Re: [LEDE-DEV] fstools testing help
Am 13.08.2016 um 20:16 schrieb Daniel Dickinson: > On Tue, 2016-08-09 at 17:13 +0200, Josua Mayer wrote: >> Hi John, >> >> I have finally found the time to look into this, and managed to >> reproduce the segfault that Álvaro noticed. >> It is caused by a call to fclose(0x0); This is very unexpected, >> shouldn't the C library check for NULL and return quietly? > > C generally doesn't do any prevent shooting yourself in the foot type > checks for things related to memory/fd/handle management. Part of the > reason is fast and small is that it leaves most things up to the > programmer. I've head RUST is supposed to be a new thing that might be > small and fast but safer, but that's just what I'm told, I haven't > looked into it, and don't really know anything about it. I did some research, and apparently fclose(0) is undefined. To each C library its own behaviour. Luckily that call is now properly prevented! > > Regards, > > Daniel > ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] fstools testing help
On Tue, 2016-08-09 at 17:13 +0200, Josua Mayer wrote: > Hi John, > > I have finally found the time to look into this, and managed to > reproduce the segfault that Álvaro noticed. > It is caused by a call to fclose(0x0); This is very unexpected, > shouldn't the C library check for NULL and return quietly? C generally doesn't do any prevent shooting yourself in the foot type checks for things related to memory/fd/handle management. Part of the reason is fast and small is that it leaves most things up to the programmer. I've head RUST is supposed to be a new thing that might be small and fast but safer, but that's just what I'm told, I haven't looked into it, and don't really know anything about it. Regards, Daniel ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] fstools testing help
Hi John, I have finally found the time to look into this, and managed to reproduce the segfault that Álvaro noticed. It is caused by a call to fclose(0x0); This is very unexpected, shouldn't the C library check for NULL and return quietly? Either way, I submitted the new patchset for you to review. br Josua Mayer Am 08.07.2016 um 07:47 schrieb John Crispin: > > On 07/07/2016 21:36, Josua Mayer wrote: >> Hi John, >> >> Could you please point me at where you merged my 2 patches? I want to >> ensure I am testing the same code that was causing segfaults. >> >> br >> Josua Mayer >> > > I already dropped them. > > John > >> Am 05.07.2016 um 14:33 schrieb Álvaro Fernández Rojas: >>> Hi Josua >>> >>> Not mounted rw. >>> Unfortunately I didn't have time to check it thoroughly and I didn't see >>> any remarkable messages. >>> >>> – Álvaro. >>> >>> El 5/7/16 a las 14:31, Josua Mayer escribió: Hi Álvaro, Am 05.07.2016 um 14:17 schrieb Álvaro Fernández Rojas: > It fails when trying to boot a Raspberry Pi, and then it segfaults when > calling mount_root without any parameters.\ How do you see it fail when it boots? Messages? / not mounted rw? > > – Álvaro > > El 5/7/16 a las 14:14, Josua Mayer escribió: >> Hi John, >> >> Am 05.07.2016 um 09:44 schrieb John Crispin: >>> On 30/06/2016 09:23, Álvaro Fernández Rojas wrote: I will try to test it on RPi tonight. Regards, Álvaro. El 30/6/16 a las 9:21, John Crispin escribió: > Hi, > > i have just merged to fstools patches into my staging tree that > address > a double mount issue and add back ext4 support. the ext4 patch > previously broke rPi and other ext4 targets, but i thini with josua's > patch applied it should work. i have done some basic testing already > my > end but would like further etst coverage to make sure there are no > hidden regressions > > things that should be tested are > > * does ext4 mount work on eMMC > * does rootfs mount work on rPi > * does extroot still work with flat mounts > * does extroot still work with flat overlay mounts > > i have only tested the last one so far. > > John > >>> Hi Josua, >>> >>> after applying your patches, users of flat ext4 are seeing segfaults >>> when calling mount_root >> when manually calling mount_root? With which parameters? >> I will certainly look into it though. >>> John >> >> ___ >> Lede-dev mailing list >> Lede-dev@lists.infradead.org >> http://lists.infradead.org/mailman/listinfo/lede-dev >> ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] fstools testing help
