* [PATCH v1] erofs-utils: lib: Fix 8MB bug on uncompressed extent size
@ 2021-12-22 2:03 Kelvin Zhang via Linux-erofs
2021-12-22 3:21 ` Gao Xiang
2022-02-06 23:59 ` Gao Xiang
0 siblings, 2 replies; 11+ messages in thread
From: Kelvin Zhang via Linux-erofs @ 2021-12-22 2:03 UTC (permalink / raw)
To: linux-erofs mailing list, Miao Xie, Fang Wei; +Cc: Kelvin Zhang
Previously, uncompressed extent can be at most 8MB before mkfs.erofs
crashes on some error condition. This is due to a minor bug in how
compressed indices are encoded. This patch fixes the issue.
Signed-off-by: Kelvin Zhang <zhangkelvin@google.com>
---
include/erofs_fs.h | 2 +-
lib/compress.c | 21 ++++++++++++++++++++-
2 files changed, 21 insertions(+), 2 deletions(-)
diff --git a/include/erofs_fs.h b/include/erofs_fs.h
index 9a91877..13eaf24 100644
--- a/include/erofs_fs.h
+++ b/include/erofs_fs.h
@@ -353,7 +353,7 @@ enum {
* compressed block count of a compressed extent (in logical clusters, aka.
* block count of a pcluster).
*/
-#define Z_EROFS_VLE_DI_D0_CBLKCNT (1 << 11)
+#define Z_EROFS_VLE_DI_D0_CBLKCNT (1U << 11)
struct z_erofs_vle_decompressed_index {
__le16 di_advise;
diff --git a/lib/compress.c b/lib/compress.c
index 98be7a2..23e571c 100644
--- a/lib/compress.c
+++ b/lib/compress.c
@@ -97,7 +97,26 @@ static void vle_write_indexes(struct z_erofs_vle_compress_ctx *ctx,
} else if (d0) {
type = Z_EROFS_VLE_CLUSTER_TYPE_NONHEAD;
- di.di_u.delta[0] = cpu_to_le16(d0);
+ /* If the |Z_EROFS_VLE_DI_D0_CBLKCNT| bit is set, parser
+ * will interpret |delta[0]| as size of pcluster, rather
+ * than distance to last head cluster. Normally this
+ * isn't a problem, because uncompressed extent size are
+ * below Z_EROFS_VLE_DI_D0_CBLKCNT * BLOCK_SIZE = 8MB.
+ * But with large pcluster it's possible to go over this
+ * number, resulting in corrupted compressed indices.
+ * To solve this, we use Z_EROFS_VLE_DI_D0_CBLKCNT-1 if
+ * the uncompressed extent size goes above 8MB. This is
+ * OK because if kernel sees another non-head cluster
+ * after going back by |delta[0]| blocks, kernel will
+ * just keep looking back.
+ */
+ if (d0 & Z_EROFS_VLE_DI_D0_CBLKCNT) {
+ di.di_u.delta[0] = max(
+ d0 & (~Z_EROFS_VLE_DI_D0_CBLKCNT),
+ Z_EROFS_VLE_DI_D0_CBLKCNT-1);
+ } else {
+ di.di_u.delta[0] = cpu_to_le16(d0);
+ }
di.di_u.delta[1] = cpu_to_le16(d1);
} else {
type = raw ? Z_EROFS_VLE_CLUSTER_TYPE_PLAIN :
--
2.34.1.448.ga2b2bfdf31-goog
^ permalink raw reply related [flat|nested] 11+ messages in thread
* Re: [PATCH v1] erofs-utils: lib: Fix 8MB bug on uncompressed extent size
2021-12-22 2:03 [PATCH v1] erofs-utils: lib: Fix 8MB bug on uncompressed extent size Kelvin Zhang via Linux-erofs
@ 2021-12-22 3:21 ` Gao Xiang
2022-01-18 23:46 ` Kelvin Zhang via Linux-erofs
2022-02-06 23:59 ` Gao Xiang
1 sibling, 1 reply; 11+ messages in thread
From: Gao Xiang @ 2021-12-22 3:21 UTC (permalink / raw)
To: Kelvin Zhang; +Cc: Miao Xie, linux-erofs mailing list
On Tue, Dec 21, 2021 at 06:03:07PM -0800, Kelvin Zhang wrote:
> Previously, uncompressed extent can be at most 8MB before mkfs.erofs
> crashes on some error condition. This is due to a minor bug in how
> compressed indices are encoded. This patch fixes the issue.
>
> Signed-off-by: Kelvin Zhang <zhangkelvin@google.com>
I have to hold this for a while and look into (evaluate) that when I
get a full free time... (maybe a week later.) I understand it's not
quite urgent for all of us currently..
I still stuck into ztailpacking inline feature for now...
Thanks,
Gao Xiang
> ---
> include/erofs_fs.h | 2 +-
> lib/compress.c | 21 ++++++++++++++++++++-
> 2 files changed, 21 insertions(+), 2 deletions(-)
>
> diff --git a/include/erofs_fs.h b/include/erofs_fs.h
> index 9a91877..13eaf24 100644
> --- a/include/erofs_fs.h
> +++ b/include/erofs_fs.h
> @@ -353,7 +353,7 @@ enum {
> * compressed block count of a compressed extent (in logical clusters, aka.
> * block count of a pcluster).
> */
> -#define Z_EROFS_VLE_DI_D0_CBLKCNT (1 << 11)
> +#define Z_EROFS_VLE_DI_D0_CBLKCNT (1U << 11)
>
> struct z_erofs_vle_decompressed_index {
> __le16 di_advise;
> diff --git a/lib/compress.c b/lib/compress.c
> index 98be7a2..23e571c 100644
> --- a/lib/compress.c
> +++ b/lib/compress.c
> @@ -97,7 +97,26 @@ static void vle_write_indexes(struct z_erofs_vle_compress_ctx *ctx,
> } else if (d0) {
> type = Z_EROFS_VLE_CLUSTER_TYPE_NONHEAD;
>
> - di.di_u.delta[0] = cpu_to_le16(d0);
> + /* If the |Z_EROFS_VLE_DI_D0_CBLKCNT| bit is set, parser
> + * will interpret |delta[0]| as size of pcluster, rather
> + * than distance to last head cluster. Normally this
> + * isn't a problem, because uncompressed extent size are
> + * below Z_EROFS_VLE_DI_D0_CBLKCNT * BLOCK_SIZE = 8MB.
> + * But with large pcluster it's possible to go over this
> + * number, resulting in corrupted compressed indices.
> + * To solve this, we use Z_EROFS_VLE_DI_D0_CBLKCNT-1 if
> + * the uncompressed extent size goes above 8MB. This is
> + * OK because if kernel sees another non-head cluster
> + * after going back by |delta[0]| blocks, kernel will
> + * just keep looking back.
> + */
> + if (d0 & Z_EROFS_VLE_DI_D0_CBLKCNT) {
> + di.di_u.delta[0] = max(
> + d0 & (~Z_EROFS_VLE_DI_D0_CBLKCNT),
> + Z_EROFS_VLE_DI_D0_CBLKCNT-1);
> + } else {
> + di.di_u.delta[0] = cpu_to_le16(d0);
> + }
> di.di_u.delta[1] = cpu_to_le16(d1);
> } else {
> type = raw ? Z_EROFS_VLE_CLUSTER_TYPE_PLAIN :
> --
> 2.34.1.448.ga2b2bfdf31-goog
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH v1] erofs-utils: lib: Fix 8MB bug on uncompressed extent size
2021-12-22 3:21 ` Gao Xiang
@ 2022-01-18 23:46 ` Kelvin Zhang via Linux-erofs
2022-01-19 3:49 ` Gao Xiang
0 siblings, 1 reply; 11+ messages in thread
From: Kelvin Zhang via Linux-erofs @ 2022-01-18 23:46 UTC (permalink / raw)
To: Gao Xiang; +Cc: Miao Xie, linux-erofs mailing list
friendly ping : )
On Tue, Dec 21, 2021 at 7:21 PM Gao Xiang <hsiangkao@linux.alibaba.com> wrote:
>
> On Tue, Dec 21, 2021 at 06:03:07PM -0800, Kelvin Zhang wrote:
> > Previously, uncompressed extent can be at most 8MB before mkfs.erofs
> > crashes on some error condition. This is due to a minor bug in how
> > compressed indices are encoded. This patch fixes the issue.
> >
> > Signed-off-by: Kelvin Zhang <zhangkelvin@google.com>
>
> I have to hold this for a while and look into (evaluate) that when I
> get a full free time... (maybe a week later.) I understand it's not
> quite urgent for all of us currently..
>
> I still stuck into ztailpacking inline feature for now...
>
> Thanks,
> Gao Xiang
>
> > ---
> > include/erofs_fs.h | 2 +-
> > lib/compress.c | 21 ++++++++++++++++++++-
> > 2 files changed, 21 insertions(+), 2 deletions(-)
> >
> > diff --git a/include/erofs_fs.h b/include/erofs_fs.h
> > index 9a91877..13eaf24 100644
> > --- a/include/erofs_fs.h
> > +++ b/include/erofs_fs.h
> > @@ -353,7 +353,7 @@ enum {
> > * compressed block count of a compressed extent (in logical clusters, aka.
> > * block count of a pcluster).
> > */
> > -#define Z_EROFS_VLE_DI_D0_CBLKCNT (1 << 11)
> > +#define Z_EROFS_VLE_DI_D0_CBLKCNT (1U << 11)
> >
> > struct z_erofs_vle_decompressed_index {
> > __le16 di_advise;
> > diff --git a/lib/compress.c b/lib/compress.c
> > index 98be7a2..23e571c 100644
> > --- a/lib/compress.c
> > +++ b/lib/compress.c
> > @@ -97,7 +97,26 @@ static void vle_write_indexes(struct z_erofs_vle_compress_ctx *ctx,
> > } else if (d0) {
> > type = Z_EROFS_VLE_CLUSTER_TYPE_NONHEAD;
> >
> > - di.di_u.delta[0] = cpu_to_le16(d0);
> > + /* If the |Z_EROFS_VLE_DI_D0_CBLKCNT| bit is set, parser
> > + * will interpret |delta[0]| as size of pcluster, rather
> > + * than distance to last head cluster. Normally this
> > + * isn't a problem, because uncompressed extent size are
> > + * below Z_EROFS_VLE_DI_D0_CBLKCNT * BLOCK_SIZE = 8MB.
> > + * But with large pcluster it's possible to go over this
> > + * number, resulting in corrupted compressed indices.
> > + * To solve this, we use Z_EROFS_VLE_DI_D0_CBLKCNT-1 if
> > + * the uncompressed extent size goes above 8MB. This is
> > + * OK because if kernel sees another non-head cluster
> > + * after going back by |delta[0]| blocks, kernel will
> > + * just keep looking back.
> > + */
> > + if (d0 & Z_EROFS_VLE_DI_D0_CBLKCNT) {
> > + di.di_u.delta[0] = max(
> > + d0 & (~Z_EROFS_VLE_DI_D0_CBLKCNT),
> > + Z_EROFS_VLE_DI_D0_CBLKCNT-1);
> > + } else {
> > + di.di_u.delta[0] = cpu_to_le16(d0);
> > + }
> > di.di_u.delta[1] = cpu_to_le16(d1);
> > } else {
> > type = raw ? Z_EROFS_VLE_CLUSTER_TYPE_PLAIN :
> > --
> > 2.34.1.448.ga2b2bfdf31-goog
--
Sincerely,
Kelvin Zhang
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH v1] erofs-utils: lib: Fix 8MB bug on uncompressed extent size
2022-01-18 23:46 ` Kelvin Zhang via Linux-erofs
@ 2022-01-19 3:49 ` Gao Xiang
0 siblings, 0 replies; 11+ messages in thread
From: Gao Xiang @ 2022-01-19 3:49 UTC (permalink / raw)
To: Kelvin Zhang; +Cc: Miao Xie, linux-erofs mailing list
Hi Kelvin,
On Tue, Jan 18, 2022 at 03:46:14PM -0800, Kelvin Zhang wrote:
> friendly ping : )
Thanks for your reminder.
It's still on my TODO list with many other stuffs along with many
internal paperwork.
I will finish it no later than the end of our Spring Festival.
Thanks,
Gao Xiang
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH v1] erofs-utils: lib: Fix 8MB bug on uncompressed extent size
2021-12-22 2:03 [PATCH v1] erofs-utils: lib: Fix 8MB bug on uncompressed extent size Kelvin Zhang via Linux-erofs
2021-12-22 3:21 ` Gao Xiang
@ 2022-02-06 23:59 ` Gao Xiang
2022-02-07 2:07 ` Gao Xiang
1 sibling, 1 reply; 11+ messages in thread
From: Gao Xiang @ 2022-02-06 23:59 UTC (permalink / raw)
To: Kelvin Zhang; +Cc: Miao Xie, linux-erofs mailing list
Hi Kelvin,
On Tue, Dec 21, 2021 at 06:03:07PM -0800, Kelvin Zhang wrote:
> Previously, uncompressed extent can be at most 8MB before mkfs.erofs
> crashes on some error condition. This is due to a minor bug in how
> compressed indices are encoded. This patch fixes the issue.
>
> Signed-off-by: Kelvin Zhang <zhangkelvin@google.com>
> ---
> include/erofs_fs.h | 2 +-
> lib/compress.c | 21 ++++++++++++++++++++-
> 2 files changed, 21 insertions(+), 2 deletions(-)
>
> diff --git a/include/erofs_fs.h b/include/erofs_fs.h
> index 9a91877..13eaf24 100644
> --- a/include/erofs_fs.h
> +++ b/include/erofs_fs.h
> @@ -353,7 +353,7 @@ enum {
> * compressed block count of a compressed extent (in logical clusters, aka.
> * block count of a pcluster).
> */
> -#define Z_EROFS_VLE_DI_D0_CBLKCNT (1 << 11)
> +#define Z_EROFS_VLE_DI_D0_CBLKCNT (1U << 11)
If erofs_fs.h update is necessary, I prefer to update in-kernel
header first. Would you mind making a kernel patch for this if needed?
>
> struct z_erofs_vle_decompressed_index {
> __le16 di_advise;
> diff --git a/lib/compress.c b/lib/compress.c
> index 98be7a2..23e571c 100644
> --- a/lib/compress.c
> +++ b/lib/compress.c
> @@ -97,7 +97,26 @@ static void vle_write_indexes(struct z_erofs_vle_compress_ctx *ctx,
> } else if (d0) {
> type = Z_EROFS_VLE_CLUSTER_TYPE_NONHEAD;
>
> - di.di_u.delta[0] = cpu_to_le16(d0);
> + /* If the |Z_EROFS_VLE_DI_D0_CBLKCNT| bit is set, parser
> + * will interpret |delta[0]| as size of pcluster, rather
> + * than distance to last head cluster. Normally this
> + * isn't a problem, because uncompressed extent size are
> + * below Z_EROFS_VLE_DI_D0_CBLKCNT * BLOCK_SIZE = 8MB.
> + * But with large pcluster it's possible to go over this
> + * number, resulting in corrupted compressed indices.
> + * To solve this, we use Z_EROFS_VLE_DI_D0_CBLKCNT-1 if
> + * the uncompressed extent size goes above 8MB. This is
> + * OK because if kernel sees another non-head cluster
> + * after going back by |delta[0]| blocks, kernel will
> + * just keep looking back.
> + */
Would you mind updating this into the kernel comment style, I mean
/*
* ...
*/
Instead?
> + if (d0 & Z_EROFS_VLE_DI_D0_CBLKCNT) {
> + di.di_u.delta[0] = max(
> + d0 & (~Z_EROFS_VLE_DI_D0_CBLKCNT),
> + Z_EROFS_VLE_DI_D0_CBLKCNT-1);
May I ask if it's actually tested with big pcluster feature? It's
lack of cpu_to_le16() convert and even the original
Z_EROFS_VLE_DI_D0_CBLKCNT flag.
Thanks,
Gao Xiang
> + } else {
> + di.di_u.delta[0] = cpu_to_le16(d0);
> + }
> di.di_u.delta[1] = cpu_to_le16(d1);
> } else {
> type = raw ? Z_EROFS_VLE_CLUSTER_TYPE_PLAIN :
> --
> 2.34.1.448.ga2b2bfdf31-goog
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH v1] erofs-utils: lib: Fix 8MB bug on uncompressed extent size
2022-02-06 23:59 ` Gao Xiang
@ 2022-02-07 2:07 ` Gao Xiang
2022-02-07 17:38 ` Kelvin Zhang via Linux-erofs
2022-02-07 17:39 ` [PATCH v2] " Kelvin Zhang via Linux-erofs
0 siblings, 2 replies; 11+ messages in thread
From: Gao Xiang @ 2022-02-07 2:07 UTC (permalink / raw)
To: Kelvin Zhang; +Cc: Miao Xie, linux-erofs mailing list
On Mon, Feb 07, 2022 at 07:59:50AM +0800, Gao Xiang wrote:
> Hi Kelvin,
>
> On Tue, Dec 21, 2021 at 06:03:07PM -0800, Kelvin Zhang wrote:
> > Previously, uncompressed extent can be at most 8MB before mkfs.erofs
> > crashes on some error condition. This is due to a minor bug in how
> > compressed indices are encoded. This patch fixes the issue.
> >
> > Signed-off-by: Kelvin Zhang <zhangkelvin@google.com>
> > ---
> > include/erofs_fs.h | 2 +-
> > lib/compress.c | 21 ++++++++++++++++++++-
> > 2 files changed, 21 insertions(+), 2 deletions(-)
> >
> > diff --git a/include/erofs_fs.h b/include/erofs_fs.h
> > index 9a91877..13eaf24 100644
> > --- a/include/erofs_fs.h
> > +++ b/include/erofs_fs.h
> > @@ -353,7 +353,7 @@ enum {
> > * compressed block count of a compressed extent (in logical clusters, aka.
> > * block count of a pcluster).
> > */
> > -#define Z_EROFS_VLE_DI_D0_CBLKCNT (1 << 11)
> > +#define Z_EROFS_VLE_DI_D0_CBLKCNT (1U << 11)
>
> If erofs_fs.h update is necessary, I prefer to update in-kernel
> header first. Would you mind making a kernel patch for this if needed?
>
> >
> > struct z_erofs_vle_decompressed_index {
> > __le16 di_advise;
> > diff --git a/lib/compress.c b/lib/compress.c
> > index 98be7a2..23e571c 100644
> > --- a/lib/compress.c
> > +++ b/lib/compress.c
> > @@ -97,7 +97,26 @@ static void vle_write_indexes(struct z_erofs_vle_compress_ctx *ctx,
> > } else if (d0) {
> > type = Z_EROFS_VLE_CLUSTER_TYPE_NONHEAD;
> >
> > - di.di_u.delta[0] = cpu_to_le16(d0);
> > + /* If the |Z_EROFS_VLE_DI_D0_CBLKCNT| bit is set, parser
> > + * will interpret |delta[0]| as size of pcluster, rather
> > + * than distance to last head cluster. Normally this
> > + * isn't a problem, because uncompressed extent size are
> > + * below Z_EROFS_VLE_DI_D0_CBLKCNT * BLOCK_SIZE = 8MB.
> > + * But with large pcluster it's possible to go over this
> > + * number, resulting in corrupted compressed indices.
> > + * To solve this, we use Z_EROFS_VLE_DI_D0_CBLKCNT-1 if
> > + * the uncompressed extent size goes above 8MB. This is
> > + * OK because if kernel sees another non-head cluster
> > + * after going back by |delta[0]| blocks, kernel will
> > + * just keep looking back.
> > + */
>
> Would you mind updating this into the kernel comment style, I mean
> /*
> * ...
> */
> Instead?
>
> > + if (d0 & Z_EROFS_VLE_DI_D0_CBLKCNT) {
> > + di.di_u.delta[0] = max(
> > + d0 & (~Z_EROFS_VLE_DI_D0_CBLKCNT),
> > + Z_EROFS_VLE_DI_D0_CBLKCNT-1);
>
> May I ask if it's actually tested with big pcluster feature? It's
> lack of cpu_to_le16() convert and even the original
> Z_EROFS_VLE_DI_D0_CBLKCNT flag.
Sorry this part shouldn't have Z_EROFS_VLE_DI_D0_CBLKCNT flag.
Btw, I think a proper change for this might be just:
if (d0 > Z_EROFS_VLE_DI_D0_CBLKCNT - 1)
di.di_u.delta[0] = le16_to_cpu(Z_EROFS_VLE_DI_D0_CBLKCNT - 1);
else
di.di_u.delta[0] = cpu_to_le16(d0);
Or using max() to simplify above even more a bit.
Thanks,
Gao Xiang
>
> Thanks,
> Gao Xiang
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH v1] erofs-utils: lib: Fix 8MB bug on uncompressed extent size
2022-02-07 2:07 ` Gao Xiang
@ 2022-02-07 17:38 ` Kelvin Zhang via Linux-erofs
2022-02-07 21:50 ` Gao Xiang
2022-02-07 17:39 ` [PATCH v2] " Kelvin Zhang via Linux-erofs
1 sibling, 1 reply; 11+ messages in thread
From: Kelvin Zhang via Linux-erofs @ 2022-02-07 17:38 UTC (permalink / raw)
To: Gao Xiang; +Cc: Miao Xie, linux-erofs mailing list
-1
On Sun, Feb 6, 2022 at 6:08 PM Gao Xiang <hsiangkao@linux.alibaba.com> wrote:
>
> On Mon, Feb 07, 2022 at 07:59:50AM +0800, Gao Xiang wrote:
> > Hi Kelvin,
> >
> > On Tue, Dec 21, 2021 at 06:03:07PM -0800, Kelvin Zhang wrote:
> > > Previously, uncompressed extent can be at most 8MB before mkfs.erofs
> > > crashes on some error condition. This is due to a minor bug in how
> > > compressed indices are encoded. This patch fixes the issue.
> > >
> > > Signed-off-by: Kelvin Zhang <zhangkelvin@google.com>
> > > ---
> > > include/erofs_fs.h | 2 +-
> > > lib/compress.c | 21 ++++++++++++++++++++-
> > > 2 files changed, 21 insertions(+), 2 deletions(-)
> > >
> > > diff --git a/include/erofs_fs.h b/include/erofs_fs.h
> > > index 9a91877..13eaf24 100644
> > > --- a/include/erofs_fs.h
> > > +++ b/include/erofs_fs.h
> > > @@ -353,7 +353,7 @@ enum {
> > > * compressed block count of a compressed extent (in logical clusters, aka.
> > > * block count of a pcluster).
> > > */
> > > -#define Z_EROFS_VLE_DI_D0_CBLKCNT (1 << 11)
> > > +#define Z_EROFS_VLE_DI_D0_CBLKCNT (1U << 11)
> >
> > If erofs_fs.h update is necessary, I prefer to update in-kernel
> > header first. Would you mind making a kernel patch for this if needed?
Reverted erofs_fs.h change.
> >
> > >
> > > struct z_erofs_vle_decompressed_index {
> > > __le16 di_advise;
> > > diff --git a/lib/compress.c b/lib/compress.c
> > > index 98be7a2..23e571c 100644
> > > --- a/lib/compress.c
> > > +++ b/lib/compress.c
> > > @@ -97,7 +97,26 @@ static void vle_write_indexes(struct z_erofs_vle_compress_ctx *ctx,
> > > } else if (d0) {
> > > type = Z_EROFS_VLE_CLUSTER_TYPE_NONHEAD;
> > >
> > > - di.di_u.delta[0] = cpu_to_le16(d0);
> > > + /* If the |Z_EROFS_VLE_DI_D0_CBLKCNT| bit is set, parser
> > > + * will interpret |delta[0]| as size of pcluster, rather
> > > + * than distance to last head cluster. Normally this
> > > + * isn't a problem, because uncompressed extent size are
> > > + * below Z_EROFS_VLE_DI_D0_CBLKCNT * BLOCK_SIZE = 8MB.
> > > + * But with large pcluster it's possible to go over this
> > > + * number, resulting in corrupted compressed indices.
> > > + * To solve this, we use Z_EROFS_VLE_DI_D0_CBLKCNT-1 if
> > > + * the uncompressed extent size goes above 8MB. This is
> > > + * OK because if kernel sees another non-head cluster
> > > + * after going back by |delta[0]| blocks, kernel will
> > > + * just keep looking back.
> > > + */
> >
> > Would you mind updating this into the kernel comment style, I mean
> > /*
> > * ...
> > */
> > Instead?
Done
> >
> > > + if (d0 & Z_EROFS_VLE_DI_D0_CBLKCNT) {
> > > + di.di_u.delta[0] = max(
> > > + d0 & (~Z_EROFS_VLE_DI_D0_CBLKCNT),
> > > + Z_EROFS_VLE_DI_D0_CBLKCNT-1);
> >
> > May I ask if it's actually tested with big pcluster feature? It's
> > lack of cpu_to_le16() convert and even the original
> > Z_EROFS_VLE_DI_D0_CBLKCNT flag.
Sorry.. It was tested on a Little Endian machine, so I didn't discover
the missing cpu_to_le16. Added now.
>
> Sorry this part shouldn't have Z_EROFS_VLE_DI_D0_CBLKCNT flag.
>
> Btw, I think a proper change for this might be just:
> if (d0 > Z_EROFS_VLE_DI_D0_CBLKCNT - 1)
> di.di_u.delta[0] = le16_to_cpu(Z_EROFS_VLE_DI_D0_CBLKCNT - 1);
> else
> di.di_u.delta[0] = cpu_to_le16(d0);
> Or using max() to simplify above even more a bit.
This would work, but it's not optimal. For example,
Z_EROFS_VLE_DI_D0_CBLKCNT << 1
is greater than Z_EROFS_VLE_DI_D0_CBLKCNT, but it does not have the
11th bit set.
Using Z_EROFS_VLE_DI_D0_CBLKCNT-1 in this case would cause the kernel to
take more hops than necessary when finding the head cluster. A better
change would be:
if (d0 > Z_EROFS_VLE_DI_D0_CBLKCNT - 1)
di.di_u.delta[0] = le16_to_cpu(largest number smaller
than d0 that does not have Z_EROFS_VLE_DI_D0_CBLKCNT bit set);
else
di.di_u.delta[0] = cpu_to_le16(d0);
But how do we find "largest number smaller than d0 that does not have
Z_EROFS_VLE_DI_D0_CBLKCNT bit set" ?
Simple, clear the Z_EROFS_VLE_DI_D0_CBLKCNT bit, and set all bits
before that to 1. In code:
d0 & (~ Z_EROFS_VLE_DI_D0_CBLKCNT) | (Z_EROFS_VLE_DI_D0_CBLKCNT-1)
So final answer:
if (d0 > Z_EROFS_VLE_DI_D0_CBLKCNT - 1)
di.di_u.delta[0] = le16_to_cpu(d0 & (~
Z_EROFS_VLE_DI_D0_CBLKCNT) | (Z_EROFS_VLE_DI_D0_CBLKCNT-1));
else
di.di_u.delta[0] = cpu_to_le16(d0);
>
> Thanks,
> Gao Xiang
>
> >
> > Thanks,
> > Gao Xiang
--
Sincerely,
Kelvin Zhang
^ permalink raw reply [flat|nested] 11+ messages in thread
* [PATCH v2] erofs-utils: lib: Fix 8MB bug on uncompressed extent size
2022-02-07 2:07 ` Gao Xiang
2022-02-07 17:38 ` Kelvin Zhang via Linux-erofs
@ 2022-02-07 17:39 ` Kelvin Zhang via Linux-erofs
1 sibling, 0 replies; 11+ messages in thread
From: Kelvin Zhang via Linux-erofs @ 2022-02-07 17:39 UTC (permalink / raw)
To: linux-erofs mailing list, Miao Xie, Fang Wei; +Cc: Kelvin Zhang
Previously, uncompressed extent can be at most 8MB before mkfs.erofs
crashes on some error condition. This is due to a minor bug in how
compressed indices are encoded. This patch fixes the issue.
Signed-off-by: Kelvin Zhang <zhangkelvin@google.com>
---
lib/compress.c | 32 +++++++++++++++++++++++++++++++-
1 file changed, 31 insertions(+), 1 deletion(-)
diff --git a/lib/compress.c b/lib/compress.c
index 98be7a2..2f7ffa7 100644
--- a/lib/compress.c
+++ b/lib/compress.c
@@ -97,7 +97,37 @@ static void vle_write_indexes(struct z_erofs_vle_compress_ctx *ctx,
} else if (d0) {
type = Z_EROFS_VLE_CLUSTER_TYPE_NONHEAD;
- di.di_u.delta[0] = cpu_to_le16(d0);
+ /*
+ * If the |Z_EROFS_VLE_DI_D0_CBLKCNT| bit is set, parser
+ * will interpret |delta[0]| as size of pcluster, rather
+ * than distance to last head cluster. Normally this
+ * isn't a problem, because uncompressed extent size are
+ * below Z_EROFS_VLE_DI_D0_CBLKCNT * BLOCK_SIZE = 8MB.
+ * But with large pcluster it's possible to go over this
+ * number, resulting in corrupted compressed indices.
+ * To solve this, we replace d0 with a number that's
+ * smaller and doesn't have the
+ * Z_EROFS_VLE_DI_D0_CBLKCNT bit set if
+ * the uncompressed extent size goes above 8MB. This is
+ * OK because if kernel sees another non-head cluster
+ * after going back by |delta[0]| blocks, kernel will
+ * just keep looking back.
+ * The largest number smaller than d0 that doesn't have
+ * Z_EROFS_VLE_DI_D0_CBLKCNT bit set is obtained by
+ * first clearing Z_EROFS_VLE_DI_D0_CBLKCNT bit, then
+ * set all bits before Z_EROFS_VLE_DI_D0_CBLKCNT to 1.
+ * Using Z_EROFS_VLE_DI_D0_CBLKCNT-1 would work, but it
+ * produces suboptimal indices in certain cases. e.g.
+ * (Z_EROFS_VLE_DI_D0_CBLKCNT<<4)|
+ * (Z_EROFS_VLE_DI_D0_CBLKCNT)
+ */
+ if (d0 & Z_EROFS_VLE_DI_D0_CBLKCNT) {
+ di.di_u.delta[0] = cpu_to_le16(
+ (d0 & (~Z_EROFS_VLE_DI_D0_CBLKCNT)) |
+ (Z_EROFS_VLE_DI_D0_CBLKCNT-1));
+ } else {
+ di.di_u.delta[0] = cpu_to_le16(d0);
+ }
di.di_u.delta[1] = cpu_to_le16(d1);
} else {
type = raw ? Z_EROFS_VLE_CLUSTER_TYPE_PLAIN :
--
2.35.0.263.gb82422642f-goog
^ permalink raw reply related [flat|nested] 11+ messages in thread
* Re: [PATCH v1] erofs-utils: lib: Fix 8MB bug on uncompressed extent size
2022-02-07 17:38 ` Kelvin Zhang via Linux-erofs
@ 2022-02-07 21:50 ` Gao Xiang
2022-02-08 18:43 ` [PATCH v3] " Kelvin Zhang via Linux-erofs
0 siblings, 1 reply; 11+ messages in thread
From: Gao Xiang @ 2022-02-07 21:50 UTC (permalink / raw)
To: Kelvin Zhang; +Cc: Miao Xie, linux-erofs mailing list
On Mon, Feb 07, 2022 at 09:38:45AM -0800, Kelvin Zhang wrote:
> -1
>
> On Sun, Feb 6, 2022 at 6:08 PM Gao Xiang <hsiangkao@linux.alibaba.com> wrote:
> >
> > On Mon, Feb 07, 2022 at 07:59:50AM +0800, Gao Xiang wrote:
> > > Hi Kelvin,
> > >
> > > On Tue, Dec 21, 2021 at 06:03:07PM -0800, Kelvin Zhang wrote:
> > > > Previously, uncompressed extent can be at most 8MB before mkfs.erofs
> > > > crashes on some error condition. This is due to a minor bug in how
> > > > compressed indices are encoded. This patch fixes the issue.
> > > >
> > > > Signed-off-by: Kelvin Zhang <zhangkelvin@google.com>
> > > > ---
> > > > include/erofs_fs.h | 2 +-
> > > > lib/compress.c | 21 ++++++++++++++++++++-
> > > > 2 files changed, 21 insertions(+), 2 deletions(-)
> > > >
> > > > diff --git a/include/erofs_fs.h b/include/erofs_fs.h
> > > > index 9a91877..13eaf24 100644
> > > > --- a/include/erofs_fs.h
> > > > +++ b/include/erofs_fs.h
> > > > @@ -353,7 +353,7 @@ enum {
> > > > * compressed block count of a compressed extent (in logical clusters, aka.
> > > > * block count of a pcluster).
> > > > */
> > > > -#define Z_EROFS_VLE_DI_D0_CBLKCNT (1 << 11)
> > > > +#define Z_EROFS_VLE_DI_D0_CBLKCNT (1U << 11)
> > >
> > > If erofs_fs.h update is necessary, I prefer to update in-kernel
> > > header first. Would you mind making a kernel patch for this if needed?
>
> Reverted erofs_fs.h change.
>
> > >
> > > >
> > > > struct z_erofs_vle_decompressed_index {
> > > > __le16 di_advise;
> > > > diff --git a/lib/compress.c b/lib/compress.c
> > > > index 98be7a2..23e571c 100644
> > > > --- a/lib/compress.c
> > > > +++ b/lib/compress.c
> > > > @@ -97,7 +97,26 @@ static void vle_write_indexes(struct z_erofs_vle_compress_ctx *ctx,
> > > > } else if (d0) {
> > > > type = Z_EROFS_VLE_CLUSTER_TYPE_NONHEAD;
> > > >
> > > > - di.di_u.delta[0] = cpu_to_le16(d0);
> > > > + /* If the |Z_EROFS_VLE_DI_D0_CBLKCNT| bit is set, parser
> > > > + * will interpret |delta[0]| as size of pcluster, rather
> > > > + * than distance to last head cluster. Normally this
> > > > + * isn't a problem, because uncompressed extent size are
> > > > + * below Z_EROFS_VLE_DI_D0_CBLKCNT * BLOCK_SIZE = 8MB.
> > > > + * But with large pcluster it's possible to go over this
> > > > + * number, resulting in corrupted compressed indices.
> > > > + * To solve this, we use Z_EROFS_VLE_DI_D0_CBLKCNT-1 if
> > > > + * the uncompressed extent size goes above 8MB. This is
> > > > + * OK because if kernel sees another non-head cluster
> > > > + * after going back by |delta[0]| blocks, kernel will
> > > > + * just keep looking back.
> > > > + */
> > >
> > > Would you mind updating this into the kernel comment style, I mean
> > > /*
> > > * ...
> > > */
> > > Instead?
>
> Done
>
> > >
> > > > + if (d0 & Z_EROFS_VLE_DI_D0_CBLKCNT) {
> > > > + di.di_u.delta[0] = max(
> > > > + d0 & (~Z_EROFS_VLE_DI_D0_CBLKCNT),
> > > > + Z_EROFS_VLE_DI_D0_CBLKCNT-1);
> > >
> > > May I ask if it's actually tested with big pcluster feature? It's
> > > lack of cpu_to_le16() convert and even the original
> > > Z_EROFS_VLE_DI_D0_CBLKCNT flag.
>
> Sorry.. It was tested on a Little Endian machine, so I didn't discover
> the missing cpu_to_le16. Added now.
>
> >
> > Sorry this part shouldn't have Z_EROFS_VLE_DI_D0_CBLKCNT flag.
> >
> > Btw, I think a proper change for this might be just:
> > if (d0 > Z_EROFS_VLE_DI_D0_CBLKCNT - 1)
> > di.di_u.delta[0] = le16_to_cpu(Z_EROFS_VLE_DI_D0_CBLKCNT - 1);
> > else
> > di.di_u.delta[0] = cpu_to_le16(d0);
> > Or using max() to simplify above even more a bit.
>
> This would work, but it's not optimal. For example,
> Z_EROFS_VLE_DI_D0_CBLKCNT << 1
> is greater than Z_EROFS_VLE_DI_D0_CBLKCNT, but it does not have the
> 11th bit set.
> Using Z_EROFS_VLE_DI_D0_CBLKCNT-1 in this case would cause the kernel to
> take more hops than necessary when finding the head cluster. A better
> change would be:
>
> if (d0 > Z_EROFS_VLE_DI_D0_CBLKCNT - 1)
> di.di_u.delta[0] = le16_to_cpu(largest number smaller
> than d0 that does not have Z_EROFS_VLE_DI_D0_CBLKCNT bit set);
> else
> di.di_u.delta[0] = cpu_to_le16(d0);
>
> But how do we find "largest number smaller than d0 that does not have
> Z_EROFS_VLE_DI_D0_CBLKCNT bit set" ?
> Simple, clear the Z_EROFS_VLE_DI_D0_CBLKCNT bit, and set all bits
> before that to 1. In code:
> d0 & (~ Z_EROFS_VLE_DI_D0_CBLKCNT) | (Z_EROFS_VLE_DI_D0_CBLKCNT-1)
>
> So final answer:
>
> if (d0 > Z_EROFS_VLE_DI_D0_CBLKCNT - 1)
> di.di_u.delta[0] = le16_to_cpu(d0 & (~
> Z_EROFS_VLE_DI_D0_CBLKCNT) | (Z_EROFS_VLE_DI_D0_CBLKCNT-1));
> else
> di.di_u.delta[0] = cpu_to_le16(d0);
>
That may work for non-compact indexes, but it's somewhat unsafe for
compact indexes (especially compact 2B), since the valid bits for
each lcluster are 14 (12 plus 2-bit lcluster type):
for head lclusters, it stores lclusterofs;
for non-head lclusters, it stores delta0 (lookback distance) or
delta1 (lookforward distance)
conditionally.
That is also why Z_EROFS_VLE_DI_D0_CBLKCNT is set as (1 << 11). So in
order to make them unique, I suggest just don't reuse higher bits even
for non-compact indexes... (I think we could stand just looking back
multiple times instead for such large logical extents...)
Actually "* eg. for 4k page-sized cluster, maximum 4K*64k = 256M)" is
somewhat outdated now after compact-indexes was introduced. If you
have time, could you submit a kernel patch to fix the description
together?
Thanks,
Gao Xiang
^ permalink raw reply [flat|nested] 11+ messages in thread
* [PATCH v3] erofs-utils: lib: Fix 8MB bug on uncompressed extent size
2022-02-07 21:50 ` Gao Xiang
@ 2022-02-08 18:43 ` Kelvin Zhang via Linux-erofs
2022-02-09 1:47 ` Gao Xiang
0 siblings, 1 reply; 11+ messages in thread
From: Kelvin Zhang via Linux-erofs @ 2022-02-08 18:43 UTC (permalink / raw)
To: linux-erofs mailing list, Miao Xie, Fang Wei; +Cc: Kelvin Zhang
Previously, uncompressed extent can be at most 8MB before mkfs.erofs
crashes on some error condition. This is due to a minor bug in how
compressed indices are encoded. This patch fixes the issue.
Signed-off-by: Kelvin Zhang <zhangkelvin@google.com>
---
lib/compress.c | 18 +++++++++++++++++-
1 file changed, 17 insertions(+), 1 deletion(-)
diff --git a/lib/compress.c b/lib/compress.c
index 98be7a2..add95f5 100644
--- a/lib/compress.c
+++ b/lib/compress.c
@@ -97,7 +97,23 @@ static void vle_write_indexes(struct z_erofs_vle_compress_ctx *ctx,
} else if (d0) {
type = Z_EROFS_VLE_CLUSTER_TYPE_NONHEAD;
- di.di_u.delta[0] = cpu_to_le16(d0);
+ /*
+ * If the |Z_EROFS_VLE_DI_D0_CBLKCNT| bit is set, parser
+ * will interpret |delta[0]| as size of pcluster, rather
+ * than distance to last head cluster. Normally this
+ * isn't a problem, because uncompressed extent size are
+ * below Z_EROFS_VLE_DI_D0_CBLKCNT * BLOCK_SIZE = 8MB.
+ * But with large pcluster it's possible to go over this
+ * number, resulting in corrupted compressed indices.
+ * To solve this, we replace d0 with
+ * Z_EROFS_VLE_DI_D0_CBLKCNT-1.
+ */
+ if (d0 > Z_EROFS_VLE_DI_D0_CBLKCNT) {
+ di.di_u.delta[0] = cpu_to_le16(
+ Z_EROFS_VLE_DI_D0_CBLKCNT-1);
+ } else {
+ di.di_u.delta[0] = cpu_to_le16(d0);
+ }
di.di_u.delta[1] = cpu_to_le16(d1);
} else {
type = raw ? Z_EROFS_VLE_CLUSTER_TYPE_PLAIN :
--
2.35.0.263.gb82422642f-goog
^ permalink raw reply related [flat|nested] 11+ messages in thread
* Re: [PATCH v3] erofs-utils: lib: Fix 8MB bug on uncompressed extent size
2022-02-08 18:43 ` [PATCH v3] " Kelvin Zhang via Linux-erofs
@ 2022-02-09 1:47 ` Gao Xiang
0 siblings, 0 replies; 11+ messages in thread
From: Gao Xiang @ 2022-02-09 1:47 UTC (permalink / raw)
To: Kelvin Zhang; +Cc: Miao Xie, linux-erofs mailing list
On Tue, Feb 08, 2022 at 10:43:17AM -0800, Kelvin Zhang wrote:
> Previously, uncompressed extent can be at most 8MB before mkfs.erofs
> crashes on some error condition. This is due to a minor bug in how
> compressed indices are encoded. This patch fixes the issue.
>
> Signed-off-by: Kelvin Zhang <zhangkelvin@google.com>
> ---
> lib/compress.c | 18 +++++++++++++++++-
> 1 file changed, 17 insertions(+), 1 deletion(-)
>
> diff --git a/lib/compress.c b/lib/compress.c
> index 98be7a2..add95f5 100644
> --- a/lib/compress.c
> +++ b/lib/compress.c
> @@ -97,7 +97,23 @@ static void vle_write_indexes(struct z_erofs_vle_compress_ctx *ctx,
> } else if (d0) {
> type = Z_EROFS_VLE_CLUSTER_TYPE_NONHEAD;
>
> - di.di_u.delta[0] = cpu_to_le16(d0);
> + /*
> + * If the |Z_EROFS_VLE_DI_D0_CBLKCNT| bit is set, parser
> + * will interpret |delta[0]| as size of pcluster, rather
> + * than distance to last head cluster. Normally this
> + * isn't a problem, because uncompressed extent size are
> + * below Z_EROFS_VLE_DI_D0_CBLKCNT * BLOCK_SIZE = 8MB.
> + * But with large pcluster it's possible to go over this
> + * number, resulting in corrupted compressed indices.
> + * To solve this, we replace d0 with
> + * Z_EROFS_VLE_DI_D0_CBLKCNT-1.
> + */
> + if (d0 > Z_EROFS_VLE_DI_D0_CBLKCNT) {
Thanks for the new version.
I think this part would be "if (d0 >= Z_EROFS_VLE_DI_D0_CBLKCNT)",
and already applied with minor styling changes:
https://git.kernel.org/pub/scm/linux/kernel/git/xiang/erofs-utils.git/commit/?id=789ac9b03c2c0d27c5be81cb8d026e2300ae822e
Thanks,
Gao Xiang
> + di.di_u.delta[0] = cpu_to_le16(
> + Z_EROFS_VLE_DI_D0_CBLKCNT-1);
> + } else {
> + di.di_u.delta[0] = cpu_to_le16(d0);
> + }
> di.di_u.delta[1] = cpu_to_le16(d1);
> } else {
> type = raw ? Z_EROFS_VLE_CLUSTER_TYPE_PLAIN :
> --
> 2.35.0.263.gb82422642f-goog
^ permalink raw reply [flat|nested] 11+ messages in thread
end of thread, other threads:[~2022-02-09 1:47 UTC | newest]
Thread overview: 11+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-12-22 2:03 [PATCH v1] erofs-utils: lib: Fix 8MB bug on uncompressed extent size Kelvin Zhang via Linux-erofs
2021-12-22 3:21 ` Gao Xiang
2022-01-18 23:46 ` Kelvin Zhang via Linux-erofs
2022-01-19 3:49 ` Gao Xiang
2022-02-06 23:59 ` Gao Xiang
2022-02-07 2:07 ` Gao Xiang
2022-02-07 17:38 ` Kelvin Zhang via Linux-erofs
2022-02-07 21:50 ` Gao Xiang
2022-02-08 18:43 ` [PATCH v3] " Kelvin Zhang via Linux-erofs
2022-02-09 1:47 ` Gao Xiang
2022-02-07 17:39 ` [PATCH v2] " Kelvin Zhang via Linux-erofs
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.