ -4368,15 +4364,16 @@ static void dgap_do_bios_load(struct board_t *brd, > >> uchar __user *ubios, int len) > >> /* > >> * Checks to see if the BIOS completed running on the card. > >> */ > >> -static void dgap_do_wait_for_bios(struct board_t *brd) > >> +static int dgap_do_wait_for_bios(struct board_t *brd) > > > > I wish this funciton returned negative error codes on error. It is > > poorly named for a boolean function. > > > >> { > >>uchar *addr; > >>u16 word; > >>u16 err1; > >>u16 err2; > >> + int ret = 0; > > > > The ret variable is not needed. Replace it with zero literal for better > > readability. > > > >> @@ -4455,15 +4452,16 @@ static void dgap_do_fep_load(struct board_t *brd, > >> uchar *ufep, int len) > >> /* > >> * Waits for the FEP to report thats its ready for us to use. > >> */ > >> -static void dgap_do_wait_for_fep(struct board_t *brd) > >> +static int dgap_do_wait_for_fep(struct board_t *brd) > > > > Same as dgap_do_wait_for_bios(). > > > > Yes, they were not originally boolean functions. Would names like > dgap_test_bios and dgap_test_fep be better names? And returns of > -EIO if they fail and 0 if good? What I'm saying is that by default kernel functions return zero on success and negative error codes. If you're going to make a boolean function then the name has to be clear. if (!dgap_read_bios_is_wonderful(...)) return -ENXIO; If it returns negative error codes then the current name is fine. This isn't the kind of thing where you have to redo the function, I try not to get too nit picky for naming in staging stuff because we can redo it later anyway. It's just a comment for later consideration. But yes, the new patch looks fine. regards, dan carpenter ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
Re: [PATCH 03/03] staging: dgap: remove more unneeded brd-state states
On 03/28/2014 09:08 AM, Mark Hounschell wrote: > On 03/28/2014 07:34 AM, Dan Carpenter wrote: >> These patches are fine and they were applied already. >> >> On Wed, Mar 12, 2014 at 12:50:55PM -0400, Mark Hounschell wrote: >>> @@ -4368,15 +4364,16 @@ static void dgap_do_bios_load(struct board_t *brd, >>> uchar __user *ubios, int len) >>> /* >>> * Checks to see if the BIOS completed running on the card. >>> */ >>> -static void dgap_do_wait_for_bios(struct board_t *brd) >>> +static int dgap_do_wait_for_bios(struct board_t *brd) >> >> I wish this funciton returned negative error codes on error. It is >> poorly named for a boolean function. >> >>> { >>> uchar *addr; >>> u16 word; >>> u16 err1; >>> u16 err2; >>> + int ret = 0; >> >> The ret variable is not needed. Replace it with zero literal for better >> readability. >> >>> @@ -4455,15 +4452,16 @@ static void dgap_do_fep_load(struct board_t *brd, >>> uchar *ufep, int len) >>> /* >>> * Waits for the FEP to report thats its ready for us to use. >>> */ >>> -static void dgap_do_wait_for_fep(struct board_t *brd) >>> +static int dgap_do_wait_for_fep(struct board_t *brd) >> >> Same as dgap_do_wait_for_bios(). >> > > Yes, they were not originally boolean functions. Would names like > dgap_test_bios and dgap_test_fep be better names? And returns of > -EIO if they fail and 0 if good? > I'll just post a new patch for review and fix as required. Mark ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
Re: [PATCH 03/03] staging: dgap: remove more unneeded brd-state states
On 03/28/2014 07:34 AM, Dan Carpenter wrote: > These patches are fine and they were applied already. > > On Wed, Mar 12, 2014 at 12:50:55PM -0400, Mark Hounschell wrote: >> @@ -4368,15 +4364,16 @@ static void dgap_do_bios_load(struct board_t *brd, >> uchar __user *ubios, int len) >> /* >> * Checks to see if the BIOS completed running on the card. >> */ >> -static void dgap_do_wait_for_bios(struct board_t *brd) >> +static int dgap_do_wait_for_bios(struct board_t *brd) > > I wish this funciton returned negative error codes on error. It is > poorly named for a boolean function. > >> { >> uchar *addr; >> u16 word; >> u16 err1; >> u16 err2; >> +int ret = 0; > > The ret variable is not needed. Replace it with zero literal for better > readability. > >> @@ -4455,15 +4452,16 @@ static void dgap_do_fep_load(struct board_t *brd, >> uchar *ufep, int len) >> /* >> * Waits for the FEP to report thats its ready for us to use. >> */ >> -static void dgap_do_wait_for_fep(struct board_t *brd) >> +static int dgap_do_wait_for_fep(struct board_t *brd) > > Same as dgap_do_wait_for_bios(). > Yes, they were not originally boolean functions. Would names like dgap_test_bios and dgap_test_fep be better names? And returns of -EIO if they fail and 0 if good? Sample patch: diff --git a/drivers/staging/dgap/dgap.c b/drivers/staging/dgap/dgap.c index 4bbedae..d0e486b 100644 --- a/drivers/staging/dgap/dgap.c +++ b/drivers/staging/dgap/dgap.c @@ -194,8 +194,8 @@ static int dgap_finalize_board_init(struct board_t *brd); static void dgap_get_vpd(struct board_t *brd); static void dgap_do_reset_board(struct board_t *brd); -static int dgap_do_wait_for_bios(struct board_t *brd); -static int dgap_do_wait_for_fep(struct board_t *brd); +static int dgap_test_bios(struct board_t *brd); +static int dgap_test_fep(struct board_t *brd); static int dgap_tty_register_ports(struct board_t *brd); static int dgap_firmware_load(struct pci_dev *pdev, int card_type); @@ -890,8 +890,9 @@ static int dgap_firmware_load(struct pci_dev *pdev, int card_type) release_firmware(fw); /* Wait for BIOS to test board... */ - if (!dgap_do_wait_for_bios(brd)) - return -ENXIO; + ret = dgap_test_bios(brd) + if (ret) + return ret; } if (fw_info[card_type].fep_name) { @@ -906,8 +907,9 @@ static int dgap_firmware_load(struct pci_dev *pdev, int card_type) release_firmware(fw); /* Wait for FEP to load on board... */ - if (!dgap_do_wait_for_fep(brd)) - return -ENXIO; + ret = dgap_test_fep(brd) + if (ret) + return ret; } #ifdef DIGI_CONCENTRATORS_SUPPORTED @@ -4332,13 +4334,12 @@ static void dgap_do_bios_load(struct board_t *brd, const uchar *ubios, int len) /* * Checks to see if the BIOS completed running on the card. */ -static int dgap_do_wait_for_bios(struct board_t *brd) +static int dgap_test_bios(struct board_t *brd) { uchar *addr; u16 word; u16 err1; u16 err2; - int ret = 0; if (!brd || (brd->magic != DGAP_BOARD_MAGIC) || !brd->re_map_membase) return ret; @@ -4355,7 +4356,7 @@ static int dgap_do_wait_for_bios(struct board_t *brd) while (brd->wait_for_bios < 1000) { /* Check to see if BIOS thinks board is good. (GD). */ if (word == *(u16 *) "GD") - return 1; + return 0; msleep_interruptible(10); brd->wait_for_bios++; word = readw(addr + POSTAREA); @@ -4369,7 +4370,7 @@ static int dgap_do_wait_for_bios(struct board_t *brd) brd->state = BOARD_FAILED; brd->dpastatus = BD_NOBIOS; - return ret; + return -EIO; } /* @@ -4420,13 +4421,12 @@ static void dgap_do_fep_load(struct board_t *brd, const uchar *ufep, int len) /* * Waits for the FEP to report thats its ready for us to use. */ -static int dgap_do_wait_for_fep(struct board_t *brd) +static int dgap_test_fep(struct board_t *brd) { uchar *addr; u16 word; u16 err1; u16 err2; - int ret = 0; if (!brd || (brd->magic != DGAP_BOARD_MAGIC) || !brd->re_map_membase) return ret; @@ -4449,7 +4449,7 @@ static int dgap_do_wait_for_fep(struct board_t *brd) if (word == *(u16 *) "5A") brd->bd_flags |= BD_FEP5PLUS; - return 1; + return 0; } msleep_interruptible(10); brd->wait_for_fep++; @@ -4464,7 +4464,7 @@ static int dgap_do_wait_for_fep(struct board_t *brd) brd->state = BOARD_FAILED; brd->dpastatus = BD_NOFEP; - return ret; + return -EIO; }
Re: [PATCH 03/03] staging: dgap: remove more unneeded brd-state states
These patches are fine and they were applied already. On Wed, Mar 12, 2014 at 12:50:55PM -0400, Mark Hounschell wrote: > @@ -4368,15 +4364,16 @@ static void dgap_do_bios_load(struct board_t *brd, > uchar __user *ubios, int len) > /* > * Checks to see if the BIOS completed running on the card. > */ > -static void dgap_do_wait_for_bios(struct board_t *brd) > +static int dgap_do_wait_for_bios(struct board_t *brd) I wish this funciton returned negative error codes on error. It is poorly named for a boolean function. > { > uchar *addr; > u16 word; > u16 err1; > u16 err2; > + int ret = 0; The ret variable is not needed. Replace it with zero literal for better readability. > @@ -4455,15 +4452,16 @@ static void dgap_do_fep_load(struct board_t *brd, > uchar *ufep, int len) > /* > * Waits for the FEP to report thats its ready for us to use. > */ > -static void dgap_do_wait_for_fep(struct board_t *brd) > +static int dgap_do_wait_for_fep(struct board_t *brd) Same as dgap_do_wait_for_bios(). regards, dan carpenter ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel