| Message ID | 20200624032712.23263-1-jsavitz@redhat.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | mm/page_alloc: fix documentation error and remove magic numbers | expand |
On Tue, Jun 23, 2020 at 11:27:12PM -0400, Joel Savitz wrote: > In addition, this patch replaces the magic number bounds with symbolic > constants to clarify the logic. Why do people think this kind of thing makes the code easier to read? It actually makes it harder. Unless the constants are used in more than one place, just leave the numbers where they are. > @@ -7852,6 +7852,9 @@ void setup_per_zone_wmarks(void) > * 8192MB: 11584k > * 16384MB: 16384k > */ > +static const int MIN_FREE_KBYTES_LOWER_BOUND = 1 << 7; > +static const int MIN_FREE_KBYTES_UPPER_BOUND = 1 << 18; > + > int __meminit init_per_zone_wmark_min(void) > { > unsigned long lowmem_kbytes; > @@ -7862,10 +7865,10 @@ int __meminit init_per_zone_wmark_min(void) > > if (new_min_free_kbytes > user_min_free_kbytes) { > min_free_kbytes = new_min_free_kbytes; > - if (min_free_kbytes < 128) > - min_free_kbytes = 128; > - if (min_free_kbytes > 262144) > - min_free_kbytes = 262144; > + if (min_free_kbytes < MIN_FREE_KBYTES_LOWER_BOUND) > + min_free_kbytes = MIN_FREE_KBYTES_LOWER_BOUND; > + if (min_free_kbytes > MIN_FREE_KBYTES_UPPER_BOUND) > + min_free_kbytes = MIN_FREE_KBYTES_UPPER_BOUND; The only thing I'd consider changing there is replacing 262144 with 256 * 1024. 1 << 18 is not clearer!
On Wed, Jun 24, 2020 at 12:12:55PM +0100, Matthew Wilcox wrote: > On Tue, Jun 23, 2020 at 11:27:12PM -0400, Joel Savitz wrote: > > In addition, this patch replaces the magic number bounds with symbolic > > constants to clarify the logic. > > Why do people think this kind of thing makes the code easier to read? > It actually makes it harder. Unless the constants are used in more > than one place, just leave the numbers where they are. > > > @@ -7852,6 +7852,9 @@ void setup_per_zone_wmarks(void) > > * 8192MB: 11584k > > * 16384MB: 16384k > > */ > > +static const int MIN_FREE_KBYTES_LOWER_BOUND = 1 << 7; > > +static const int MIN_FREE_KBYTES_UPPER_BOUND = 1 << 18; > > + I think these constants would look better if declared as an enum. > > int __meminit init_per_zone_wmark_min(void) > > { > > unsigned long lowmem_kbytes; > > @@ -7862,10 +7865,10 @@ int __meminit init_per_zone_wmark_min(void) > > > > if (new_min_free_kbytes > user_min_free_kbytes) { > > min_free_kbytes = new_min_free_kbytes; > > - if (min_free_kbytes < 128) > > - min_free_kbytes = 128; > > - if (min_free_kbytes > 262144) > > - min_free_kbytes = 262144; > > + if (min_free_kbytes < MIN_FREE_KBYTES_LOWER_BOUND) > > + min_free_kbytes = MIN_FREE_KBYTES_LOWER_BOUND; > > + if (min_free_kbytes > MIN_FREE_KBYTES_UPPER_BOUND) > > + min_free_kbytes = MIN_FREE_KBYTES_UPPER_BOUND; > > The only thing I'd consider changing there is replacing 262144 with 256 > * 1024. 1 << 18 is not clearer! >
On Wed, Jun 24, 2020 at 10:07:27AM -0400, Rafael Aquini wrote: > On Wed, Jun 24, 2020 at 12:12:55PM +0100, Matthew Wilcox wrote: > > On Tue, Jun 23, 2020 at 11:27:12PM -0400, Joel Savitz wrote: > > > In addition, this patch replaces the magic number bounds with symbolic > > > constants to clarify the logic. > > > > Why do people think this kind of thing makes the code easier to read? > > It actually makes it harder. Unless the constants are used in more > > than one place, just leave the numbers where they are. > > > > > @@ -7852,6 +7852,9 @@ void setup_per_zone_wmarks(void) > > > * 8192MB: 11584k > > > * 16384MB: 16384k > > > */ > > > +static const int MIN_FREE_KBYTES_LOWER_BOUND = 1 << 7; > > > +static const int MIN_FREE_KBYTES_UPPER_BOUND = 1 << 18; > > > + > > I think these constants would look better if declared as an enum. Why does having to look in two different places make the code clearer?
On Wed, Jun 24, 2020 at 03:09:58PM +0100, Matthew Wilcox wrote: > On Wed, Jun 24, 2020 at 10:07:27AM -0400, Rafael Aquini wrote: > > On Wed, Jun 24, 2020 at 12:12:55PM +0100, Matthew Wilcox wrote: > > > On Tue, Jun 23, 2020 at 11:27:12PM -0400, Joel Savitz wrote: > > > > In addition, this patch replaces the magic number bounds with symbolic > > > > constants to clarify the logic. > > > > > > Why do people think this kind of thing makes the code easier to read? > > > It actually makes it harder. Unless the constants are used in more > > > than one place, just leave the numbers where they are. > > > > > > > @@ -7852,6 +7852,9 @@ void setup_per_zone_wmarks(void) > > > > * 8192MB: 11584k > > > > * 16384MB: 16384k > > > > */ > > > > +static const int MIN_FREE_KBYTES_LOWER_BOUND = 1 << 7; > > > > +static const int MIN_FREE_KBYTES_UPPER_BOUND = 1 << 18; > > > > + > > > > I think these constants would look better if declared as an enum. > > Why does having to look in two different places make the code clearer? > It might not make it clearer in this particular case, because it was easy to take the meaning from the code, but it also doesn't make it harder to read, so I don't have any strong opinion on this case. Joel's approach, however, makes sense if you consider it's generally a good practice to get rid of the unnamed magic numbers anti-pattern. Cheers, -- Rafael
On Wed, Jun 24, 2020 at 10:26:26AM -0400, Rafael Aquini wrote: > Joel's approach, however, makes sense if you consider it's generally a > good practice to get rid of the unnamed magic numbers anti-pattern. But I don't. I care about how easy it is to understand the code, not conforming to some bullshit trendy word salad.
diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 48eb0f1410d4..f725addc2748 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -7832,7 +7832,7 @@ void setup_per_zone_wmarks(void) * Initialise min_free_kbytes. * * For small machines we want it small (128k min). For large machines - * we want it large (64MB max). But it is not linear, because network + * we want it large (256MB max). But it is not linear, because network * bandwidth does not increase linearly with machine size. We use * * min_free_kbytes = 4 * sqrt(lowmem_kbytes), for better accuracy: @@ -7852,6 +7852,9 @@ void setup_per_zone_wmarks(void) * 8192MB: 11584k * 16384MB: 16384k */ +static const int MIN_FREE_KBYTES_LOWER_BOUND = 1 << 7; +static const int MIN_FREE_KBYTES_UPPER_BOUND = 1 << 18; + int __meminit init_per_zone_wmark_min(void) { unsigned long lowmem_kbytes; @@ -7862,10 +7865,10 @@ int __meminit init_per_zone_wmark_min(void) if (new_min_free_kbytes > user_min_free_kbytes) { min_free_kbytes = new_min_free_kbytes; - if (min_free_kbytes < 128) - min_free_kbytes = 128; - if (min_free_kbytes > 262144) - min_free_kbytes = 262144; + if (min_free_kbytes < MIN_FREE_KBYTES_LOWER_BOUND) + min_free_kbytes = MIN_FREE_KBYTES_LOWER_BOUND; + if (min_free_kbytes > MIN_FREE_KBYTES_UPPER_BOUND) + min_free_kbytes = MIN_FREE_KBYTES_UPPER_BOUND; } else { pr_warn("min_free_kbytes is not updated to %d because user defined value %d is preferred\n", new_min_free_kbytes, user_min_free_kbytes);
When I increased the upper bound of the min_free_kbytes value in ee8eb9a5fe863, I forgot to tweak the above comment to reflect the new value. This patch fixes that mistake. In addition, this patch replaces the magic number bounds with symbolic constants to clarify the logic. Suggested-by: John Hubbard <jhubbard@nvidia.com> Suggested-by: Vlastimil Babka <vbabka@suse.cz> Signed-off-by: Joel Savitz <jsavitz@redhat.com> --- mm/page_alloc.c | 13 ++++++++----- 1 file changed, 8 insertions(+), 5 deletions(-)