| Message ID | 20210902071951.28722-1-o.rempel@pengutronix.de (mailing list archive) |
|---|---|
| State | Superseded |
| Commit | bea6a94a279bcbe6b2cde348782b28baf12255a5 |
| Headers | show |
| Series | [v1] MIPS: Malta: fix alignment of the devicetree buffer | expand |
On Thu, Sep 02, 2021 at 09:19:51AM +0200, Oleksij Rempel wrote: > Starting with following patch MIPS Malta is not able to boot: > | commit 79edff12060fe7772af08607eff50c0e2486c5ba > | Author: Rob Herring <robh@kernel.org> > | scripts/dtc: Update to upstream version v1.6.0-51-g183df9e9c2b9 > > The reason is the alignment test added to the fdt_ro_probe_(). To fix > this issue, we need to make sure that fdt_buf is aligned. > > Since the dtc patch was designed to uncover potential issue, I handle > initial MIPS Malta patch as initial bug. > > Fixes: e81a8c7dabac ("MIPS: Malta: Setup RAM regions via DT") > Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de> > --- > arch/mips/mti-malta/malta-dtshim.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) applied to mips-next. Thomas.
On 02.09.2021 11:03:24, Thomas Bogendoerfer wrote: > On Thu, Sep 02, 2021 at 09:19:51AM +0200, Oleksij Rempel wrote: > > Starting with following patch MIPS Malta is not able to boot: > > | commit 79edff12060fe7772af08607eff50c0e2486c5ba > > | Author: Rob Herring <robh@kernel.org> > > | scripts/dtc: Update to upstream version v1.6.0-51-g183df9e9c2b9 If the above patch broke malts, IMHO the fixes tag should be: Fixes: 79edff12060f ("scripts/dtc: Update to upstream version v1.6.0-51-g183df9e9c2b9") > > The reason is the alignment test added to the fdt_ro_probe_(). To fix > > this issue, we need to make sure that fdt_buf is aligned. > > > > Since the dtc patch was designed to uncover potential issue, I handle > > initial MIPS Malta patch as initial bug. > > > > Fixes: e81a8c7dabac ("MIPS: Malta: Setup RAM regions via DT") > > Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de> > > --- > > arch/mips/mti-malta/malta-dtshim.c | 2 +- > > 1 file changed, 1 insertion(+), 1 deletion(-) > > applied to mips-next. Please add stable on Cc, as Malta is broken since v5.12 (79edff12060f == v5.12-rc1~104^2~10^2~3). regards, Marc
On Thu, 2 Sep 2021, Marc Kleine-Budde wrote: > > > Starting with following patch MIPS Malta is not able to boot: > > > | commit 79edff12060fe7772af08607eff50c0e2486c5ba > > > | Author: Rob Herring <robh@kernel.org> > > > | scripts/dtc: Update to upstream version v1.6.0-51-g183df9e9c2b9 > > If the above patch broke malts, IMHO the fixes tag should be: > > Fixes: 79edff12060f ("scripts/dtc: Update to upstream version v1.6.0-51-g183df9e9c2b9") It's not clear either from this patch description or from the commit referred what the actual *purpose* of the alignment check is and therefore I think it's hard to say which of the two commits needs to be indicated as the offender. Without further data I tend to agree it's 79edff12060f that ought to be referred, being the immediate cause of the regression, as things obviously worked before, for 5+ years. > > > The reason is the alignment test added to the fdt_ro_probe_(). To fix > > > this issue, we need to make sure that fdt_buf is aligned. > > > > > > Since the dtc patch was designed to uncover potential issue, I handle > > > initial MIPS Malta patch as initial bug. > > > > > > Fixes: e81a8c7dabac ("MIPS: Malta: Setup RAM regions via DT") > > > Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de> > > > --- > > > arch/mips/mti-malta/malta-dtshim.c | 2 +- > > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > applied to mips-next. > > Please add stable on Cc, as Malta is broken since v5.12 > (79edff12060f == v5.12-rc1~104^2~10^2~3). That for sure. Maciej
On Sun, Oct 10, 2021 at 11:58:24AM +0200, Maciej W. Rozycki wrote: > On Thu, 2 Sep 2021, Marc Kleine-Budde wrote: > > > > > Starting with following patch MIPS Malta is not able to boot: > > > > | commit 79edff12060fe7772af08607eff50c0e2486c5ba > > > > | Author: Rob Herring <robh@kernel.org> > > > > | scripts/dtc: Update to upstream version v1.6.0-51-g183df9e9c2b9 > > > > If the above patch broke malts, IMHO the fixes tag should be: > > > > Fixes: 79edff12060f ("scripts/dtc: Update to upstream version v1.6.0-51-g183df9e9c2b9") > > It's not clear either from this patch description or from the commit > referred what the actual *purpose* of the alignment check is and therefore > I think it's hard to say which of the two commits needs to be indicated as > the offender. Without further data I tend to agree it's 79edff12060f that > ought to be referred, being the immediate cause of the regression, as > things obviously worked before, for 5+ years. This patch was already taken. Do I need to do any thing? Regards, Oleksij
On Mon, 11 Oct 2021, Oleksij Rempel wrote: > > It's not clear either from this patch description or from the commit > > referred what the actual *purpose* of the alignment check is and therefore > > I think it's hard to say which of the two commits needs to be indicated as > > the offender. Without further data I tend to agree it's 79edff12060f that > > ought to be referred, being the immediate cause of the regression, as > > things obviously worked before, for 5+ years. > > This patch was already taken. Do I need to do any thing? No, unless Thomas chimes in, especially since you've posted v2 already (I hope he's sorted the stable part). Overall we need better change descriptions though. Maciej
diff --git a/arch/mips/mti-malta/malta-dtshim.c b/arch/mips/mti-malta/malta-dtshim.c index 0ddf03df6268..f451268f6c38 100644 --- a/arch/mips/mti-malta/malta-dtshim.c +++ b/arch/mips/mti-malta/malta-dtshim.c @@ -22,7 +22,7 @@ #define ROCIT_CONFIG_GEN1_MEMMAP_SHIFT 8 #define ROCIT_CONFIG_GEN1_MEMMAP_MASK (0xf << 8) -static unsigned char fdt_buf[16 << 10] __initdata; +static unsigned char fdt_buf[16 << 10] __initdata __aligned(8); /* determined physical memory size, not overridden by command line args */ extern unsigned long physical_memsize;
Starting with following patch MIPS Malta is not able to boot: | commit 79edff12060fe7772af08607eff50c0e2486c5ba | Author: Rob Herring <robh@kernel.org> | scripts/dtc: Update to upstream version v1.6.0-51-g183df9e9c2b9 The reason is the alignment test added to the fdt_ro_probe_(). To fix this issue, we need to make sure that fdt_buf is aligned. Since the dtc patch was designed to uncover potential issue, I handle initial MIPS Malta patch as initial bug. Fixes: e81a8c7dabac ("MIPS: Malta: Setup RAM regions via DT") Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de> --- arch/mips/mti-malta/malta-dtshim.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)