Message ID | 1519150999-27982-2-git-send-email-tharvey@gateworks.com |
---|---|
State | Changes Requested |
Headers | show |
Series | [LEDE-DEV,OpenWrt-Devel,1/8] octeontx: add USB_PCI support | expand |
On 2018-02-20 19:23, Tim Harvey wrote: > We are using F2FS as our overlay read/write FS. Others can be supported > as modules. > > Signed-off-by: Tim Harvey <tharvey@gateworks.com> I think it would be a good idea to leave in ext4. If the overlay storage area is too small for f2fs, fstools chooses ext4 instead. - Felix
On Wed, Feb 21, 2018 at 2:36 PM, Felix Fietkau <nbd@nbd.name> wrote: > On 2018-02-20 19:23, Tim Harvey wrote: >> We are using F2FS as our overlay read/write FS. Others can be supported >> as modules. >> >> Signed-off-by: Tim Harvey <tharvey@gateworks.com> > I think it would be a good idea to leave in ext4. If the overlay storage > area is too small for f2fs, fstools chooses ext4 instead. > > - Felix Felix, That makes sense. I think the other patches apply if that one is skipped, or would you rather I re-submit a new sieres (assuming there is no other feedback) Tim
On 2018-02-22 01:12, Tim Harvey wrote: > On Wed, Feb 21, 2018 at 2:36 PM, Felix Fietkau <nbd@nbd.name> wrote: >> On 2018-02-20 19:23, Tim Harvey wrote: >>> We are using F2FS as our overlay read/write FS. Others can be supported >>> as modules. >>> >>> Signed-off-by: Tim Harvey <tharvey@gateworks.com> >> I think it would be a good idea to leave in ext4. If the overlay storage >> area is too small for f2fs, fstools chooses ext4 instead. >> >> - Felix > > Felix, > > That makes sense. I think the other patches apply if that one is > skipped, or would you rather I re-submit a new sieres (assuming there > is no other feedback) No need to resubmit the series, you can send an updated version of this patch (removing only ext2/ext3) separately. - Felix
On 2018-02-22 09:34, Felix Fietkau wrote: > On 2018-02-22 01:12, Tim Harvey wrote: >> On Wed, Feb 21, 2018 at 2:36 PM, Felix Fietkau <nbd@nbd.name> wrote: >>> On 2018-02-20 19:23, Tim Harvey wrote: >>>> We are using F2FS as our overlay read/write FS. Others can be supported >>>> as modules. >>>> >>>> Signed-off-by: Tim Harvey <tharvey@gateworks.com> >>> I think it would be a good idea to leave in ext4. If the overlay storage >>> area is too small for f2fs, fstools chooses ext4 instead. >>> >>> - Felix >> >> Felix, >> >> That makes sense. I think the other patches apply if that one is >> skipped, or would you rather I re-submit a new sieres (assuming there >> is no other feedback) > No need to resubmit the series, you can send an updated version of this > patch (removing only ext2/ext3) separately. No need to do that anymore. It turns out that there were a lot more bogus config overrides in that target, some of them breaking the build in various other places. I've cleaned it up and verified that a build with all kernel modules enabled now actually works ;) - Felix
On Thu, Feb 22, 2018 at 4:14 AM, Felix Fietkau <nbd@nbd.name> wrote: > On 2018-02-22 09:34, Felix Fietkau wrote: >> On 2018-02-22 01:12, Tim Harvey wrote: >>> On Wed, Feb 21, 2018 at 2:36 PM, Felix Fietkau <nbd@nbd.name> wrote: >>>> On 2018-02-20 19:23, Tim Harvey wrote: >>>>> We are using F2FS as our overlay read/write FS. Others can be supported >>>>> as modules. >>>>> >>>>> Signed-off-by: Tim Harvey <tharvey@gateworks.com> >>>> I think it would be a good idea to leave in ext4. If the overlay storage >>>> area is too small for f2fs, fstools chooses ext4 instead. >>>> >>>> - Felix >>> >>> Felix, >>> >>> That makes sense. I think the other patches apply if that one is >>> skipped, or would you rather I re-submit a new sieres (assuming there >>> is no other feedback) >> No need to resubmit the series, you can send an updated version of this >> patch (removing only ext2/ext3) separately. > No need to do that anymore. It turns out that there were a lot more > bogus config overrides in that target, some of them breaking the build > in various other places. > > I've cleaned it up and verified that a build with all kernel modules > enabled now actually works ;) > > - Felix Thanks Felix! Can you explain what I could have done to see or catch these? Is this something that came up because it was a new target arch or is it something to look out for when bumping from one kernel to another as well? I'm not clear what the best practices are for creating a kernel config for OpenWrt. Regards, Tim
On 2018-02-22 16:39, Tim Harvey wrote: > On Thu, Feb 22, 2018 at 4:14 AM, Felix Fietkau <nbd@nbd.name> wrote: >> On 2018-02-22 09:34, Felix Fietkau wrote: >>> On 2018-02-22 01:12, Tim Harvey wrote: >>>> On Wed, Feb 21, 2018 at 2:36 PM, Felix Fietkau <nbd@nbd.name> wrote: >>>>> On 2018-02-20 19:23, Tim Harvey wrote: >>>>>> We are using F2FS as our overlay read/write FS. Others can be supported >>>>>> as modules. >>>>>> >>>>>> Signed-off-by: Tim Harvey <tharvey@gateworks.com> >>>>> I think it would be a good idea to leave in ext4. If the overlay storage >>>>> area is too small for f2fs, fstools chooses ext4 instead. >>>>> >>>>> - Felix >>>> >>>> Felix, >>>> >>>> That makes sense. I think the other patches apply if that one is >>>> skipped, or would you rather I re-submit a new sieres (assuming there >>>> is no other feedback) >>> No need to resubmit the series, you can send an updated version of this >>> patch (removing only ext2/ext3) separately. >> No need to do that anymore. It turns out that there were a lot more >> bogus config overrides in that target, some of them breaking the build >> in various other places. >> >> I've cleaned it up and verified that a build with all kernel modules >> enabled now actually works ;) >> >> - Felix > > Thanks Felix! > > Can you explain what I could have done to see or catch these? Is this > something that came up because it was a new target arch or is it > something to look out for when bumping from one kernel to another as > well? I'm not clear what the best practices are for creating a kernel > config for OpenWrt. I open up config-* in an editor and look at every line and ask myself whether it makes sense for the target to set it. I simply delete everything that doesn't make sense to me, then run make kernel_oldconfig afterwards. - Felix
diff --git a/target/linux/octeontx/config-4.14 b/target/linux/octeontx/config-4.14 index c56a27d..34533ef 100644 --- a/target/linux/octeontx/config-4.14 +++ b/target/linux/octeontx/config-4.14 @@ -221,12 +221,6 @@ CONFIG_EDAC_THUNDERX=y CONFIG_EEPROM_AT24=y # CONFIG_EVM is not set CONFIG_EXPORTFS=y -CONFIG_EXT2_FS=y -CONFIG_EXT3_FS=y -# CONFIG_EXT3_FS_POSIX_ACL is not set -# CONFIG_EXT3_FS_SECURITY is not set -CONFIG_EXT4_FS=y -CONFIG_EXT4_FS_POSIX_ACL=y # CONFIG_F2FS_CHECK_FS is not set CONFIG_F2FS_FS=y # CONFIG_F2FS_FS_SECURITY is not set @@ -239,8 +233,6 @@ CONFIG_FIXED_PHY=y CONFIG_FIX_EARLYCON_MEM=y CONFIG_FRAME_POINTER=y CONFIG_FREEZER=y -CONFIG_FS_MBCACHE=y -CONFIG_FS_POSIX_ACL=y CONFIG_FUTEX_PI=y CONFIG_GENERIC_ALLOCATOR=y CONFIG_GENERIC_ARCH_TOPOLOGY=y @@ -363,7 +355,6 @@ CONFIG_IRQ_FORCED_THREADING=y CONFIG_IRQ_TIME_ACCOUNTING=y CONFIG_IRQ_WORK=y # CONFIG_ISDN is not set -CONFIG_JBD2=y CONFIG_JUMP_LABEL=y CONFIG_KEXEC=y CONFIG_KEXEC_CORE=y
We are using F2FS as our overlay read/write FS. Others can be supported as modules. Signed-off-by: Tim Harvey <tharvey@gateworks.com> --- target/linux/octeontx/config-4.14 | 9 --------- 1 file changed, 9 deletions(-)