Hi John, Am 08.07.2016 um 07:47 schrieb John Crispin: > > On 07/07/2016 21:36, Josua Mayer wrote: >> Hi John, >> >> Could you please point me at where you merged my 2 patches? I want to >> ensure I am testing the same code that was causing segfaults. >> >> br >> Josua Mayer >> > > I already dropped them. : >_< > > John > Well short story: given following setup: mmclko0p1: fat32 (zImage, dtb) mmcblk0p2: ext4 (rootfs) bootargs root=/dev/mmcblk0p2 ro no segfaults when running mount_root no memleaks reported by valgrind Any ideas what might be going wrong? >> Am 05.07.2016 um 14:33 schrieb Álvaro Fernández Rojas: >>> Hi Josua >>> >>> Not mounted rw. >>> Unfortunately I didn't have time to check it thoroughly and I didn't see >>> any remarkable messages. >>> >>> – Álvaro. >>> >>> El 5/7/16 a las 14:31, Josua Mayer escribió: Hi Álvaro, Am 05.07.2016 um 14:17 schrieb Álvaro Fernández Rojas: > It fails when trying to boot a Raspberry Pi, and then it segfaults when > calling mount_root without any parameters.\ How do you see it fail when it boots? Messages? / not mounted rw? > > – Álvaro > > El 5/7/16 a las 14:14, Josua Mayer escribió: >> Hi John, >> >> Am 05.07.2016 um 09:44 schrieb John Crispin: >>> On 30/06/2016 09:23, Álvaro Fernández Rojas wrote: I will try to test it on RPi tonight. Regards, Álvaro. El 30/6/16 a las 9:21, John Crispin escribió: > Hi, > > i have just merged to fstools patches into my staging tree that > address > a double mount issue and add back ext4 support. the ext4 patch > previously broke rPi and other ext4 targets, but i thini with josua's > patch applied it should work. i have done some basic testing already > my > end but would like further etst coverage to make sure there are no > hidden regressions > > things that should be tested are > > * does ext4 mount work on eMMC > * does rootfs mount work on rPi > * does extroot still work with flat mounts > * does extroot still work with flat overlay mounts > > i have only tested the last one so far. > > John > >>> Hi Josua, >>> >>> after applying your patches, users of flat ext4 are seeing segfaults >>> when calling mount_root >> when manually calling mount_root? With which parameters? >> I will certainly look into it though. >>> John >> >> ___ >> Lede-dev mailing list >> Lede-dev@lists.infradead.org >> http://lists.infradead.org/mailman/listinfo/lede-dev >> ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] fstools testing help
On 07/07/2016 21:36, Josua Mayer wrote: > Hi John, > > Could you please point me at where you merged my 2 patches? I want to > ensure I am testing the same code that was causing segfaults. > > br > Josua Mayer > I already dropped them. John > Am 05.07.2016 um 14:33 schrieb Álvaro Fernández Rojas: >> Hi Josua >> >> Not mounted rw. >> Unfortunately I didn't have time to check it thoroughly and I didn't see any >> remarkable messages. >> >> – Álvaro. >> >> El 5/7/16 a las 14:31, Josua Mayer escribió: >>> Hi Álvaro, >>> >>> Am 05.07.2016 um 14:17 schrieb Álvaro Fernández Rojas: It fails when trying to boot a Raspberry Pi, and then it segfaults when calling mount_root without any parameters.\ >>> How do you see it fail when it boots? Messages? / not mounted rw? – Álvaro El 5/7/16 a las 14:14, Josua Mayer escribió: > Hi John, > > Am 05.07.2016 um 09:44 schrieb John Crispin: >> On 30/06/2016 09:23, Álvaro Fernández Rojas wrote: >>> I will try to test it on RPi tonight. >>> >>> Regards, >>> Álvaro. >>> >>> El 30/6/16 a las 9:21, John Crispin escribió: Hi, i have just merged to fstools patches into my staging tree that address a double mount issue and add back ext4 support. the ext4 patch previously broke rPi and other ext4 targets, but i thini with josua's patch applied it should work. i have done some basic testing already my end but would like further etst coverage to make sure there are no hidden regressions things that should be tested are * does ext4 mount work on eMMC * does rootfs mount work on rPi * does extroot still work with flat mounts * does extroot still work with flat overlay mounts i have only tested the last one so far. John >> Hi Josua, >> >> after applying your patches, users of flat ext4 are seeing segfaults >> when calling mount_root > when manually calling mount_root? With which parameters? > I will certainly look into it though. >> John >>> > > ___ > Lede-dev mailing list > Lede-dev@lists.infradead.org > http://lists.infradead.org/mailman/listinfo/lede-dev > ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] fstools testing help
Hi John, Could you please point me at where you merged my 2 patches? I want to ensure I am testing the same code that was causing segfaults. br Josua Mayer Am 05.07.2016 um 14:33 schrieb Álvaro Fernández Rojas: > Hi Josua > > Not mounted rw. > Unfortunately I didn't have time to check it thoroughly and I didn't see any > remarkable messages. > > – Álvaro. > > El 5/7/16 a las 14:31, Josua Mayer escribió: >> Hi Álvaro, >> >> Am 05.07.2016 um 14:17 schrieb Álvaro Fernández Rojas: >>> It fails when trying to boot a Raspberry Pi, and then it segfaults when >>> calling mount_root without any parameters.\ >> How do you see it fail when it boots? Messages? / not mounted rw? >>> >>> – Álvaro >>> >>> El 5/7/16 a las 14:14, Josua Mayer escribió: Hi John, Am 05.07.2016 um 09:44 schrieb John Crispin: > On 30/06/2016 09:23, Álvaro Fernández Rojas wrote: >> I will try to test it on RPi tonight. >> >> Regards, >> Álvaro. >> >> El 30/6/16 a las 9:21, John Crispin escribió: >>> Hi, >>> >>> i have just merged to fstools patches into my staging tree that address >>> a double mount issue and add back ext4 support. the ext4 patch >>> previously broke rPi and other ext4 targets, but i thini with josua's >>> patch applied it should work. i have done some basic testing already my >>> end but would like further etst coverage to make sure there are no >>> hidden regressions >>> >>> things that should be tested are >>> >>> * does ext4 mount work on eMMC >>> * does rootfs mount work on rPi >>> * does extroot still work with flat mounts >>> * does extroot still work with flat overlay mounts >>> >>> i have only tested the last one so far. >>> >>> John >>> > Hi Josua, > > after applying your patches, users of flat ext4 are seeing segfaults > when calling mount_root when manually calling mount_root? With which parameters? I will certainly look into it though. > John >> ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] fstools testing help
Hi Josua Not mounted rw. Unfortunately I didn't have time to check it thoroughly and I didn't see any remarkable messages. – Álvaro. El 5/7/16 a las 14:31, Josua Mayer escribió: > Hi Álvaro, > > Am 05.07.2016 um 14:17 schrieb Álvaro Fernández Rojas: >> It fails when trying to boot a Raspberry Pi, and then it segfaults when >> calling mount_root without any parameters.\ > How do you see it fail when it boots? Messages? / not mounted rw? >> >> – Álvaro >> >> El 5/7/16 a las 14:14, Josua Mayer escribió: >>> Hi John, >>> >>> Am 05.07.2016 um 09:44 schrieb John Crispin: On 30/06/2016 09:23, Álvaro Fernández Rojas wrote: > I will try to test it on RPi tonight. > > Regards, > Álvaro. > > El 30/6/16 a las 9:21, John Crispin escribió: >> Hi, >> >> i have just merged to fstools patches into my staging tree that address >> a double mount issue and add back ext4 support. the ext4 patch >> previously broke rPi and other ext4 targets, but i thini with josua's >> patch applied it should work. i have done some basic testing already my >> end but would like further etst coverage to make sure there are no >> hidden regressions >> >> things that should be tested are >> >> * does ext4 mount work on eMMC >> * does rootfs mount work on rPi >> * does extroot still work with flat mounts >> * does extroot still work with flat overlay mounts >> >> i have only tested the last one so far. >> >> John >> Hi Josua, after applying your patches, users of flat ext4 are seeing segfaults when calling mount_root >>> when manually calling mount_root? With which parameters? >>> I will certainly look into it though. John > ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] fstools testing help
Hi Álvaro, Am 05.07.2016 um 14:17 schrieb Álvaro Fernández Rojas: > It fails when trying to boot a Raspberry Pi, and then it segfaults when > calling mount_root without any parameters.\ How do you see it fail when it boots? Messages? / not mounted rw? > > – Álvaro > > El 5/7/16 a las 14:14, Josua Mayer escribió: >> Hi John, >> >> Am 05.07.2016 um 09:44 schrieb John Crispin: >>> On 30/06/2016 09:23, Álvaro Fernández Rojas wrote: I will try to test it on RPi tonight. Regards, Álvaro. El 30/6/16 a las 9:21, John Crispin escribió: > Hi, > > i have just merged to fstools patches into my staging tree that address > a double mount issue and add back ext4 support. the ext4 patch > previously broke rPi and other ext4 targets, but i thini with josua's > patch applied it should work. i have done some basic testing already my > end but would like further etst coverage to make sure there are no > hidden regressions > > things that should be tested are > > * does ext4 mount work on eMMC > * does rootfs mount work on rPi > * does extroot still work with flat mounts > * does extroot still work with flat overlay mounts > > i have only tested the last one so far. > > John > >>> Hi Josua, >>> >>> after applying your patches, users of flat ext4 are seeing segfaults >>> when calling mount_root >> when manually calling mount_root? With which parameters? >> I will certainly look into it though. >>> John ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] fstools testing help
It fails when trying to boot a Raspberry Pi, and then it segfaults when calling mount_root without any parameters. – Álvaro El 5/7/16 a las 14:14, Josua Mayer escribió: > Hi John, > > Am 05.07.2016 um 09:44 schrieb John Crispin: >> >> On 30/06/2016 09:23, Álvaro Fernández Rojas wrote: >>> I will try to test it on RPi tonight. >>> >>> Regards, >>> Álvaro. >>> >>> El 30/6/16 a las 9:21, John Crispin escribió: Hi, i have just merged to fstools patches into my staging tree that address a double mount issue and add back ext4 support. the ext4 patch previously broke rPi and other ext4 targets, but i thini with josua's patch applied it should work. i have done some basic testing already my end but would like further etst coverage to make sure there are no hidden regressions things that should be tested are * does ext4 mount work on eMMC * does rootfs mount work on rPi * does extroot still work with flat mounts * does extroot still work with flat overlay mounts i have only tested the last one so far. John >> Hi Josua, >> >> after applying your patches, users of flat ext4 are seeing segfaults >> when calling mount_root > when manually calling mount_root? With which parameters? > I will certainly look into it though. >> >> John > ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] fstools testing help
Hi John, Am 05.07.2016 um 09:44 schrieb John Crispin: > > On 30/06/2016 09:23, Álvaro Fernández Rojas wrote: >> I will try to test it on RPi tonight. >> >> Regards, >> Álvaro. >> >> El 30/6/16 a las 9:21, John Crispin escribió: >>> Hi, >>> >>> i have just merged to fstools patches into my staging tree that address >>> a double mount issue and add back ext4 support. the ext4 patch >>> previously broke rPi and other ext4 targets, but i thini with josua's >>> patch applied it should work. i have done some basic testing already my >>> end but would like further etst coverage to make sure there are no >>> hidden regressions >>> >>> things that should be tested are >>> >>> * does ext4 mount work on eMMC >>> * does rootfs mount work on rPi >>> * does extroot still work with flat mounts >>> * does extroot still work with flat overlay mounts >>> >>> i have only tested the last one so far. >>> >>> John >>> > Hi Josua, > > after applying your patches, users of flat ext4 are seeing segfaults > when calling mount_root when manually calling mount_root? With which parameters? I will certainly look into it though. > > John ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] fstools testing help
On 30/06/2016 09:23, Álvaro Fernández Rojas wrote: > I will try to test it on RPi tonight. > > Regards, > Álvaro. > > El 30/6/16 a las 9:21, John Crispin escribió: >> Hi, >> >> i have just merged to fstools patches into my staging tree that address >> a double mount issue and add back ext4 support. the ext4 patch >> previously broke rPi and other ext4 targets, but i thini with josua's >> patch applied it should work. i have done some basic testing already my >> end but would like further etst coverage to make sure there are no >> hidden regressions >> >> things that should be tested are >> >> * does ext4 mount work on eMMC >> * does rootfs mount work on rPi >> * does extroot still work with flat mounts >> * does extroot still work with flat overlay mounts >> >> i have only tested the last one so far. >> >> John >> Hi Josua, after applying your patches, users of flat ext4 are seeing segfaults when calling mount_root John ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] fstools testing help
I will try to test it on RPi tonight. Regards, Álvaro. El 30/6/16 a las 9:21, John Crispin escribió: > Hi, > > i have just merged to fstools patches into my staging tree that address > a double mount issue and add back ext4 support. the ext4 patch > previously broke rPi and other ext4 targets, but i thini with josua's > patch applied it should work. i have done some basic testing already my > end but would like further etst coverage to make sure there are no > hidden regressions > > things that should be tested are > > * does ext4 mount work on eMMC > * does rootfs mount work on rPi > * does extroot still work with flat mounts > * does extroot still work with flat overlay mounts > > i have only tested the last one so far. > > John > ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
[LEDE-DEV] fstools testing help
Hi, i have just merged to fstools patches into my staging tree that address a double mount issue and add back ext4 support. the ext4 patch previously broke rPi and other ext4 targets, but i thini with josua's patch applied it should work. i have done some basic testing already my end but would like further etst coverage to make sure there are no hidden regressions things that should be tested are * does ext4 mount work on eMMC * does rootfs mount work on rPi * does extroot still work with flat mounts * does extroot still work with flat overlay mounts i have only tested the last one so far. John ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev