| Message ID | 20220316155455.162362-3-ikegami.t@gmail.com |
|---|---|
| State | Superseded |
| Delegated to: | Vignesh R |
| Headers | show |
| Series | mtd: cfi_cmdset_0002: Use chip_ready() for write on S29GL064N | expand |
Hi Tokunori, ikegami.t@gmail.com wrote on Thu, 17 Mar 2022 00:54:54 +0900: > As pointed out by this bug report [1], buffered writes are now broken on > S29GL064N. This issue comes from a rework which switched from using chip_good() > to chip_ready(), because DQ true data 0xFF is read on S29GL064N and an error > returned by chip_good(). Vignesh, I believe you understand this issue better than I do, can you propose an improved commit log? > One way to solve the issue is to revert the change > partially to use chip_ready for S29GL064N. > > [1] https://lore.kernel.org/r/b687c259-6413-26c9-d4c9-b3afa69ea124@pengutronix.de/ > > Fixes: dfeae1073583("mtd: cfi_cmdset_0002: Change write buffer to check correct value") > Signed-off-by: Tokunori Ikegami <ikegami.t@gmail.com> > Tested-by: Ahmad Fatoum <a.fatoum@pengutronix.de> > Cc: stable@vger.kernel.org > --- > drivers/mtd/chips/cfi_cmdset_0002.c | 25 +++++++++++++++++++++---- > 1 file changed, 21 insertions(+), 4 deletions(-) > > diff --git a/drivers/mtd/chips/cfi_cmdset_0002.c b/drivers/mtd/chips/cfi_cmdset_0002.c > index e68ddf0f7fc0..6c57f85e1b8e 100644 > --- a/drivers/mtd/chips/cfi_cmdset_0002.c > +++ b/drivers/mtd/chips/cfi_cmdset_0002.c > @@ -866,6 +866,23 @@ static int __xipram chip_check(struct map_info *map, struct flchip *chip, > chip_check(map, chip, addr, &datum); \ > }) > > +static bool __xipram cfi_use_chip_ready_for_write(struct map_info *map) At the very least I would call this function: cfi_use_chip_ready_for_writes() Yet, I still don't fully get what chip_ready is versus chip_good. > +{ > + struct cfi_private *cfi = map->fldrv_priv; > + > + return cfi->mfr == CFI_MFR_AMD && cfi->id == 0x0c01; > +} > + > +static int __xipram chip_good_for_write(struct map_info *map, > + struct flchip *chip, unsigned long addr, > + map_word expected) > +{ > + if (cfi_use_chip_ready_for_write(map)) > + return chip_ready(map, chip, addr); If possible and not too invasive I would definitely add a "quirks" flag somewhere instead of this cfi_use_chip_ready_for_write() check. Anyway, I would move this to the chip_good() implementation directly so we partially hide the quirks complexity from the core. > + > + return chip_good(map, chip, addr, expected); > +} > + > static int get_chip(struct map_info *map, struct flchip *chip, unsigned long adr, int mode) > { > DECLARE_WAITQUEUE(wait, current); > @@ -1686,7 +1703,7 @@ static int __xipram do_write_oneword_once(struct map_info *map, > * "chip_good" to avoid the failure due to scheduling. > */ > if (time_after(jiffies, timeo) && > - !chip_good(map, chip, adr, datum)) { > + !chip_good_for_write(map, chip, adr, datum)) { > xip_enable(map, chip, adr); > printk(KERN_WARNING "MTD %s(): software timeout\n", __func__); > xip_disable(map, chip, adr); > @@ -1694,7 +1711,7 @@ static int __xipram do_write_oneword_once(struct map_info *map, > break; > } > > - if (chip_good(map, chip, adr, datum)) { > + if (chip_good_for_write(map, chip, adr, datum)) { > if (cfi_check_err_status(map, chip, adr)) > ret = -EIO; > break; > @@ -1966,14 +1983,14 @@ static int __xipram do_write_buffer_wait(struct map_info *map, > * "chip_good" to avoid the failure due to scheduling. > */ > if (time_after(jiffies, timeo) && > - !chip_good(map, chip, adr, datum)) { > + !chip_good_for_write(map, chip, adr, datum)) { > pr_err("MTD %s(): software timeout, address:0x%.8lx.\n", > __func__, adr); > ret = -EIO; > break; > } > > - if (chip_good(map, chip, adr, datum)) { > + if (chip_good_for_write(map, chip, adr, datum)) { > if (cfi_check_err_status(map, chip, adr)) > ret = -EIO; > break; Thanks, Miquèl
On 16/03/22 10:51 pm, Miquel Raynal wrote: > Hi Tokunori, > > ikegami.t@gmail.com wrote on Thu, 17 Mar 2022 00:54:54 +0900: > >> As pointed out by this bug report [1], buffered writes are now broken on >> S29GL064N. This issue comes from a rework which switched from using chip_good() >> to chip_ready(), because DQ true data 0xFF is read on S29GL064N and an error >> returned by chip_good(). > > Vignesh, I believe you understand this issue better than I do, can you > propose an improved commit log? How about: Since commit dfeae1073583("mtd: cfi_cmdset_0002: Change write buffer to check correct value") buffered writes fail on S29GL064N. This is because, on S29GL064N, reads return 0xFF at the end of DQ polling for write completion, where as, chip_good() check expects actual data written to the last location to be returned post DQ polling completion. Fix is to revert to using chip_good() for S29GL064N which only checks for DQ lines to settle down to determine write completion. > >> One way to solve the issue is to revert the change >> partially to use chip_ready for S29GL064N. >> >> [1] https://lore.kernel.org/r/b687c259-6413-26c9-d4c9-b3afa69ea124@pengutronix.de/ >> >> Fixes: dfeae1073583("mtd: cfi_cmdset_0002: Change write buffer to check correct value") >> Signed-off-by: Tokunori Ikegami <ikegami.t@gmail.com> >> Tested-by: Ahmad Fatoum <a.fatoum@pengutronix.de> >> Cc: stable@vger.kernel.org >> --- >> drivers/mtd/chips/cfi_cmdset_0002.c | 25 +++++++++++++++++++++---- >> 1 file changed, 21 insertions(+), 4 deletions(-) >> >> diff --git a/drivers/mtd/chips/cfi_cmdset_0002.c b/drivers/mtd/chips/cfi_cmdset_0002.c >> index e68ddf0f7fc0..6c57f85e1b8e 100644 >> --- a/drivers/mtd/chips/cfi_cmdset_0002.c >> +++ b/drivers/mtd/chips/cfi_cmdset_0002.c >> @@ -866,6 +866,23 @@ static int __xipram chip_check(struct map_info *map, struct flchip *chip, >> chip_check(map, chip, addr, &datum); \ >> }) >> >> +static bool __xipram cfi_use_chip_ready_for_write(struct map_info *map) > > At the very least I would call this function: > cfi_use_chip_ready_for_writes() > > Yet, I still don't fully get what chip_ready is versus chip_good. > >> +{ >> + struct cfi_private *cfi = map->fldrv_priv; >> + >> + return cfi->mfr == CFI_MFR_AMD && cfi->id == 0x0c01; >> +} >> + >> +static int __xipram chip_good_for_write(struct map_info *map, >> + struct flchip *chip, unsigned long addr, >> + map_word expected) >> +{ >> + if (cfi_use_chip_ready_for_write(map)) >> + return chip_ready(map, chip, addr); > > If possible and not too invasive I would definitely add a "quirks" flag > somewhere instead of this cfi_use_chip_ready_for_write() check. > > Anyway, I would move this to the chip_good() implementation directly so > we partially hide the quirks complexity from the core. Yeah, unfortunately this driver does not use quirk flags and tends to hide quirks behind bool functions like above Regards Vignesh
Hello Vignesh, On 17.03.22 11:01, Vignesh Raghavendra wrote: > > > On 16/03/22 10:51 pm, Miquel Raynal wrote: >> Hi Tokunori, >> >> ikegami.t@gmail.com wrote on Thu, 17 Mar 2022 00:54:54 +0900: >> >>> As pointed out by this bug report [1], buffered writes are now broken on >>> S29GL064N. This issue comes from a rework which switched from using chip_good() >>> to chip_ready(), because DQ true data 0xFF is read on S29GL064N and an error >>> returned by chip_good(). >> >> Vignesh, I believe you understand this issue better than I do, can you >> propose an improved commit log? > > How about: > > Since commit dfeae1073583("mtd: cfi_cmdset_0002: Change write buffer to > check correct value") buffered writes fail on S29GL064N. This is > because, on S29GL064N, reads return 0xFF at the end of DQ polling for > write completion, where as, chip_good() check expects actual data > written to the last location to be returned post DQ polling completion. > Fix is to revert to using chip_good() for S29GL064N which only checks > for DQ lines to settle down to determine write completion. Message sounds good to me with one remark: The issue is independent of whether buffered writes are used or not. It's just because buffered writes are the default, that it was broken by dfeae1073583 ("mtd: cfi_cmdset_0002: Change write buffer to check correct value"). The word write case was broken by 37c673ade35c ("mtd: cfi_cmdset_0002: Use chip_good() to retry in do_write_oneword()"), so the commit message should probably reference both. as this commit indeed fixes both FORCE_WORD_WRITE == 0 and == 1. Thanks, Ahmad > >> >>> One way to solve the issue is to revert the change >>> partially to use chip_ready for S29GL064N. >>> >>> [1] https://lore.kernel.org/r/b687c259-6413-26c9-d4c9-b3afa69ea124@pengutronix.de/ >>> >>> Fixes: dfeae1073583("mtd: cfi_cmdset_0002: Change write buffer to check correct value") >>> Signed-off-by: Tokunori Ikegami <ikegami.t@gmail.com> >>> Tested-by: Ahmad Fatoum <a.fatoum@pengutronix.de> >>> Cc: stable@vger.kernel.org >>> --- >>> drivers/mtd/chips/cfi_cmdset_0002.c | 25 +++++++++++++++++++++---- >>> 1 file changed, 21 insertions(+), 4 deletions(-) >>> >>> diff --git a/drivers/mtd/chips/cfi_cmdset_0002.c b/drivers/mtd/chips/cfi_cmdset_0002.c >>> index e68ddf0f7fc0..6c57f85e1b8e 100644 >>> --- a/drivers/mtd/chips/cfi_cmdset_0002.c >>> +++ b/drivers/mtd/chips/cfi_cmdset_0002.c >>> @@ -866,6 +866,23 @@ static int __xipram chip_check(struct map_info *map, struct flchip *chip, >>> chip_check(map, chip, addr, &datum); \ >>> }) >>> >>> +static bool __xipram cfi_use_chip_ready_for_write(struct map_info *map) >> >> At the very least I would call this function: >> cfi_use_chip_ready_for_writes() >> >> Yet, I still don't fully get what chip_ready is versus chip_good. >> >>> +{ >>> + struct cfi_private *cfi = map->fldrv_priv; >>> + >>> + return cfi->mfr == CFI_MFR_AMD && cfi->id == 0x0c01; >>> +} >>> + >>> +static int __xipram chip_good_for_write(struct map_info *map, >>> + struct flchip *chip, unsigned long addr, >>> + map_word expected) >>> +{ >>> + if (cfi_use_chip_ready_for_write(map)) >>> + return chip_ready(map, chip, addr); >> >> If possible and not too invasive I would definitely add a "quirks" flag >> somewhere instead of this cfi_use_chip_ready_for_write() check. >> >> Anyway, I would move this to the chip_good() implementation directly so >> we partially hide the quirks complexity from the core. > > Yeah, unfortunately this driver does not use quirk flags and tends to > hide quirks behind bool functions like above > > Regards > Vignesh >
On 16.03.22 16:54, Tokunori Ikegami wrote: > As pointed out by this bug report [1], buffered writes are now broken on > S29GL064N. This issue comes from a rework which switched from using chip_good() > to chip_ready(), because DQ true data 0xFF is read on S29GL064N and an error > returned by chip_good(). One way to solve the issue is to revert the change > partially to use chip_ready for S29GL064N. > > [1] https://lore.kernel.org/r/b687c259-6413-26c9-d4c9-b3afa69ea124@pengutronix.de/ Why did you switch from the documented format for links you added on my request (see https://lore.kernel.org/stable/f1b44e87-e457-7783-d46e-0d577cea3b72@leemhuis.info/ ) to v2 to something else that is not recognized by tools and scripts that rely on proper link tags? You are making my and maybe other peoples life unnecessary hard. :-(( FWIW, the proper style should support footnote style like this: Link: https://lore.kernel.org/r/b687c259-6413-26c9-d4c9-b3afa69ea124@pengutronix.de/ [1] Ciao, Thorsten #regzbot ^backmonitor: https://lore.kernel.org/r/b687c259-6413-26c9-d4c9-b3afa69ea124@pengutronix.de/
Hi Thorsten, regressions@leemhuis.info wrote on Mon, 21 Mar 2022 12:48:11 +0100: > On 16.03.22 16:54, Tokunori Ikegami wrote: > > As pointed out by this bug report [1], buffered writes are now broken on > > S29GL064N. This issue comes from a rework which switched from using chip_good() > > to chip_ready(), because DQ true data 0xFF is read on S29GL064N and an error > > returned by chip_good(). One way to solve the issue is to revert the change > > partially to use chip_ready for S29GL064N. > > > > [1] https://lore.kernel.org/r/b687c259-6413-26c9-d4c9-b3afa69ea124@pengutronix.de/ > > Why did you switch from the documented format for links you added on my > request (see > https://lore.kernel.org/stable/f1b44e87-e457-7783-d46e-0d577cea3b72@leemhuis.info/ > > ) to v2 to something else that is not recognized by tools and scripts > that rely on proper link tags? You are making my and maybe other peoples > life unnecessary hard. :-(( > > FWIW, the proper style should support footnote style like this: > > Link: > https://lore.kernel.org/r/b687c259-6413-26c9-d4c9-b3afa69ea124@pengutronix.de/ > [1] > > Ciao, Thorsten > > #regzbot ^backmonitor: > https://lore.kernel.org/r/b687c259-6413-26c9-d4c9-b3afa69ea124@pengutronix.de/ > Because today's requirement from maintainers is to provide a Link tag that points to the mail discussion of the patch being applied. I then asked to use the above form instead to point to the bug report because I don't see the point of having a "Link" tag for it? Thanks, Miquèl
On 21.03.22 13:35, Miquel Raynal wrote: > > regressions@leemhuis.info wrote on Mon, 21 Mar 2022 12:48:11 +0100: > >> On 16.03.22 16:54, Tokunori Ikegami wrote: >>> As pointed out by this bug report [1], buffered writes are now broken on >>> S29GL064N. This issue comes from a rework which switched from using chip_good() >>> to chip_ready(), because DQ true data 0xFF is read on S29GL064N and an error >>> returned by chip_good(). One way to solve the issue is to revert the change >>> partially to use chip_ready for S29GL064N. >>> >>> [1] https://lore.kernel.org/r/b687c259-6413-26c9-d4c9-b3afa69ea124@pengutronix.de/ >> >> Why did you switch from the documented format for links you added on my >> request (see >> https://lore.kernel.org/stable/f1b44e87-e457-7783-d46e-0d577cea3b72@leemhuis.info/ >> >> ) to v2 to something else that is not recognized by tools and scripts >> that rely on proper link tags? You are making my and maybe other peoples >> life unnecessary hard. :-(( >> >> FWIW, the proper style should support footnote style like this: >> >> Link: >> https://lore.kernel.org/r/b687c259-6413-26c9-d4c9-b3afa69ea124@pengutronix.de/ >> [1] >> >> Ciao, Thorsten >> >> #regzbot ^backmonitor: >> https://lore.kernel.org/r/b687c259-6413-26c9-d4c9-b3afa69ea124@pengutronix.de/ >> > > Because today's requirement from maintainers is to provide a Link > tag that points to the mail discussion of the patch being applied. That can be an additional Link tag, that is done all the time. > I > then asked to use the above form instead to point to the bug report > because I don't see the point of having a "Link" tag for it? But it's not your own project, we are all working with thousands of people together on this project on various different fronts. That needs coordination, as some things otherwise become hard or impossible. That's why we have documentation that explains how to do some things. Not following it just because you don't like it is not helpful and in this case makes my life as a volunteer a lot harder. If you don't like the approach explained by the documentation, submit a patch adjusting the documentation and then we can talk about this. But until that is applied please stick to the format explained by the documentation. Ciao, Thorsten
Thorsten, regressions@leemhuis.info wrote on Mon, 21 Mar 2022 13:51:10 +0100: > On 21.03.22 13:35, Miquel Raynal wrote: > > > > regressions@leemhuis.info wrote on Mon, 21 Mar 2022 12:48:11 +0100: > > > >> On 16.03.22 16:54, Tokunori Ikegami wrote: > >>> As pointed out by this bug report [1], buffered writes are now broken on > >>> S29GL064N. This issue comes from a rework which switched from using chip_good() > >>> to chip_ready(), because DQ true data 0xFF is read on S29GL064N and an error > >>> returned by chip_good(). One way to solve the issue is to revert the change > >>> partially to use chip_ready for S29GL064N. > >>> > >>> [1] https://lore.kernel.org/r/b687c259-6413-26c9-d4c9-b3afa69ea124@pengutronix.de/ > >> > >> Why did you switch from the documented format for links you added on my > >> request (see > >> https://lore.kernel.org/stable/f1b44e87-e457-7783-d46e-0d577cea3b72@leemhuis.info/ > >> > >> ) to v2 to something else that is not recognized by tools and scripts > >> that rely on proper link tags? You are making my and maybe other peoples > >> life unnecessary hard. :-(( > >> > >> FWIW, the proper style should support footnote style like this: > >> > >> Link: > >> https://lore.kernel.org/r/b687c259-6413-26c9-d4c9-b3afa69ea124@pengutronix.de/ > >> [1] > >> > >> Ciao, Thorsten > >> > >> #regzbot ^backmonitor: > >> https://lore.kernel.org/r/b687c259-6413-26c9-d4c9-b3afa69ea124@pengutronix.de/ > >> > > > > Because today's requirement from maintainers is to provide a Link > > tag that points to the mail discussion of the patch being applied. > > That can be an additional Link tag, that is done all the time. > > > I > > then asked to use the above form instead to point to the bug report > > because I don't see the point of having a "Link" tag for it? Perhaps I should emphasize that I don't remember your initial request regarding the use of a Link tag and my original idea was to help this contributor, not kill your tools which I actually know very little about. > But it's not your own project, we are all working with thousands of > people together on this project on various different fronts. That needs > coordination, as some things otherwise become hard or impossible. That's > why we have documentation that explains how to do some things. Not > following it just because you don't like it is not helpful and in this > case makes my life as a volunteer a lot harder. Let's be honest, you are referring to a Documentation patch that *you* wrote and was merged into Linus' tree mid January. How often do you think people used to the contribution workflow monitor these files? I am totally fine enforcing the use of Link: tags if this is what has been decided, just don't expect everybody to switch to a style rather than another over a night. > If you don't like the approach explained by the documentation, submit a > patch adjusting the documentation and then we can talk about this. But > until that is applied please stick to the format explained by the > documentation. This is uselessly condescending. Thanks, Miquèl
On 21.03.22 14:41, Miquel Raynal wrote: > regressions@leemhuis.info wrote on Mon, 21 Mar 2022 13:51:10 +0100: >> On 21.03.22 13:35, Miquel Raynal wrote: >>> regressions@leemhuis.info wrote on Mon, 21 Mar 2022 12:48:11 +0100: >>> >>>> On 16.03.22 16:54, Tokunori Ikegami wrote: >>>>> As pointed out by this bug report [1], buffered writes are now broken on >>>>> S29GL064N. This issue comes from a rework which switched from using chip_good() >>>>> to chip_ready(), because DQ true data 0xFF is read on S29GL064N and an error >>>>> returned by chip_good(). One way to solve the issue is to revert the change >>>>> partially to use chip_ready for S29GL064N. >>>>> >>>>> [1] https://lore.kernel.org/r/b687c259-6413-26c9-d4c9-b3afa69ea124@pengutronix.de/ >>>> >>>> Why did you switch from the documented format for links you added on my >>>> request (see >>>> https://lore.kernel.org/stable/f1b44e87-e457-7783-d46e-0d577cea3b72@leemhuis.info/ >>>> >>>> ) to v2 to something else that is not recognized by tools and scripts >>>> that rely on proper link tags? You are making my and maybe other peoples >>>> life unnecessary hard. :-(( >>>> >>>> FWIW, the proper style should support footnote style like this: >>>> >>>> Link: >>>> https://lore.kernel.org/r/b687c259-6413-26c9-d4c9-b3afa69ea124@pengutronix.de/ >>>> [1] >>>> >>>> Ciao, Thorsten >>>> >>>> #regzbot ^backmonitor: >>>> https://lore.kernel.org/r/b687c259-6413-26c9-d4c9-b3afa69ea124@pengutronix.de/ >>>> >>> >>> Because today's requirement from maintainers is to provide a Link >>> tag that points to the mail discussion of the patch being applied. >> >> That can be an additional Link tag, that is done all the time. >> >>> I >>> then asked to use the above form instead to point to the bug report >>> because I don't see the point of having a "Link" tag for it? > > Perhaps I should emphasize that I don't remember your initial request > regarding the use of a Link tag Happen, no worries. > and my original idea was to help this > contributor, not kill your tools which I actually know very little > about. >>> But it's not your own project, we are all working with thousands of >> people together on this project on various different fronts. That needs >> coordination, as some things otherwise become hard or impossible. That's >> why we have documentation that explains how to do some things. Not >> following it just because you don't like it is not helpful and in this >> case makes my life as a volunteer a lot harder. > > Let's be honest, you are referring to a Documentation patch that *you* > wrote Correct, but in case of submitting-patches it was just a clarification how to place links; why the whole aspect was missing in the other is kinda odd and likely lost in history... > and was merged into Linus' tree mid January. How often do you > think people used to the contribution workflow monitor these files? Not often, that's why I have no problem pointing it out, even if that's slightly annoying. But you can imagine that it felt kinda odd on my side when asking someone to set the links (with references to the docs explaining how to set them) and seeing them added then in v2, just so see they vanished again in v3 of the same patch. :-/ > I am totally fine enforcing the use of Link: tags if this is what has > been decided, just don't expect everybody to switch to a style rather > than another over a night. I don't. >> If you don't like the approach explained by the documentation, submit a >> patch adjusting the documentation and then we can talk about this. But >> until that is applied please stick to the format explained by the >> documentation. > This is uselessly condescending. I apologize, it wasn't meant that way. I had to many discussions already where people were not setting any links and it seems the topic is slowly hitting a nerve here. Sorry. Ciao, Thorsten
Hi Thorsten, regressions@leemhuis.info wrote on Mon, 21 Mar 2022 15:17:50 +0100: > On 21.03.22 14:41, Miquel Raynal wrote: > > regressions@leemhuis.info wrote on Mon, 21 Mar 2022 13:51:10 +0100: > >> On 21.03.22 13:35, Miquel Raynal wrote: > >>> regressions@leemhuis.info wrote on Mon, 21 Mar 2022 12:48:11 +0100: > >>> > >>>> On 16.03.22 16:54, Tokunori Ikegami wrote: > >>>>> As pointed out by this bug report [1], buffered writes are now broken on > >>>>> S29GL064N. This issue comes from a rework which switched from using chip_good() > >>>>> to chip_ready(), because DQ true data 0xFF is read on S29GL064N and an error > >>>>> returned by chip_good(). One way to solve the issue is to revert the change > >>>>> partially to use chip_ready for S29GL064N. > >>>>> > >>>>> [1] https://lore.kernel.org/r/b687c259-6413-26c9-d4c9-b3afa69ea124@pengutronix.de/ > >>>> > >>>> Why did you switch from the documented format for links you added on my > >>>> request (see > >>>> https://lore.kernel.org/stable/f1b44e87-e457-7783-d46e-0d577cea3b72@leemhuis.info/ > >>>> > >>>> ) to v2 to something else that is not recognized by tools and scripts > >>>> that rely on proper link tags? You are making my and maybe other peoples > >>>> life unnecessary hard. :-(( > >>>> > >>>> FWIW, the proper style should support footnote style like this: > >>>> > >>>> Link: > >>>> https://lore.kernel.org/r/b687c259-6413-26c9-d4c9-b3afa69ea124@pengutronix.de/ > >>>> [1] > >>>> > >>>> Ciao, Thorsten > >>>> > >>>> #regzbot ^backmonitor: > >>>> https://lore.kernel.org/r/b687c259-6413-26c9-d4c9-b3afa69ea124@pengutronix.de/ > >>>> > >>> > >>> Because today's requirement from maintainers is to provide a Link > >>> tag that points to the mail discussion of the patch being applied. > >> > >> That can be an additional Link tag, that is done all the time. > >> > >>> I > >>> then asked to use the above form instead to point to the bug report > >>> because I don't see the point of having a "Link" tag for it? > > > > Perhaps I should emphasize that I don't remember your initial request > > regarding the use of a Link tag > > Happen, no worries. > > > and my original idea was to help this > > contributor, not kill your tools which I actually know very little > > about. > >>> But it's not your own project, we are all working with thousands of > >> people together on this project on various different fronts. That needs > >> coordination, as some things otherwise become hard or impossible. That's > >> why we have documentation that explains how to do some things. Not > >> following it just because you don't like it is not helpful and in this > >> case makes my life as a volunteer a lot harder. > > > > Let's be honest, you are referring to a Documentation patch that *you* > > wrote > > Correct, but in case of submitting-patches it was just a clarification > how to place links; why the whole aspect was missing in the other is > kinda odd and likely lost in history... > > > and was merged into Linus' tree mid January. How often do you > > think people used to the contribution workflow monitor these files? > > Not often, that's why I have no problem pointing it out, even if that's > slightly annoying. But you can imagine that it felt kinda odd on my side > when asking someone to set the links (with references to the docs > explaining how to set them) and seeing them added then in v2, just so > see they vanished again in v3 of the same patch. :-/ I fully understand. I actually learned that these tags had to be used for this purpose, so I will actually enforce their use in my next reviews. Just a side question, should the Documentation also mention how to refer to links for people not used to it? Something like [5.Posting.rst]: "Link: <link> [1] Link: <link> [2]" My original point was that maintainers would almost always add a Link tag at the end, containing the mailing-list thread about the patch being applied. Just saying in the commit log "see the link below" then becomes misleading. > > I am totally fine enforcing the use of Link: tags if this is what has > > been decided, just don't expect everybody to switch to a style rather > > than another over a night. > > I don't. > > >> If you don't like the approach explained by the documentation, submit a > >> patch adjusting the documentation and then we can talk about this. But > >> until that is applied please stick to the format explained by the > >> documentation. > > This is uselessly condescending. > > I apologize, it wasn't meant that way. No worries, thanks :-) > I had to many discussions already > where people were not setting any links and it seems the topic is slowly > hitting a nerve here. Sorry. I also feel like I am repeating myself sometimes. And then I remember Rob and the ton of e-mails where he has to repeat himself hundreds of times a day and I feel slightly better :-p Cheers, Miquèl
On 21.03.22 15:56, Miquel Raynal wrote: > regressions@leemhuis.info wrote on Mon, 21 Mar 2022 15:17:50 +0100: >> On 21.03.22 14:41, Miquel Raynal wrote: >>> regressions@leemhuis.info wrote on Mon, 21 Mar 2022 13:51:10 +0100: >>>> On 21.03.22 13:35, Miquel Raynal wrote: >>>>> regressions@leemhuis.info wrote on Mon, 21 Mar 2022 12:48:11 +0100: >>>>> >>>>>> On 16.03.22 16:54, Tokunori Ikegami wrote: >>>>>>> As pointed out by this bug report [1], buffered writes are now broken on >>>>>>> S29GL064N. This issue comes from a rework which switched from using chip_good() >>>>>>> to chip_ready(), because DQ true data 0xFF is read on S29GL064N and an error >>>>>>> returned by chip_good(). One way to solve the issue is to revert the change >>>>>>> partially to use chip_ready for S29GL064N. >>>>>>> >>>>>>> [1] https://lore.kernel.org/r/b687c259-6413-26c9-d4c9-b3afa69ea124@pengutronix.de/ >>>>>> >>>>>> Why did you switch from the documented format for links you added on my >>>>>> request (see >>>>>> https://lore.kernel.org/stable/f1b44e87-e457-7783-d46e-0d577cea3b72@leemhuis.info/ >>>>>> >>>>>> ) to v2 to something else that is not recognized by tools and scripts >>>>>> that rely on proper link tags? You are making my and maybe other peoples >>>>>> life unnecessary hard. :-(( >>>>>> >>>>>> FWIW, the proper style should support footnote style like this: >>>>>> >>>>>> Link: >>>>>> https://lore.kernel.org/r/b687c259-6413-26c9-d4c9-b3afa69ea124@pengutronix.de/ >>>>>> [1] >>>>>> >>>>>> Ciao, Thorsten >>>>>> >>>>>> #regzbot ^backmonitor: >>>>>> https://lore.kernel.org/r/b687c259-6413-26c9-d4c9-b3afa69ea124@pengutronix.de/ >>>>>> >>>>> >>>>> Because today's requirement from maintainers is to provide a Link >>>>> tag that points to the mail discussion of the patch being applied. >>>> >>>> That can be an additional Link tag, that is done all the time. >>>> >>>>> I >>>>> then asked to use the above form instead to point to the bug report >>>>> because I don't see the point of having a "Link" tag for it? >>> >>> Perhaps I should emphasize that I don't remember your initial request >>> regarding the use of a Link tag >> >> Happen, no worries. >> >>> and my original idea was to help this >>> contributor, not kill your tools which I actually know very little >>> about. >>>>> But it's not your own project, we are all working with thousands of >>>> people together on this project on various different fronts. That needs >>>> coordination, as some things otherwise become hard or impossible. That's >>>> why we have documentation that explains how to do some things. Not >>>> following it just because you don't like it is not helpful and in this >>>> case makes my life as a volunteer a lot harder. >>> >>> Let's be honest, you are referring to a Documentation patch that *you* >>> wrote >> >> Correct, but in case of submitting-patches it was just a clarification >> how to place links; why the whole aspect was missing in the other is >> kinda odd and likely lost in history... >> >>> and was merged into Linus' tree mid January. How often do you >>> think people used to the contribution workflow monitor these files? >> >> Not often, that's why I have no problem pointing it out, even if that's >> slightly annoying. But you can imagine that it felt kinda odd on my side >> when asking someone to set the links (with references to the docs >> explaining how to set them) and seeing them added then in v2, just so >> see they vanished again in v3 of the same patch. :-/ > > I fully understand. I actually learned that these tags had to be used > for this purpose, so I will actually enforce their use in my next > reviews. > > Just a side question, should the Documentation also mention how > to refer to links for people not used to it? Something like > [5.Posting.rst]: > > "Link: <link> [1] > Link: <link> [2]" Maybe. But I think the better approach would be: introduce more specify tags like "Reported:" (and maybe drop "Reported-by" at the same time?) or "BugLink" (some people use that already!) would be better -- and then maybe "Posted:", "Reviewposting", or something like that for the link to the patch that is being applied; and leave "Link" for the rest. I proposed that a while ago, but that didn't get any traction. > My original point was that maintainers would almost always add > a Link tag at the end, containing the mailing-list thread about the > patch being applied. Just saying in the commit log "see the link below" > then becomes misleading. Maybe, but OTOH that link is normally at the end, which kinda makes it obvious. > [...] Ciao, Thorsten
Hi Miquèl-san, On 2022/03/17 2:21, Miquel Raynal wrote: > Hi Tokunori, > > ikegami.t@gmail.com wrote on Thu, 17 Mar 2022 00:54:54 +0900: > >> As pointed out by this bug report [1], buffered writes are now broken on >> S29GL064N. This issue comes from a rework which switched from using chip_good() >> to chip_ready(), because DQ true data 0xFF is read on S29GL064N and an error >> returned by chip_good(). > Vignesh, I believe you understand this issue better than I do, can you > propose an improved commit log? > >> One way to solve the issue is to revert the change >> partially to use chip_ready for S29GL064N. >> >> [1] https://lore.kernel.org/r/b687c259-6413-26c9-d4c9-b3afa69ea124@pengutronix.de/ >> >> Fixes: dfeae1073583("mtd: cfi_cmdset_0002: Change write buffer to check correct value") >> Signed-off-by: Tokunori Ikegami <ikegami.t@gmail.com> >> Tested-by: Ahmad Fatoum <a.fatoum@pengutronix.de> >> Cc: stable@vger.kernel.org >> --- >> drivers/mtd/chips/cfi_cmdset_0002.c | 25 +++++++++++++++++++++---- >> 1 file changed, 21 insertions(+), 4 deletions(-) >> >> diff --git a/drivers/mtd/chips/cfi_cmdset_0002.c b/drivers/mtd/chips/cfi_cmdset_0002.c >> index e68ddf0f7fc0..6c57f85e1b8e 100644 >> --- a/drivers/mtd/chips/cfi_cmdset_0002.c >> +++ b/drivers/mtd/chips/cfi_cmdset_0002.c >> @@ -866,6 +866,23 @@ static int __xipram chip_check(struct map_info *map, struct flchip *chip, >> chip_check(map, chip, addr, &datum); \ >> }) >> >> +static bool __xipram cfi_use_chip_ready_for_write(struct map_info *map) > At the very least I would call this function: > cfi_use_chip_ready_for_writes() > > Yet, I still don't fully get what chip_ready is versus chip_good. This was deleted as to use the quirks flag instead. > >> +{ >> + struct cfi_private *cfi = map->fldrv_priv; >> + >> + return cfi->mfr == CFI_MFR_AMD && cfi->id == 0x0c01; >> +} >> + >> +static int __xipram chip_good_for_write(struct map_info *map, >> + struct flchip *chip, unsigned long addr, >> + map_word expected) >> +{ >> + if (cfi_use_chip_ready_for_write(map)) >> + return chip_ready(map, chip, addr); > If possible and not too invasive I would definitely add a "quirks" flag > somewhere instead of this cfi_use_chip_ready_for_write() check. Added the quirks flag by the version 5 patch. > > Anyway, I would move this to the chip_good() implementation directly so > we partially hide the quirks complexity from the core. Yes also added the chip_good to check the quirks flag. Regards, Ikegami > >> + >> + return chip_good(map, chip, addr, expected); >> +} >> + >> static int get_chip(struct map_info *map, struct flchip *chip, unsigned long adr, int mode) >> { >> DECLARE_WAITQUEUE(wait, current); >> @@ -1686,7 +1703,7 @@ static int __xipram do_write_oneword_once(struct map_info *map, >> * "chip_good" to avoid the failure due to scheduling. >> */ >> if (time_after(jiffies, timeo) && >> - !chip_good(map, chip, adr, datum)) { >> + !chip_good_for_write(map, chip, adr, datum)) { >> xip_enable(map, chip, adr); >> printk(KERN_WARNING "MTD %s(): software timeout\n", __func__); >> xip_disable(map, chip, adr); >> @@ -1694,7 +1711,7 @@ static int __xipram do_write_oneword_once(struct map_info *map, >> break; >> } >> >> - if (chip_good(map, chip, adr, datum)) { >> + if (chip_good_for_write(map, chip, adr, datum)) { >> if (cfi_check_err_status(map, chip, adr)) >> ret = -EIO; >> break; >> @@ -1966,14 +1983,14 @@ static int __xipram do_write_buffer_wait(struct map_info *map, >> * "chip_good" to avoid the failure due to scheduling. >> */ >> if (time_after(jiffies, timeo) && >> - !chip_good(map, chip, adr, datum)) { >> + !chip_good_for_write(map, chip, adr, datum)) { >> pr_err("MTD %s(): software timeout, address:0x%.8lx.\n", >> __func__, adr); >> ret = -EIO; >> break; >> } >> >> - if (chip_good(map, chip, adr, datum)) { >> + if (chip_good_for_write(map, chip, adr, datum)) { >> if (cfi_check_err_status(map, chip, adr)) >> ret = -EIO; >> break; > > Thanks, > Miquèl
Hi, On 2022/03/17 19:01, Vignesh Raghavendra wrote: > > On 16/03/22 10:51 pm, Miquel Raynal wrote: >> Hi Tokunori, >> >> ikegami.t@gmail.com wrote on Thu, 17 Mar 2022 00:54:54 +0900: >> >>> As pointed out by this bug report [1], buffered writes are now broken on >>> S29GL064N. This issue comes from a rework which switched from using chip_good() >>> to chip_ready(), because DQ true data 0xFF is read on S29GL064N and an error >>> returned by chip_good(). >> Vignesh, I believe you understand this issue better than I do, can you >> propose an improved commit log? > How about: > > Since commit dfeae1073583("mtd: cfi_cmdset_0002: Change write buffer to > check correct value") buffered writes fail on S29GL064N. This is > because, on S29GL064N, reads return 0xFF at the end of DQ polling for > write completion, where as, chip_good() check expects actual data > written to the last location to be returned post DQ polling completion. > Fix is to revert to using chip_good() for S29GL064N which only checks > for DQ lines to settle down to determine write completion. Fixed the commit message as suggested by the version 5 patch. >>> One way to solve the issue is to revert the change >>> partially to use chip_ready for S29GL064N. >>> >>> [1] https://lore.kernel.org/r/b687c259-6413-26c9-d4c9-b3afa69ea124@pengutronix.de/ >>> >>> Fixes: dfeae1073583("mtd: cfi_cmdset_0002: Change write buffer to check correct value") >>> Signed-off-by: Tokunori Ikegami <ikegami.t@gmail.com> >>> Tested-by: Ahmad Fatoum <a.fatoum@pengutronix.de> >>> Cc: stable@vger.kernel.org >>> --- >>> drivers/mtd/chips/cfi_cmdset_0002.c | 25 +++++++++++++++++++++---- >>> 1 file changed, 21 insertions(+), 4 deletions(-) >>> >>> diff --git a/drivers/mtd/chips/cfi_cmdset_0002.c b/drivers/mtd/chips/cfi_cmdset_0002.c >>> index e68ddf0f7fc0..6c57f85e1b8e 100644 >>> --- a/drivers/mtd/chips/cfi_cmdset_0002.c >>> +++ b/drivers/mtd/chips/cfi_cmdset_0002.c >>> @@ -866,6 +866,23 @@ static int __xipram chip_check(struct map_info *map, struct flchip *chip, >>> chip_check(map, chip, addr, &datum); \ >>> }) >>> >>> +static bool __xipram cfi_use_chip_ready_for_write(struct map_info *map) >> At the very least I would call this function: >> cfi_use_chip_ready_for_writes() >> >> Yet, I still don't fully get what chip_ready is versus chip_good. >> >>> +{ >>> + struct cfi_private *cfi = map->fldrv_priv; >>> + >>> + return cfi->mfr == CFI_MFR_AMD && cfi->id == 0x0c01; >>> +} >>> + >>> +static int __xipram chip_good_for_write(struct map_info *map, >>> + struct flchip *chip, unsigned long addr, >>> + map_word expected) >>> +{ >>> + if (cfi_use_chip_ready_for_write(map)) >>> + return chip_ready(map, chip, addr); >> If possible and not too invasive I would definitely add a "quirks" flag >> somewhere instead of this cfi_use_chip_ready_for_write() check. >> >> Anyway, I would move this to the chip_good() implementation directly so >> we partially hide the quirks complexity from the core. > Yeah, unfortunately this driver does not use quirk flags and tends to > hide quirks behind bool functions like above Added the quirks flag as mentioned. Regards, Ikegami > > Regards > Vignesh
Hi Ahmad-san, On 2022/03/17 23:16, Ahmad Fatoum wrote: > Hello Vignesh, > > On 17.03.22 11:01, Vignesh Raghavendra wrote: >> >> On 16/03/22 10:51 pm, Miquel Raynal wrote: >>> Hi Tokunori, >>> >>> ikegami.t@gmail.com wrote on Thu, 17 Mar 2022 00:54:54 +0900: >>> >>>> As pointed out by this bug report [1], buffered writes are now broken on >>>> S29GL064N. This issue comes from a rework which switched from using chip_good() >>>> to chip_ready(), because DQ true data 0xFF is read on S29GL064N and an error >>>> returned by chip_good(). >>> Vignesh, I believe you understand this issue better than I do, can you >>> propose an improved commit log? >> How about: >> >> Since commit dfeae1073583("mtd: cfi_cmdset_0002: Change write buffer to >> check correct value") buffered writes fail on S29GL064N. This is >> because, on S29GL064N, reads return 0xFF at the end of DQ polling for >> write completion, where as, chip_good() check expects actual data >> written to the last location to be returned post DQ polling completion. >> Fix is to revert to using chip_good() for S29GL064N which only checks >> for DQ lines to settle down to determine write completion. > Message sounds good to me with one remark: The issue is independent of > whether buffered writes are used or not. It's just because buffered writes > are the default, that it was broken by dfeae1073583 ("mtd: cfi_cmdset_0002: > Change write buffer to check correct value"). The word write case was broken > by 37c673ade35c ("mtd: cfi_cmdset_0002: Use chip_good() to retry in > do_write_oneword()"), so the commit message should probably reference > both. as this commit indeed fixes both FORCE_WORD_WRITE == 0 and == 1. Is this really caused the error on do_write_oneword by the changed? Actually it was changed to use chip_good instead of chip_ready. But before the change still do_write_oneword uses both chip_ready and chip_good. So it seems that it is possible to be caused the error before the change also. By the way could you please try to test the version 5 patches again? Regards, Ikegami > > Thanks, > Ahmad > > >>>> One way to solve the issue is to revert the change >>>> partially to use chip_ready for S29GL064N. >>>> >>>> [1] https://lore.kernel.org/r/b687c259-6413-26c9-d4c9-b3afa69ea124@pengutronix.de/ >>>> >>>> Fixes: dfeae1073583("mtd: cfi_cmdset_0002: Change write buffer to check correct value") >>>> Signed-off-by: Tokunori Ikegami <ikegami.t@gmail.com> >>>> Tested-by: Ahmad Fatoum <a.fatoum@pengutronix.de> >>>> Cc: stable@vger.kernel.org >>>> --- >>>> drivers/mtd/chips/cfi_cmdset_0002.c | 25 +++++++++++++++++++++---- >>>> 1 file changed, 21 insertions(+), 4 deletions(-) >>>> >>>> diff --git a/drivers/mtd/chips/cfi_cmdset_0002.c b/drivers/mtd/chips/cfi_cmdset_0002.c >>>> index e68ddf0f7fc0..6c57f85e1b8e 100644 >>>> --- a/drivers/mtd/chips/cfi_cmdset_0002.c >>>> +++ b/drivers/mtd/chips/cfi_cmdset_0002.c >>>> @@ -866,6 +866,23 @@ static int __xipram chip_check(struct map_info *map, struct flchip *chip, >>>> chip_check(map, chip, addr, &datum); \ >>>> }) >>>> >>>> +static bool __xipram cfi_use_chip_ready_for_write(struct map_info *map) >>> At the very least I would call this function: >>> cfi_use_chip_ready_for_writes() >>> >>> Yet, I still don't fully get what chip_ready is versus chip_good. >>> >>>> +{ >>>> + struct cfi_private *cfi = map->fldrv_priv; >>>> + >>>> + return cfi->mfr == CFI_MFR_AMD && cfi->id == 0x0c01; >>>> +} >>>> + >>>> +static int __xipram chip_good_for_write(struct map_info *map, >>>> + struct flchip *chip, unsigned long addr, >>>> + map_word expected) >>>> +{ >>>> + if (cfi_use_chip_ready_for_write(map)) >>>> + return chip_ready(map, chip, addr); >>> If possible and not too invasive I would definitely add a "quirks" flag >>> somewhere instead of this cfi_use_chip_ready_for_write() check. >>> >>> Anyway, I would move this to the chip_good() implementation directly so >>> we partially hide the quirks complexity from the core. >> Yeah, unfortunately this driver does not use quirk flags and tends to >> hide quirks behind bool functions like above >> >> Regards >> Vignesh >> >
Hi, On 2022/03/22 0:16, Thorsten Leemhuis wrote: > On 21.03.22 15:56, Miquel Raynal wrote: >> regressions@leemhuis.info wrote on Mon, 21 Mar 2022 15:17:50 +0100: >>> On 21.03.22 14:41, Miquel Raynal wrote: >>>> regressions@leemhuis.info wrote on Mon, 21 Mar 2022 13:51:10 +0100: >>>>> On 21.03.22 13:35, Miquel Raynal wrote: >>>>>> regressions@leemhuis.info wrote on Mon, 21 Mar 2022 12:48:11 +0100: >>>>>> >>>>>>> On 16.03.22 16:54, Tokunori Ikegami wrote: >>>>>>>> As pointed out by this bug report [1], buffered writes are now broken on >>>>>>>> S29GL064N. This issue comes from a rework which switched from using chip_good() >>>>>>>> to chip_ready(), because DQ true data 0xFF is read on S29GL064N and an error >>>>>>>> returned by chip_good(). One way to solve the issue is to revert the change >>>>>>>> partially to use chip_ready for S29GL064N. >>>>>>>> >>>>>>>> [1] https://lore.kernel.org/r/b687c259-6413-26c9-d4c9-b3afa69ea124@pengutronix.de/ >>>>>>> Why did you switch from the documented format for links you added on my >>>>>>> request (see >>>>>>> https://lore.kernel.org/stable/f1b44e87-e457-7783-d46e-0d577cea3b72@leemhuis.info/ >>>>>>> >>>>>>> ) to v2 to something else that is not recognized by tools and scripts >>>>>>> that rely on proper link tags? You are making my and maybe other peoples >>>>>>> life unnecessary hard. :-(( >>>>>>> >>>>>>> FWIW, the proper style should support footnote style like this: >>>>>>> >>>>>>> Link: >>>>>>> https://lore.kernel.org/r/b687c259-6413-26c9-d4c9-b3afa69ea124@pengutronix.de/ >>>>>>> [1] >>>>>>> >>>>>>> Ciao, Thorsten >>>>>>> >>>>>>> #regzbot ^backmonitor: >>>>>>> https://lore.kernel.org/r/b687c259-6413-26c9-d4c9-b3afa69ea124@pengutronix.de/ >>>>>>> >>>>>> Because today's requirement from maintainers is to provide a Link >>>>>> tag that points to the mail discussion of the patch being applied. >>>>> That can be an additional Link tag, that is done all the time. >>>>> >>>>>> I >>>>>> then asked to use the above form instead to point to the bug report >>>>>> because I don't see the point of having a "Link" tag for it? >>>> Perhaps I should emphasize that I don't remember your initial request >>>> regarding the use of a Link tag >>> Happen, no worries. >>> >>>> and my original idea was to help this >>>> contributor, not kill your tools which I actually know very little >>>> about. >>>>>> But it's not your own project, we are all working with thousands of >>>>> people together on this project on various different fronts. That needs >>>>> coordination, as some things otherwise become hard or impossible. That's >>>>> why we have documentation that explains how to do some things. Not >>>>> following it just because you don't like it is not helpful and in this >>>>> case makes my life as a volunteer a lot harder. >>>> Let's be honest, you are referring to a Documentation patch that *you* >>>> wrote >>> Correct, but in case of submitting-patches it was just a clarification >>> how to place links; why the whole aspect was missing in the other is >>> kinda odd and likely lost in history... >>> >>>> and was merged into Linus' tree mid January. How often do you >>>> think people used to the contribution workflow monitor these files? >>> Not often, that's why I have no problem pointing it out, even if that's >>> slightly annoying. But you can imagine that it felt kinda odd on my side >>> when asking someone to set the links (with references to the docs >>> explaining how to set them) and seeing them added then in v2, just so >>> see they vanished again in v3 of the same patch. :-/ >> I fully understand. I actually learned that these tags had to be used >> for this purpose, so I will actually enforce their use in my next >> reviews. >> >> Just a side question, should the Documentation also mention how >> to refer to links for people not used to it? Something like >> [5.Posting.rst]: >> >> "Link: <link> [1] >> Link: <link> [2]" > Maybe. But I think the better approach would be: introduce more specify > tags like "Reported:" (and maybe drop "Reported-by" at the same time?) > or "BugLink" (some people use that already!) would be better -- and then > maybe "Posted:", "Reviewposting", or something like that for the link to > the patch that is being applied; and leave "Link" for the rest. I > proposed that a while ago, but that didn't get any traction. Fixed to use Link tag as before by the version 5 patches instead of [1]. Regards, Ikegami > >> My original point was that maintainers would almost always add >> a Link tag at the end, containing the mailing-list thread about the >> patch being applied. Just saying in the commit log "see the link below" >> then becomes misleading. > Maybe, but OTOH that link is normally at the end, which kinda makes it > obvious. > >> [...] > Ciao, Thorsten
On 22.03.22 03:49, Tokunori Ikegami wrote: > Hi Ahmad-san, > > On 2022/03/17 23:16, Ahmad Fatoum wrote: >> Hello Vignesh, >> >> On 17.03.22 11:01, Vignesh Raghavendra wrote: >>> >>> On 16/03/22 10:51 pm, Miquel Raynal wrote: >>>> Hi Tokunori, >>>> >>>> ikegami.t@gmail.com wrote on Thu, 17 Mar 2022 00:54:54 +0900: >>>> >>>>> As pointed out by this bug report [1], buffered writes are now broken on >>>>> S29GL064N. This issue comes from a rework which switched from using chip_good() >>>>> to chip_ready(), because DQ true data 0xFF is read on S29GL064N and an error >>>>> returned by chip_good(). >>>> Vignesh, I believe you understand this issue better than I do, can you >>>> propose an improved commit log? >>> How about: >>> >>> Since commit dfeae1073583("mtd: cfi_cmdset_0002: Change write buffer to >>> check correct value") buffered writes fail on S29GL064N. This is >>> because, on S29GL064N, reads return 0xFF at the end of DQ polling for >>> write completion, where as, chip_good() check expects actual data >>> written to the last location to be returned post DQ polling completion. >>> Fix is to revert to using chip_good() for S29GL064N which only checks >>> for DQ lines to settle down to determine write completion. >> Message sounds good to me with one remark: The issue is independent of >> whether buffered writes are used or not. It's just because buffered writes >> are the default, that it was broken by dfeae1073583 ("mtd: cfi_cmdset_0002: >> Change write buffer to check correct value"). The word write case was broken >> by 37c673ade35c ("mtd: cfi_cmdset_0002: Use chip_good() to retry in >> do_write_oneword()"), so the commit message should probably reference >> both. as this commit indeed fixes both FORCE_WORD_WRITE == 0 and == 1. > > Is this really caused the error on do_write_oneword by the changed? > Actually it was changed to use chip_good instead of chip_ready. > But before the change still do_write_oneword uses both chip_ready and chip_good. > So it seems that it is possible to be caused the error before the change also. Oh, I think you're right. Disregard my suggestion for the other Fixes: entry then. > By the way could you please try to test the version 5 patches again? Just did so for v7. Sorry for the delay. Cheers, Ahmad > > Regards, > Ikegami > >> >> Thanks, >> Ahmad >> >> >>>>> One way to solve the issue is to revert the change >>>>> partially to use chip_ready for S29GL064N. >>>>> >>>>> [1] https://lore.kernel.org/r/b687c259-6413-26c9-d4c9-b3afa69ea124@pengutronix.de/ >>>>> >>>>> Fixes: dfeae1073583("mtd: cfi_cmdset_0002: Change write buffer to check correct value") >>>>> Signed-off-by: Tokunori Ikegami <ikegami.t@gmail.com> >>>>> Tested-by: Ahmad Fatoum <a.fatoum@pengutronix.de> >>>>> Cc: stable@vger.kernel.org >>>>> --- >>>>> drivers/mtd/chips/cfi_cmdset_0002.c | 25 +++++++++++++++++++++---- >>>>> 1 file changed, 21 insertions(+), 4 deletions(-) >>>>> >>>>> diff --git a/drivers/mtd/chips/cfi_cmdset_0002.c b/drivers/mtd/chips/cfi_cmdset_0002.c >>>>> index e68ddf0f7fc0..6c57f85e1b8e 100644 >>>>> --- a/drivers/mtd/chips/cfi_cmdset_0002.c >>>>> +++ b/drivers/mtd/chips/cfi_cmdset_0002.c >>>>> @@ -866,6 +866,23 @@ static int __xipram chip_check(struct map_info *map, struct flchip *chip, >>>>> chip_check(map, chip, addr, &datum); \ >>>>> }) >>>>> +static bool __xipram cfi_use_chip_ready_for_write(struct map_info *map) >>>> At the very least I would call this function: >>>> cfi_use_chip_ready_for_writes() >>>> >>>> Yet, I still don't fully get what chip_ready is versus chip_good. >>>> >>>>> +{ >>>>> + struct cfi_private *cfi = map->fldrv_priv; >>>>> + >>>>> + return cfi->mfr == CFI_MFR_AMD && cfi->id == 0x0c01; >>>>> +} >>>>> + >>>>> +static int __xipram chip_good_for_write(struct map_info *map, >>>>> + struct flchip *chip, unsigned long addr, >>>>> + map_word expected) >>>>> +{ >>>>> + if (cfi_use_chip_ready_for_write(map)) >>>>> + return chip_ready(map, chip, addr); >>>> If possible and not too invasive I would definitely add a "quirks" flag >>>> somewhere instead of this cfi_use_chip_ready_for_write() check. >>>> >>>> Anyway, I would move this to the chip_good() implementation directly so >>>> we partially hide the quirks complexity from the core. >>> Yeah, unfortunately this driver does not use quirk flags and tends to >>> hide quirks behind bool functions like above >>> >>> Regards >>> Vignesh >>> >> >
Hi Ahmad-san, On 2022/03/28 19:49, Ahmad Fatoum wrote: > On 22.03.22 03:49, Tokunori Ikegami wrote: >> Hi Ahmad-san, >> >> On 2022/03/17 23:16, Ahmad Fatoum wrote: >>> Hello Vignesh, >>> >>> On 17.03.22 11:01, Vignesh Raghavendra wrote: >>>> On 16/03/22 10:51 pm, Miquel Raynal wrote: >>>>> Hi Tokunori, >>>>> >>>>> ikegami.t@gmail.com wrote on Thu, 17 Mar 2022 00:54:54 +0900: >>>>> >>>>>> As pointed out by this bug report [1], buffered writes are now broken on >>>>>> S29GL064N. This issue comes from a rework which switched from using chip_good() >>>>>> to chip_ready(), because DQ true data 0xFF is read on S29GL064N and an error >>>>>> returned by chip_good(). >>>>> Vignesh, I believe you understand this issue better than I do, can you >>>>> propose an improved commit log? >>>> How about: >>>> >>>> Since commit dfeae1073583("mtd: cfi_cmdset_0002: Change write buffer to >>>> check correct value") buffered writes fail on S29GL064N. This is >>>> because, on S29GL064N, reads return 0xFF at the end of DQ polling for >>>> write completion, where as, chip_good() check expects actual data >>>> written to the last location to be returned post DQ polling completion. >>>> Fix is to revert to using chip_good() for S29GL064N which only checks >>>> for DQ lines to settle down to determine write completion. >>> Message sounds good to me with one remark: The issue is independent of >>> whether buffered writes are used or not. It's just because buffered writes >>> are the default, that it was broken by dfeae1073583 ("mtd: cfi_cmdset_0002: >>> Change write buffer to check correct value"). The word write case was broken >>> by 37c673ade35c ("mtd: cfi_cmdset_0002: Use chip_good() to retry in >>> do_write_oneword()"), so the commit message should probably reference >>> both. as this commit indeed fixes both FORCE_WORD_WRITE == 0 and == 1. >> Is this really caused the error on do_write_oneword by the changed? >> Actually it was changed to use chip_good instead of chip_ready. >> But before the change still do_write_oneword uses both chip_ready and chip_good. >> So it seems that it is possible to be caused the error before the change also. > Oh, I think you're right. Disregard my suggestion for the other > Fixes: entry then. Thanks for the confirmation. > >> By the way could you please try to test the version 5 patches again? > Just did so for v7. Sorry for the delay. Thank you so much for your help. Sorry I missed to cc on version 5 to version 7 patches. Regards, Ikegami > > Cheers, > Ahmad > >> Regards, >> Ikegami >> >>> Thanks, >>> Ahmad >>> >>> >>>>>> One way to solve the issue is to revert the change >>>>>> partially to use chip_ready for S29GL064N. >>>>>> >>>>>> [1] https://lore.kernel.org/r/b687c259-6413-26c9-d4c9-b3afa69ea124@pengutronix.de/ >>>>>> >>>>>> Fixes: dfeae1073583("mtd: cfi_cmdset_0002: Change write buffer to check correct value") >>>>>> Signed-off-by: Tokunori Ikegami <ikegami.t@gmail.com> >>>>>> Tested-by: Ahmad Fatoum <a.fatoum@pengutronix.de> >>>>>> Cc: stable@vger.kernel.org >>>>>> --- >>>>>> drivers/mtd/chips/cfi_cmdset_0002.c | 25 +++++++++++++++++++++---- >>>>>> 1 file changed, 21 insertions(+), 4 deletions(-) >>>>>> >>>>>> diff --git a/drivers/mtd/chips/cfi_cmdset_0002.c b/drivers/mtd/chips/cfi_cmdset_0002.c >>>>>> index e68ddf0f7fc0..6c57f85e1b8e 100644 >>>>>> --- a/drivers/mtd/chips/cfi_cmdset_0002.c >>>>>> +++ b/drivers/mtd/chips/cfi_cmdset_0002.c >>>>>> @@ -866,6 +866,23 @@ static int __xipram chip_check(struct map_info *map, struct flchip *chip, >>>>>> chip_check(map, chip, addr, &datum); \ >>>>>> }) >>>>>> +static bool __xipram cfi_use_chip_ready_for_write(struct map_info *map) >>>>> At the very least I would call this function: >>>>> cfi_use_chip_ready_for_writes() >>>>> >>>>> Yet, I still don't fully get what chip_ready is versus chip_good. >>>>> >>>>>> +{ >>>>>> + struct cfi_private *cfi = map->fldrv_priv; >>>>>> + >>>>>> + return cfi->mfr == CFI_MFR_AMD && cfi->id == 0x0c01; >>>>>> +} >>>>>> + >>>>>> +static int __xipram chip_good_for_write(struct map_info *map, >>>>>> + struct flchip *chip, unsigned long addr, >>>>>> + map_word expected) >>>>>> +{ >>>>>> + if (cfi_use_chip_ready_for_write(map)) >>>>>> + return chip_ready(map, chip, addr); >>>>> If possible and not too invasive I would definitely add a "quirks" flag >>>>> somewhere instead of this cfi_use_chip_ready_for_write() check. >>>>> >>>>> Anyway, I would move this to the chip_good() implementation directly so >>>>> we partially hide the quirks complexity from the core. >>>> Yeah, unfortunately this driver does not use quirk flags and tends to >>>> hide quirks behind bool functions like above >>>> >>>> Regards >>>> Vignesh >>>> >
diff --git a/drivers/mtd/chips/cfi_cmdset_0002.c b/drivers/mtd/chips/cfi_cmdset_0002.c index e68ddf0f7fc0..6c57f85e1b8e 100644 --- a/drivers/mtd/chips/cfi_cmdset_0002.c +++ b/drivers/mtd/chips/cfi_cmdset_0002.c @@ -866,6 +866,23 @@ static int __xipram chip_check(struct map_info *map, struct flchip *chip, chip_check(map, chip, addr, &datum); \ }) +static bool __xipram cfi_use_chip_ready_for_write(struct map_info *map) +{ + struct cfi_private *cfi = map->fldrv_priv; + + return cfi->mfr == CFI_MFR_AMD && cfi->id == 0x0c01; +} + +static int __xipram chip_good_for_write(struct map_info *map, + struct flchip *chip, unsigned long addr, + map_word expected) +{ + if (cfi_use_chip_ready_for_write(map)) + return chip_ready(map, chip, addr); + + return chip_good(map, chip, addr, expected); +} + static int get_chip(struct map_info *map, struct flchip *chip, unsigned long adr, int mode) { DECLARE_WAITQUEUE(wait, current); @@ -1686,7 +1703,7 @@ static int __xipram do_write_oneword_once(struct map_info *map, * "chip_good" to avoid the failure due to scheduling. */ if (time_after(jiffies, timeo) && - !chip_good(map, chip, adr, datum)) { + !chip_good_for_write(map, chip, adr, datum)) { xip_enable(map, chip, adr); printk(KERN_WARNING "MTD %s(): software timeout\n", __func__); xip_disable(map, chip, adr); @@ -1694,7 +1711,7 @@ static int __xipram do_write_oneword_once(struct map_info *map, break; } - if (chip_good(map, chip, adr, datum)) { + if (chip_good_for_write(map, chip, adr, datum)) { if (cfi_check_err_status(map, chip, adr)) ret = -EIO; break; @@ -1966,14 +1983,14 @@ static int __xipram do_write_buffer_wait(struct map_info *map, * "chip_good" to avoid the failure due to scheduling. */ if (time_after(jiffies, timeo) && - !chip_good(map, chip, adr, datum)) { + !chip_good_for_write(map, chip, adr, datum)) { pr_err("MTD %s(): software timeout, address:0x%.8lx.\n", __func__, adr); ret = -EIO; break; } - if (chip_good(map, chip, adr, datum)) { + if (chip_good_for_write(map, chip, adr, datum)) { if (cfi_check_err_status(map, chip, adr)) ret = -EIO; break;
As pointed out by this bug report [1], buffered writes are now broken on S29GL064N. This issue comes from a rework which switched from using chip_good() to chip_ready(), because DQ true data 0xFF is read on S29GL064N and an error returned by chip_good(). One way to solve the issue is to revert the change partially to use chip_ready for S29GL064N. [1] https://lore.kernel.org/r/b687c259-6413-26c9-d4c9-b3afa69ea124@pengutronix.de/ Fixes: dfeae1073583("mtd: cfi_cmdset_0002: Change write buffer to check correct value") Signed-off-by: Tokunori Ikegami <ikegami.t@gmail.com> Tested-by: Ahmad Fatoum <a.fatoum@pengutronix.de> Cc: stable@vger.kernel.org --- drivers/mtd/chips/cfi_cmdset_0002.c | 25 +++++++++++++++++++++---- 1 file changed, 21 insertions(+), 4 deletions(-)